View
2.058
Download
539
Embed Size (px)
DESCRIPTION
Citation preview
Well Control 11
C O N T E N T S
1 INTRODUCTION1.2 SYSTEMS ANALYSIS OF THE PRODUCTION
SYSTEM1.3 HYDROCARBON PHASE BEHAVIOUR1.4 RESERVOIR INFLOW PERFORMANCE
1.4.1 Liquid Inflow1.4.2 Gas Inflow1.4.3 2 Phase (Gas-Liquid) Inflow1.4.4 Examples of IPRs
1.5 TUBING (OUTFLOW) PERFORMANCE1.5.1 Tubing Pressure Traverse1.5.2 The Tubing Friction Term1.5.3 Introduction to Multiphase Flow in Vertical
Tubing1.5.4 Prediction of Multiphase Fluid Properties
1.6 “GRADIENT” OR PRESSURE TRAVERSECURVES
1.7 FLOW MAPS AND CORRELATIONS1.7.1 Duns and Ros1.7.2 Hagedoorn and Brown1.7.3 Beggs and Brill1.7.4 Gray
1.8 TEMPERATURE MODELLING1.9 SURFACE PRESSURE LOSSES
1.9.1 Surface Components1.9.2 Flow Through Chokes1.9.2.1Single Phase Liquid Flow1.9.2.2Single Phase Gas Flow1.9.2.3Multiphase Flow1.9.3 Gathering System Layout
1.10 COMPLETIONS INFLOW PERFORMANCE1.10.1 Perforated Completions1.10.1.1 Perforation Charge Performance1.10.1.2 Perforation Gun Selection1.10.2 Gravel Packed Completion1.10.2.1 Non-Darcy Turbulence Pressure Losses1.10.2.2 Restriction of Gravel Pack Drawdown
1.11 COMPUTERISED WELL PERFORMANCEPREDICTION PROGRAMS
1.12 WELL PERFORMANCE SENSITIVITYSTUDY EXERCISE1.12.1 Reservoir Inflow and Tubing Outflow
Restrictions1.12.2 Tubing Size and Liquid Loading1.12.3 Effect of Water Cut and Depletion1.12.4 Opportunities for Skin Removal by
Stimulation
1Well Performance1
1.12.5 Completion Design1.12.6 Well Head Pressure
1.13 FURTHER READING
1
2
LEARNING OUTCOMES
Having worked through this chapter the Student will be able to:
• Describe Well Inflow and tubing vertical lift performance.
• Discuss the implementation of these concepts in computerised well completiondesign programs.
• Discuss the need for artificial lift, i.e. the addition of external energy when thenatural reservoir energy is insufficient to continue economic production.
Later modules will extend these concepts by:
• Discussing the many artificial lift techniques and develops selection criteria.
• Describing the design process for a gas lift and electric Submersible Pump system.
Institute of Petroleum Engineering, Heriot-Watt University 3
1Well Performance1
1 INTRODUCTION
A simple producing system is illustrated in figure 1.
Liquid
Gas
Pwh
Vertical Tubing
Separator
Surface
Psep
Pwf
Hyrodocarbon Reservoir
Casing Annulus Isolation Packer
Radial Flow in Porous Media
Horizontal Flow Line
PR
The hydrocarbon fluid flows from the reservoir into the well, up the tubing, along thehorizontal flow line and into the oil storage tank. During this process the fluid’spressure is reduced from the reservoir pressure to atmosphere pressure in a series ofpressure loss processes (Figure 2):
(1) across the reservoir
(2) across the completion (perforation/gravel pack etc.)
(3) across the tubing and any restrictions
(4) across the sub surface safety valve
(5) across the surface choke
(6) across flowline
Figure 1
Simplified hydrocarbon
production system
1
4
These pressure losses can be grouped into three main components:
(7) summarises the total pressure losses in the reservoir and completion
(8) summarises the total pressure losses in the tubing
(9) summarises the total pressure losses at the surface
A pump or compressor are often used to aid evacuation of fluids (gas/water/oil) fromthe separator. The separator is operated under gas pressure control and liquid (oil andwater) level control. Hence it normally acts as the end point of the flowing systemsince a pump is necessary to aid evacuation of the liquids from the separator.
Separator
∆P5=(Pwh-PDSC)
Bottomhole Restriction
Gas
Liquid
∆P9=(Pwh-Psep)
∆P1=(PR-Pwfs)
∆P4=(PUSV-PDSV)
∆P8=
(Pw
f-Pw
h)
∆P3=(PUR-PDR)
∆P2=(Pwfs-Pwf)
PDSCPWH Psep
PDSV
PUSV
PDR
PUR
Pwf Pwfs PR
∆P1=(PR-Pwfs) = Loss in Hydrocarbon Reservoir (Porous Medium) ∆P2=(Pwfs-Pwf) = Loss Across Completion ∆P3=(PUR-PDR) = Loss Across Tubing and any Restrictions ∆P4=(PUSV-PDSV) = Loss Across Safety Valve∆P5=(Pwh-PDSC) = Loss Across Surface Choke∆P6=(PDSC-Psep) = Loss or DownstreamN.B. U refers to Upstream and D to Discharge or DownstreamSUMMARY PRESSURE LOSSES∆P7=(Pwf-PR) = Total Loss in Reservoir and Completion∆P8=(Pwf-Pwh) = Total Loss in Tubing∆P9=(Pwh-Psep) = Total Loss at the Surface
Safety Valve
Surface Choke
∆P6=(PDSC-Psep)
PR
Reservoir PressureP
wfsFlowing sand face Pressure
Pwf
Flowing Bottom Hole PressureP
URUpstream Restriction Pressure
PDR
Downstream Restriction PressureP
USVUpstream Safety Valve Pressure
PDSV
Downstream Safety Valve PressureP
WHWell Head Pressure
PDSC
Downstream surface Choke PressureP
sepSeparator Pressure
Figure 2
Pressure losses during
production
Institute of Petroleum Engineering, Heriot-Watt University 5
1Well Performance1
The magnitude of these individual pressure losses depend on the reservoir propertiesand pressures; fluid being produced and the well design. Production Technologists/Engineers need to understand the interplay of these various factors so as to designcompletions which maximise profitability from the oil or gas production. There areno standard “rules of thumb” which can be used. Figure 3 schematically representsthe pressure distribution across the production system shown in Figure 2. It identifiesthe most significant components, flowline, tubing and the reservoir and completionwhere pressure losses occur.
Table 1 was developed by Duns and Ross (“Vertical flow of gas and liquid mixturesin wells”, Proc Sixth World Petroleum Conference, Frankfurt, Vol 2, paper 22, 1963)to illustrate one possible distribution in a conventional land oil field developed withvertical wells.
2.5
5.0
10.0
15.0
2700
3700
4500
4800
36
25
15
11
57
68
78
82
7
7
7
7
Well Productivity Index (bopd/psi)
Production Rate(bopd)
Pressure Loss Distribution (%)
Across Reservoir and Completion
(∆ P7)
Across Tubing(∆ P8)
Across Flowline(∆ P9)
ReservoirDrainageBoundary
Reservoir
Completion
Sand Face
Wellbore
Tubing Restriction
Safety Valve
Wellhead
Choke
Psep
Pwh
Pwf
PR
Position
Pre
ssur
e
Inflow ∆P7 Well ∆P8 Surface (∆P9)
Table 1
Pressure Loss Distribution
as a Function of Well
Productivity Index
Figure 3
Pressure across production
system
1
6
The corresponding figures for a field developed with horizontal wells (much greaterwell productivity indices), subsea wells (long flowlines, possibly over hilly terrain)and pipelines would have a very different distribution. This is due to the flowoccurring mainly in a horizontal direction rather than vertical orientation associatedwith wells. The balance between the “Elevation terms” and the “Friction terms”across the pipe (i.e. contribution to ∆ P
8) change drastically as shown in table 2.
Pressure Loss Term Elevation Friction
(vertical) well 85–98% 2–15%
(horizontal) pipeline/well 0–30% 70–100%
Well Orientation
1.2 Systems Analysis of the Production SystemThe use of systems analysis to design a hydrocarbon production system was firstsuggested by Gilbert (“Flowing and Gas Lift Performance”, API Drilling andProduction Practices, 1954”). Systems analysis, which has been applied to manytypes of systems of interacting components, consists of selecting a point or nodewithin the producing system (well and surface facilities). Equations for the relationshipbetween flow rate and pressure drop are then developed for the well components bothupstream of the node (inflow) and downstream (outflow). The flow rate and pressureat the node can be calculated since:
(i) Flow into the node equals flow out of the node.
(ii) Only one pressure can exist at the node.
Further, at any time, the pressure at the end points of the system {separator (Psep
) andreservoir pressure (P
R)} are both fixed. Thus:
PR - (Pressure loss upstream components) = P
node(1)
Psep
+ (Pressure loss downstream components) = Pnode
(2)
Operating PointPressure at Node
Pre
ssur
e
Flow Rate Through Node
Flow Rate
Node Inflow
Node Outflow
Figure 4
Node flow rate and
pressure
Table 2
Typical pressure loss
Distributions
Institute of Petroleum Engineering, Heriot-Watt University 7
1Well Performance1
Typical results of such an analysis is shown in Figure 4 where the pressure-raterelationship has been plotted for both the inflow (Equation 1) and outflow (Equation 2)at the node. The intersection of these two lines is the (normally unique) operatingpoint. This defines the pressure and rate at the node. This approach forms the basisof all hand and computerised flow calculation procedures. It is frequently referred toas “nodal analysis”. (This name is also a trademark of Schlumberger TechnologyCorporation for this process).
Having established the concept of nodal analysis, the following sections will discusshow the hydrocarbon phase behaviour (Section 1.3) effects the reservoir fluid InflowPerformance into the well (Section 1.4). The outflow, or tubing performance will bereviewed (Section 1.5 et seq.) and the interaction between the in- and out-flowdiscussed in Section 1.12 which gives a number of well performance sensitivity studies.
1.3 Hydrocarbon Phase BehaviourHydrocarbon reservoir fluids are a complex mixture of hydrocarbon molecules, thecomposition of which is dependent on the source rock, degree of maturation etc. Phasechanges occur when this complex hydrocarbon fluid flows from the (high temperatureand pressure) reservoir environment to the (cool, low pressure) separator conditions.Such changes are sketched for an undersaturated oil in Figure 5. Here it can be seenthat the fluid:
Dew Point Line
BubblePoint Line
Pre
ssur
e
Temperature
Liquid Phase Only
Gas Phase Only
100% Liquid
80%
60%
40%
20%
5% 0% Liquid
Critical Point
(d)
(c)
(b)
(a)(PR,TR)
(Pwf,Twf)
(Psep,Tsep)
Reservoir
Wellbore
(e)
Separator
Two PhaseRegion
(f)
(i) is present as a single phase liquid in the reservoir {point (a)}
Figure 5
Schematic phase diagram
for an undersaturated oil
1
8
(ii) remains a single phase liquid at the wellbore (significant reduction in pressureand small change in temperature during flow in reservoir) {point (b)}
(iii) starts to evolve gas {point (c)} as temperature and pressure are reduced duringflow up the tubing
(iv) evolves increasing amounts of gas {points (d) and (e)} until the separator {point(f)} is reached.
Some or all of the flow regimes illustrated in figure 6 may occur.
The phase behaviour of the hydrocarbon fluid controls the fluid’s gas/liquid ratio asa function of bottom hole pressure. This, in turn, will effect flow rate, i.e. the InflowPerformance Relationship (IPR) discussed in section 1.4 and the outflow tubingperformance.
1.4 Reservoir Inflow PerformanceThe Inflow Performance Relationship (IPR)is routinely measured using bottomholepressure gauges at regular intervals as part of the field monitoring programme. Thisrelationship between flow rate (q) and wellbore pressure (P
wf) is one of the major
building blocks for a nodal-type analysis of well performance.
Institute of Petroleum Engineering, Heriot-Watt University 9
1Well Performance1
Reservoir
(c) Pbubble Tbubble
PRTR (a)
(d)
(e)
Wellhead
(b)
(Pwf)
Pbubble, Tbubble (Bubble Point)
Mist Flow
Annular Flow
Churn Flow
Slug Flow
Bubble Flow
Separator
(f) Psep,Tsep
Single Phase Flow
(d)
(e)
(c)
1.4.1 Liquid InflowField measurements have shown that wells producing undersaturated oil (no gas at thewellbore) or water have a straight line IPR (Figure 7).
q = PI (PR - P
wf) (3)
where q is the flow rate and PI the Productivity Index, i.e. the well inflow rate per unitof well drawdown.
Figure 6
Schematic view of possible
phase changes in tubing
1
10
q
Pwf
AOF
(Reservoir Pressure)
WellDrawdown
Liquid Flow Rate (q)
Wel
l Bor
e F
low
ing
Pre
ssur
e (P
wf)
q (max)
PR
Well Production
A theoretical basis for the straight line IPR can be derived using Darcy’s Law, radialinflow into the well along with other assumptions about rock and fluid properties. PIis a useful tool for comparing wells since it combines all the relevant rock, fluid andgeometrical properties into a single value to describe (relative) inflow performance.
The Absolute Openhole Factor (AOF or qmax
), is the flowrate at zero (bottomhole),wellbore flowing pressure. AOF, although often representing unrealistic conditions,is a useful parameter when comparing wells within a field since it combines PI andreservoir pressure in one number representative of well inflow potential.
A straight line IPR can be determined from two field measurements:
(i) the stabilised bottomhole pressure with the well shut in {reservoir pressure of (PR)}
(ii) the flowing, bottom hole, wellbore pressure (Pwf
) at one production rate
The well’s inflow potential can then be calculated at any draw-down (or Pwf
)
1.4.2 Gas InflowThe compressible nature of gas results in the IPR no longer being a straight line.However, the extension of this steady state relationship derived from Darcy’s Law,using an average value for the properties of the gas between the reservoir and wellbore,leads to
q = C (PR
2 - Pwf
2) (4)
where C is a constant
This relationship is valid at low flow rates, but becomes invalid at higher flow ratessince non-Darcy (or turbulent) flow effects begin to be observed. This can beaccounted for by use of the “Bureau of Mines” equation that was developed from fieldobservations:
Figure 7
Staightline IPR (for an
incompressible liquid)
Institute of Petroleum Engineering, Heriot-Watt University 11
1Well Performance1
q = C (PR
2 - Pwf
2)n (5)
where 0.5 <n <1.0
A log-log plot of q versus (PR
2 - Pwf
2) yields a straight line of slope n and intercept C.Standard practice for testing gas wells is to measure the bottom hole, flowing,wellbore pressure (P
wf) at four production rates. Figure 8 shows that the change of
slope from the initial value of 1 (Darcy flow, equation [4]). Non-Darcy flow effects(equation [5], n<1) are observed at the two higher rates.
Both equations [4] and [5] are illustrated in Figure 8 which shows the >50% reductionin AOF (from 1.4 to 0.9) due to these non-Darcy flow effects.
AOFPR
2
PR
2 -P
wf2
Reduction Due to Non-Darcy Flow
N
on-D
arcy
Flo
w
D
arcy
Flo
w
Gas Flow Rate (qG)(logarithmic scale).
0.1 1.0 10Test Rates
(
)
1.4.3 2 Phase (Gas-Liquid) InflowStraight line IPR (Section 1.4.1) are also not applicable to when two phase inflow istaking place, e.g. when saturated oil is being produced. Vogel (“Inflow PerformanceRelationships in Solution-Gas Drive Wells”, J Pet Tech, 1968, 83-92) proposed thefollowing equation based on a large number of well performance simulations:
Figure 8
Gas well deliverability
reduced by non-Darcy flow
pressure losses
1
12
q
qP P P Pwf R wf R
max
. / . /= − ( ) − ( )1 0 2 0 82
(6)
where qmax
is the AOF, i.e. q when Pwf
= 0
Vogel’s key contribution was the introduction of the concept of normalising theproduction rate to the AOF value (q
max). Rewriting equation [6] in this manner gives:
q
qP Pwf R
n
max
/= − ( ){ }12
(7)
which is virtually equivalent to Vogel’s equation when n = 1 (Fetkovitch, “Isochronaltesting of oil wells”, SPE 4529, Las Vegas, Sept 1973). i.e:
q
qP Pwf R
max
/= − ( )12
(8)
When multirate test data is available then equation [7] is preferred since it includeshigh rate (non-Darcy or turbulent) effects. This is best done by plotting the data in asimilar manner to Figure 8, the resulting staight line has a slope of 1/n.
Figure 9 compares the production rate as a function of drawdown for an undersaturatedoil (straight line IPR, line A) and a saturated oil showing the two phase flow effectsdiscussed above (curve B). The figure also shows the special case (curve C) when thewellbore pressure is below the bubble point while the reservoir pressure is above, i.e.(incompressible) liquid flow is occurring in the bulk of the reservoir.
Nor
mal
ised
Wel
lbor
e F
low
ing
Pre
ssur
e (P
wf/P
R)
Oil Flow Rate
qb max qc max qa max
A
C
B
Bubble Point
Reservoir
Pressure (PR) A Straight line IPR (undersaturated oil)
B Vogel or Curved IPR (saturated oil)
C Combination of A and B when reservoir pressure was above the bubble point.IPR becomes curved at the bubble point
1.0
Figure 9
Inflow performance
relationships
Institute of Petroleum Engineering, Heriot-Watt University 13
1Well Performance1
1.4.4 Examples of IPRs
(i) Reservoir Depletion
The previous section discussed the value of normalising the IPR. This is illustratedwith data on the IPR of a saturated (or solution gas drive) oil reservoir. Figure 10ashows that the IPR rapidly decreases with increasing cumulative oil recovery. Thisis not only due to reservoir pressure depletion; but is also related to the increasinggas saturation which is making oil flow progressively more difficult.
Wel
lbor
e F
low
ing
Pre
ssur
e
0.1%2%
4%
6%
8%
10%
12%14%
Production Rate
Cumulative
Oil R
ecovery
A plot of the same data in a normalised manner (Figure 10b) shows that the curvesare quite similar, the increasing gas saturation being responsible for the (relatively)greater drawdown for similar normalised production rate.
Nor
mal
ised
Wel
lbor
e F
low
ing
Pre
ssur
e (P
wf/P
R) 1.0
0 0.2 0.4 0.6 0.8 1.0
0.8
0.6
0.4
0.2
0
Normalised Production Rate (q/q max)
Cumulative Recovery = 0.1%, 2%,4%,
6%, 8%
10%
12%
14%
(ii) Crude Oil Properties
Figure 10a
IPR curves
Figure 10b
Normalised IPR curve
1
14
Consider the production of crude oil A, which is significantly more viscous thancrude oil B. Providing all other factors were kept constant, the IPR curves wouldshow substantial differences (Figure 11a) while the normalised IPR curves areessentially the same (Figure 11b).
Wel
lbo
re F
low
ing
Pre
ssu
re (
Pw
f)
Production Rate (q)
Oil A is More Viscous than Oil B
A
B
No
rmal
ised
Wel
lbo
re F
low
ing
Pre
ssu
re (
Pw
f/P
R)
1.0
0 0.2 0.4 0.6 0.8 1.0
0.8
0.6
0.4
0.2
0
Normalised Production Rate (q/q max)
A
B
Figure 11b is a simple method to estimate the future well IPR as the reservoirundergoes pressure depletion.
(iii) Heterogeneous Formations
All the above refers to a well producing from a homogeneous reservoir. Frequently,wells are completed on heterogeneous formations where production from severaldifferent zones is commingled. Reservoirs with differing permeabilities will bedepleted at different rates - the resulting composite IPR being the sum of the separateindividual IPRs (Figure 12). It will change as the well depth, fluid type, productionrate etc. alter.
Figure 11a
Actual IPR's
Figure 11b
Dimensionless IPR
Institute of Petroleum Engineering, Heriot-Watt University 15
1Well Performance1
Production Rate
Wel
lbor
e F
low
ing
Pre
ssur
e (P
wf)
0
Composite IPR forall three Zones
100-md zone
10-md zone
1-md zone
1.5 Tubing (Outflow) PerformanceChapter 1.4 discussed the inflow of reservoir fluids into the wellbore and the pressuredrop required to achieve this. The outflow pressure drop required to lift a fluid fromthe perforations to the wellhead and then the separator is the second factor whichdetermines the well production. This outflow performance will now be discussed.
Starting from the top of the well, the parameters which contribute to the pressure atthe bottom of the well are:
(i) the (back) pressure at the wellhead
(ii) the hydrostatic head between the wellbore and wellhead. This is a function ofthe change in elevation between the wellhead and the wellbore and the averagedensity of the fluid in tubing all multiplied by the acceleration due to gravity.
(iii) the pressure loss required to overcome friction losses due to viscous drag. Thisdepends on the fluid’s flow rate, flow regime and its viscous properties as well as thelength, diameter and roughness of the tubing.
NB - Pressure losses due to acceleration of the expanding fluid are normally lowand can be neglected.
Figure 13 illustrates the relative importance of these factors and their interaction ofthese components for a given well depth as a function of flow rate and fluid type.
(i) Figure 13a is for a single (incompressible) liquid production. Being dense, thehydrostatic head component is relatively large and constant (the density of water andthe heavier crudes, e.g. 20˚API, shows only minor variations with pressure andtemperature changes typically found in producing oil wells). The friction compo-nent increases rapidly, once turbulent flow is achieved, after the erratic behaviourwhen the transition region between laminar and turbulent flow has been passed.
Figure 12
Composite IPR for
heterogenuous formation
1
16
(ii) Figure 13b is for a gas well. The hydrostatic head component is now muchsmaller, but increases with depth and rate since gas density is very pressuredependent. Frictional pressure losses are normally the most important component,turbulent flow being encountered even at low flow rates.
(iii) Multiphase (gas/liquid) production is illustrated in Figure 13c. The variationof friction and hydrostatic pressure losses with production rate is complicated; theirrelative importance may change, depending on the exact conditions.
NB It has been assumed that the reader is familiar with the basic fluid mechanicsequations that describe flow in pipes through the laws of:
(i) conservation of mass and momentum (for calculating pressure changes) and;
(ii) conservation of mass and energy (for calculating enthalpy, and hence temperaturechanges).
Bot
tom
hole
Pre
ssur
e
Bot
tom
hole
Pre
ssur
e
Bot
tom
hole
Pre
ssur
e
Liquid Rate(a)
Gas Rate(b)
Liquid Rate(c)
TurbulentTransition
Laminar
Friction
Hydrostatic Head Friction
Friction Hydrostatic Head
Wellhead Pressure Wellhead Pressure Wellhead Pressure
FullyTurbulent
Liquid Gas Multiphase Liquid /Gas Mixture
Pressure(a)
Pressure(b)
Pressure(c)
Wel
lhea
d P
ress
ure
Wel
lhea
d P
ress
ure
Wel
lhea
d P
ress
ure
Hyd
rost
atic
Hea
d
Hyd
rost
atic
Hea
d
Hyd
rost
atic
Hea
d
Friction
Friction
Friction
DepthWellheadDepth Depth
Liquid Gas Multiphase Liquid /Gas Mixture
Figure 13
Components of tubing
pressure loss for different
fluids
Figure 14
Tubing pressure traverse
Institute of Petroleum Engineering, Heriot-Watt University 17
1Well Performance1
1.5.1 Tubing Pressure TraverseThese differences between the three fluid systems are also apparent when the changein pressure as a function of depth at a constant well production rate is considered(Figure 14). This plot is known as a “Tubing Pressure Traverse” and the change intotal pressure with depth is known as a “gradient” curve. The behaviour of theindividual pressure components which make up the final gradient curve is summa-rised below.
(i) Figure 14a is for a single (incompressible) liquid. The hydrostatic head andfriction components, are straight lines. This is because the fluid density and thefriction loss per unit tubing length remain constant over the complete tubing length(the latter assumes no restriction in the tubing).
(ii) Figure 14b is for a gas well. The (relative) contribution of the hydrostatic headcomponent increases with depth since the gas density increases with the totalpressure (i.e. well depth). The friction component shows the reverse behaviour - thegas velocity is greatest at shallow depth since the pressure is lowest at this point andthe same amount of gas is entering and exiting the tubing. Thus the ratio:
(friction pressure/hydrostatic head) per unit length of tubing
increases as the well depth decreases i.e. the importance of the friction component isless at greater well depths for gas flow in a constant diameter tubing.
The opposing behaviour of the friction and hydrostatic head components with depthresults in the (total) pressure traverse approximating a straight line.
This conclusion is only true for a single tubing diameter. Restrictions in gas wells canoften lead to unacceptably high pressure losses due to the consequent large increasein fluid velocity.
(iii) A simple description of multiphase flow (Figure 14c) does not exist sincesimple analytical equations etc are not available for this complex flow regime; butthe overall shape of Figure 14c is between the two earlier figures.
NB. In all the above cases the only parameter that is under the operational control ofthe production engineer is the wellhead pressure or system “back pressure”. Theremainder of the completion can only be influenced by the engineer at the design stage.This is thus the time when a wide range of sensitivity analyses should be performedin order to ensure that the installed well will be “fit for purpose” during its lifetime.Use of a standardised well design in a field can bring significant cost savings.However, these have to be balanced against the costs e.g. foregone production, extraworkovers to change tubing size, etc, that a non-optimum well completion will bring.Reduction in total, lifetime unit costs of constructing and operating the well is the aimof the production engineer, optimising the profitability of field development.
Section 1.5.2 will discuss calculation of pressure losses in pipes. Section 1.5.3discusses the importance of the phase behaviour of the hydrocarbon fluid followed bythe introduction of “Gradient Curves” as a simple means of describing multiphase(outflow) tubing performance (Section 1.6).
1
18
Laminar, Highly Non-Newtonian
Turbulant, Newtonian
Pipe Wall
V /
V m
ax
Pipe Centre
Laminar, Newtonian
1.0
1.5.2 The Tubing Friction TermFigs 13 and 14 schematically indicated the importance of the friction component whenpredicting the pressure at any point in the wellbore. This frictional pressure will bea function of the fluid characteristics (Newtonian or non-Newtonian fluid viscosities),fluid flow conditions (velocity and laminar or turbulent flow) and the properties of thetubing (diameter and roughness).
A full fluid mechanical description of all the situations that are encountered inProduction Engineering is beyond the scope of this text - however, this section isdesigned to introduce some basic concepts for the simplest case - single phase flowof an incompressible, Newtonian fluid.
(i) Reynolds Number
The Reynolds Number (Re) is the ratio of the inertial forces to the viscous forces for
fluid (density, ρ, and viscosity, µ) flowing in a circular pipe (diameter, D)
RDv
e = ≡µ
ρµ
ρ
vv / D
or inertial forcesviscous forces
2
where v is the average fluid velocity.
The velocity profile of a Newtonian fluid flowing in LAMINAR flow is shown infigure 15. Laminar flow is characterised by the individual fluid particles movingONLY in the flow direction with no fluid movement across the pipe, i.e. the fluid canbe pictured as flowing in a series of concentric tubes with the maximum velocity at thepipe centre and a minimum velocity at the pipe wall. On the other hand, TURBU-LENT flow is characterised by rapidly fluctuating flow velocity components inrandom directions.
Figure 15
Velocity profiles in laminar
and turbulent pipe flow
Institute of Petroleum Engineering, Heriot-Watt University 19
1Well Performance1
Newtonian fluids are defined as having a viscosity that is independent of shear rate.Non-Newtonian fluids, by contrast, have a shear rate dependent viscosity, e.g. a shearthinning fluid has an apparent viscosity which decreases as the shear rate increases.They also have a very different velocity profile across the tubing. The third profilein Figure 15 is for a highly shear thinning, non-Newtonian fluid. This type of fluidis encountered in hydraulic fracturing stimulations and gravel pack operations. Theirunusual behaviour allows high concentrations of solid material (gravel pack sand orproppant) to remain in suspension at low shear rates while the, apparently highlyviscous, fluid can be pumped through the tubing with lower pressure drops than wouldbe expected if pure water was being pumped.
Not only will the frictional pressure drop across a length of tubing be different betweenthe laminar and turbulent cases, but the flow velocity profiles will have significantconsequences for several production operations (Figure 16). For example:
Fluid 1
Fluid 2Fluid 2
Fluid 1
MixtureFluid 1 & 2
Fluid 1
Fluid 2
Direction of Flow
Cross Sections Both FluidsTurbulent Laminar Cross Sections
LimitedMixingZone
Extensive contamination of fluid1 by fluid 2, in a limitedmixing zonebut bulk of fluids still remain separate
Inter penetration offluid 1 by fluid 2
occurs overa largetubinglength
(i) When pumping a series of fluids into a well which should not mix e.g. for a sandcontrol treatment; the mixing zone between the fluids will be small if both fluids arein turbulent flow. On the other hand, laminar flow allows the centre portion of thetrailing fluid to penetrate a long way into the leading fluid. This results in the twofluids arriving simultaneously at the bottom of the tubing and being mixed duringinjection into the perforation.
(ii) Mixing of two fluid streams being combined in a T-piece will only occurrapidly if the flow is turbulent. Laminar flow will result in concentration gradientsoccurring in the transverse direction across the pipe for a substantial distancedownstream of the T-piece.
Figure 16
Mixing of fluid in pipe flow
as a function of flow regime
1
20
Flow Rate (bbl/d)
Rey
nold
s N
umbe
r, R
Ne
1010
100
1000
10000
100 1000
Turbulent Flow
TransitionZone
Laminar Flow
Reynolds Number, vs Flow RateFor 1.0 gm/cc Fluid
34
56
8
34
56
8
34
56
8
Pipe R
adiu
s (in
)
Pipe R
adiu
s (in
)
Pipe R
adiu
s (in
)
1 Centpoise
10 Centpoises
100 Centpoises
Flow Rate (bbl/day)
Rey
no
lds
Nu
mb
er (
Re)
Laminar flow is characterised by low Reynolds Numbers. Turbulent flow firstbecomes apparent at a Reynolds Number of 2100 with fully turbulent flow beingobserved at about 3500 and higher. Figure 17 plots the Reynolds Number as a functionof fluid viscosity, pump rate and pipe radius. It can be seen that, at the typical flowrates encountered in petroleum engineering, fluids with a water-like viscosity arenormally in turbulent flow while viscous oil is in laminar flow.
(iii) Frictional Pressure Drop
Experiments have been made to measure the pressure drop (per unit length of pipe)of a liquid flowing through a pipes of known diameter. The measurements wererepeated with pipes of differing materials and also with smooth wall pipes which hadbeen deliberately treated to create a surface of known roughness. All the possiblecombinations of variables have been studied by varying the flow rate and fluidviscosity as well.
These experiments showed that the frictional pressure drop (∆P) may be calculatedfrom the Fanning equation:
∆Pf v L
Dm= ρ 2
2
where fm is the Moody friction factor, ρ the density of the fluid flowing at a velocity
v in a pipe of length L and Diameter D.
(a) Laminar Flow (Re <2000): The frictional pressure drop is independent of
tubing roughness and is proportional to the fluid velocity. The friction factor (fm) is
inversely proportional to the Reynolds Number
Figure 17
Reynolds number with
volumetric flow rate,
viscosity, and pipe size
Institute of Petroleum Engineering, Heriot-Watt University 21
1Well Performance1
fm = 64/R
e
(b) Turbulent Flow (Re >2000): The frictional pressure drop is very sensitive to the
exact nature of the inner pipe wall as well as to the fluid flow conditions (ReynoldsNumber). Experiments showed that the important parameter was the relative piperoughness (ε/D) where ε refers to the absolute height of roughness features thatprotrude from the pipe surface into the flow stream and D the pipe diameter.
The Chen equation (Chen, “An explicit equation for friction factors in pipes”, Ind.Eng. Chem. Fund., 18, p296, 1979) is probably the easiest equation for calculating thefriction factor:
14
5 04522 8257
7 1491 098 0 8981
f R Rm e e
= − − +
log
3.7065
ε ε.log
... .
∫ =64/R
e
Turbulent Zone
Complete Turbulence, Rough Pipes
SmoothPipes
*TZ = Transition Zone between Laminar Flow and Turbulent Flow.
TZ*LZ*
*LZ = Laminar Zone
103
104
2 3 4 5 6 8 105
2 3 4 5 6 8 106
2 3 4 5 6 8 107
2 3 4 5 6 8 108
2 3 4 5 6 8.008
.009
.01
.015
.02
.025
.03
.04
.05
.06
.07
.08
.09
0.1
.00001
.00005
.0001
.0002
.0004
.0006
.0008
.001
.002
.004
.006
.008
.01
.015
.02
.03
.04
.05
MoodyFrictionFactor (f)
Re = Reynolds Number =
RelativeRoughness
KD
Dvpµ
ε =
It has a similar accuracy to the more normally quoted Colebrook-White equation,which was used to generate Figure 18 (a plot of the Moody friction factor as a functionof Reynolds Number and turbulence).
The (absolute) pipe roughness depends on many factors; the bulk of which theengineer has little control over. These include:
• Pipe metallurgy and any coating materials applied.• Fluid velocity (erosion at high rates) and fluid corrosivity (pH, the presence of
solids, CO2, H
2S etc).
• Deposits (hydrates, paraffins, asphaltenes).• Years in service.
Figure 18
Moody friction factor
diagram
1
22
It comes as no surprise to learn that, in turbulent flow where it is an importantparameter, roughness is normally treated as an empirical parameter which is used asa fitting parameter to match calculated results to actual pressure drop measurements.However, the roughness value used must be realistic. Table 3 quotes typical valuesfor use in these calculations:
Material RoughnessPlastic Pipe or Coating
New TubingDirty Well Tubing
0.00.000050.00075
1.5.3 Introduction to Multiphase Flow in Vertical TubingMultiphase flow would be greatly simplified if the two phases behaved as a homoge-neous mixture whose properties were an appropriately averaged value of the indi-vidual phase properties. However, experiments have shown that this is not the caseand that one fundamental phenomenon occurring in vertical multiphase (oil-gas,water-oil, etc) flow is the concept of SLIP and HOLD UP . These phenomenon aremost important for the gas/liquid case since the density differences are greatest:
(i) SLIP refers to the ability of the less dense (“lighter”) phase to flow at a greatervelocity than the denser (“heavier”) phase.
(ii) HOLD UP is a consequence of slip - the volume fraction of the pipe occupiedby the denser phase is greater than would be expected from the (relative) in - andoutflow of the two phases - since its flow velocity is slower than that for the lightphase.
NB. This accumulation of the denser phase in the pipe is an equilibrium phenomenoni.e. the in- and out-let flow rates of a particular phase flowing in the pipe are the same.
InsituIn/Outflow
Phase RatioPhase Ratio
Phase RatioPhase Ratio
InsituIn/Outflow
VL=VSL, HL=λL, VG=VSG
VL<VSL, HL>λL, VG>VSG
VG
VL
VG
VL
NO SLIP
SLIP
λL HL
λL HL
These concepts can be best understood with the help of the following mathematicaldescription and Figure 19.
Table 3
Typical pipe roughness
values
Figure 19
Volume fraction changes
when slip occurs during
flow
Institute of Petroleum Engineering, Heriot-Watt University 23
1Well Performance1
NB. Subscripts G, L, O and W refer to the gas, liquid, oil and water phasesrespectively, while q is the phase volume flow rate, V the velocity and A
p the cross
sectional area of the pipe.
Superficial phase velocities (VSL
and VSG
) are given by:
VSL
= qL/A
p and V
SG = q
g/A
p
In situ (or actual) velocities (VL and V
G) are given by:
VL = q
L/A
L and V
G = q
G/A
G
where AL and A
G are the actual areas of the pipe occupied by that liquid and phases
respectively.
AL and A
G under NO SLIP conditions can be calculated from the in- and out-flow
phase rates
AL = q
L/(q
L + q
G) and A
G = q
G/( q
L + q
G)
The slip condition can be quantified by the liquid Holdup (HL) defined as the Fraction
of the pipe filled with liquid:
HL = A
L/A
p and H
G = A
G/A
P = 1 - H
L
and the No Slip Holdup:
λL = q
L/(q
L + q
G)
where λL is the input liquid volume fraction.
The relationship between all these variables is illustrated in figure 19.
Hence if slip occurs, then the slip velocity, Vs, is given by:
Vs = V
G – V
L = V
G/H
G - V
L/H
L
1.5.4 Prediction of Multiphase Fluid PropertiesThe above concepts can be used to predict some of the properties of the multi phasemixture using a phase averaging mixing rule e.g. the density of a liquid/gas mixture(ρ
m) is:
ρm = ρ
Lf
L + ρ
G(1 – f
L)
where fL is the liquid volume fraction. This is equal to:
ρm = ρ
Lλ
L + ρ
G{1 – λ
L} NO SLIP
orρ
sm = ρ
LH
L + ρ
G{1 – H
L } SLIP
1
24
other properties, such as viscosity, cannot be predicted by this averaging techniqueand require the use of special correlations.
(i) Liquid/Liquid Flow
Downhole sampling and video has shown that many light oils are flowing as twoseparate phases which only form an emulsion once the fluid is subjected to a highshear rate e.g. in the surface choke. The oil and water are flowing as separated phaseswith one phase will form the continuous phase with the second phase being dispersedas small droplets within this continuous phase. The (oil) volume fraction at whichthe continuous phase will change from being oil to water depends on the oilproperties, in particular the surface tension and the amount and types of any surfaceactive chemicals that the oil contains. The size of the discontinuous phase dropletswill depend on the above oil density, the oil and water fluid properties as well as thetype and amount of shear that the mixture is subjected to.
Most properties of the liquid/liquid mixture can be calculated with the phase averagingmixing rule discussed at the beginning of this chapter. One exception is the viscosity- the emulsion formed when mixing oil and water have been experimentally observedwith viscosities up to 50 times greater than that predicted from the averaging rule(Figure 20). The height of this viscosity maximum is dependent on:
(a) the extent and type of shear imposed on the oil-water mixture and
(b) the type of emulsion formed; loose, (i.e. easy to separate), medium ortight (i.e. difficult to separate).
N.B. The increased viscosity of low API gravity crude oils often leads to greaterseparation problems in the surface facilities.
0 1.0Water Fraction (fω)
Vis
cosi
ty
µm = µo fo + µω (1- fo) µω
µo
Figure 20
Schematic representation of
the viscosity of water / oil
mixtures
Institute of Petroleum Engineering, Heriot-Watt University 25
1Well Performance1
Studies have also shown that the oil concentration at which the emulsion changes frombeing oil phase continuous to being water phase continuous is related to the oilviscosity (and hence density). It can thus be seen that the emulsion properties are quitespecific to the system being considered e.g. in one field use of an electric submersiblepump to artificial lift a viscous oil/water mixture can result in breakage of the pumpdrive shaft at oil/water ratios near the inversion point (or zone of maximum viscosity).These problems are often not observed during earlier production experience using aRod Pump - this is because of the greater shear being imposed by the rapidly rotating,centrifugal impeller of the electric submersible pump was required to form the highviscosity emulsion. Specific laboratory tests can be very useful in ensuring a properdesign is made if the emulsion viscosity is potentially a critical parameter in theproduction system.
Liquid/liquid flow in tubings shows many of the flow phenomena which are discussedin the following section on gas/liquid flow; though the range of flow regimesobservable during gas/liquid flow is much greater due to the greater density contrastbetween the phases and the greater velocity associated with gas flow.
(ii) Gas/Liquid (Oil) Flow
Figs 5 and 6 introduced the concept of phase changes to the hydrocarbon fluid as ittravels up the tubing. The following is an elaboration of these ideas and introducesthe concept of a flow map to describe multiphase tubing flow which will be combinedwith the multiphase flow concepts described in the previous chapter.
Figs 5 and 6 relate to the favourable production case when the wellbore flowingpressure (P
wf) is greater than the bubble point i.e. single phase oil is entering the well.
Single phase flow in the tubing will continue until the pressure (and temperature)reduce sufficiently that the bubble point is reached (point c).
The flow patterns in the tubing that will result from this gas bubble formation is afunction of:
(a) gas and liquid flow rates
(b) pipe angle of inclination
(c) pipe diameter
(d) phase densities
(A) FLOW IN VERTICAL TUBING (Figure 21)
(i) at the point (c) that the first gas bubbles appear the fluid mixture’s velocity inthe tubing will increase and the average fluid density decrease.
(ii) The initially formed bubbles will be widely dispersed within the liquid.Continued flow up the tubing results in a further pressure reduction, increasing thenumber of bubbles - which still remain widely dispersed in a continuous liquid phase.This is called “bubble flow” regime.
1
26
Reservoir
(c) Pbubble Tbubble
PRTR (a)
(d)
(e)
Wellhead
(b)
(Pwf)
Pbubble, Tbubble (Bubble Point)
Mist Flow
Annular Flow
Churn Flow
Slug Flow
Bubble Flow
Separator
(f) Psep,Tsep
Single Phase Flow
(d)
(e)
(c)
(iii) Further upward movement of the produced fluid generates an increasing volume(and mass) of gas phase; with a corresponding reduction in the mass (and volume) ofthe liquid phase. Intense mixing will ensure the gas and liquid phases remain inequilibrium as the pressure reduces i.e. the composition of the gas phase will changewith the evaporation of progressively higher molecular weight, hydrocarbon mol-ecules. Availability of an “equation of state” for describing the reservoir fluids PVTproperties allows “flash” calculations to be carried out, i.e. to calculate the composi-tion of the liquid and gas phases at any required temperature and pressure.
Figure 21
Schematic view of phase
changes in tubing
Institute of Petroleum Engineering, Heriot-Watt University 27
1Well Performance1
(iv) The increasing tubing volume fraction occupied by the gas allows bubblecoalescence to occur to such an extent that they fill the entire pipe cross section andform a slug i.e. a series of very large gas bubbles of roughly constant size separatedfrom each other by areas of liquid containing smaller gas bubbles. This is called the“slug flow regime”.
Gas slugs can act as an efficient mechanism to lift liquid to the surface.
(v) Velocity increases, associated with continued expansion of the available gasand further volume increase of the gas phase, eventually result in the large gas slugbreaking up into a wider range of gas bubble sizes. This is called “churn flow”, ahighly turbulent flow pattern associated with oscillating liquid flows. This trend maycontinue so that the phases become dispersed within one another, i.e. neither iscontinuous. This has been called “froth flow” (not illustrated in Figure 6).
(vi) Further upward fluid flow continues the gas liberation and expansion processesso that the phases separate into a central, high velocity core of gas with a continuousfilm of liquid on the tubing wall - the “Annular flow” regime.
(vii) Shear at the gas/liquid interface resulting from continually increasing gasvelocities will eventually destroy the annular ring of liquid on the tubing wall anddisperse it as a “mist” of small droplets - the “Mist flow” regime.
NB. The high velocities experienced near the surface can result in the increase in thefrictional pressure gradient exceeding the decrease in the hydrostatic head pressuregradient, so that the pressure in the tubing may increase as the depth (and rate)decreases (e.g. Figure 13c).
These flow regime transitions have been studied both theoretically and experimentally(by visually observing the flow regime as a function of gas and liquid velocity in atransparent, vertical column). Correlations are generated so that the boundariesbetween the various flow regimes can be plotted on a Flow Map (Figure 22).
Bubble
Churn
Mist
AnnularSlug
Dimensionless Gas Velocity
Dim
ensi
onle
ss L
iqui
d V
eloc
ity
Figure 22
Example flow map
1
28
(B) FLOW IN INCLINED TUBING
The complex description of gas/liquid multiphase flow in vertical pipes is simplifiedby the fact that the low density (gas) phase is tending to rise (due to its low density)in the same direction as the overall flow. This is not the case for inclined or horizontalflow - under these conditions it is much easier for the gas to separate from the liquidand for the difference between the actual and superficial phase velocities to becomemuch greater than for the corresponding vertical flow conditions. This naturally altersthe flow regime as the pipe’s angle of inclination (θ) increases from the vertical. Asecond effect is that the tubing length (L) becomes greater than H (the vertical depth)as θ increases.
L = H/cosθ
The hydrostatic head component of the total downhole will also tend to increase withincreasing deviation angle, θ, since, under most conditions, the average fluid densitywill increase due to an increase in the liquid hold up (H
L).
(C) FLOW IN HORIZONTAL WELLS AND FLOW LINES (Figure 23)
Stratified
Smooth
Annular
Wavy
Elongated Bubble
Slug
Mist
Bubble
Intermittent
DistributedFigure 23
Horizontal pipe liquid / gas
flow patterns
Institute of Petroleum Engineering, Heriot-Watt University 29
1Well Performance1
All the above trends (observed in inclined tubings) become progressively moreextreme as the angle θ increases to 90º - a horizontal flow line. Here the hydrostatichead component is of minor importance while the tendency for phase separation dueto density difference is at its greatest. Experiments have been carried out intransparent pipes and have identified the following regimes.
Stratified - smooth - wavy
Dispersed bubble
Intermittent - elongated bubble - slug
Annular - annular mist - annular wavy
increasingvelocity
increasingliquid
fraction
Large charges are observed in the flow pattern when the pipe inclination anglechanges from +1˚ to -1˚ under stratified or (relatively) low velocity flow conditions.This can be particularly important when considering the flow regimes in horizontalwells (which are never exactly horizontal and whose liner/casing has a greaterdiameter than normal production tubings).
The above, empirical description of multiphase flow has discussed the phenomenaidentified by experimental studies. The object of these studies was:
(i) to produce a flow map which delineated the boundaries between the differentflow regimes and
(ii) to develop a correlation between pressure drop and liquid and gas phaseproperties and velocities, as a function of tubing diameter, within that flow regime.
The combination of these two factors allow the calculation of the tubing outflowperformance.
Different investigators have published flow maps (which are associated with theirnames) and experimental correlations which, unfortunately, can be contradictory i.e.under certain flow conditions they predict different flow regimes (and pressuredrops). Flow maps and correlations (along with recommendations) will be discussedin greater detail in chapter 1.7. These pressure drop calculations were initiallyimplemented using hand calculation procedures. Nowadays, several (commercial)computer programs are available to rapidly and easily complete these complicatedcalculations (chapter 1.11). Prior to the widespread availability of computers orelectronic calculators, use was made of “gradient” curves which greatly simplified thecalculation process.
1.6 “Gradient” or Pressure Traverse CurvesGradient curves were originally proposed by Gilbert (“Flowing and Gas Lift WellPerformance”, Drilling and Production Practice, API, 1954). Gradient curvescorrelate pressure drop as a function of tubing length (Figure 24). Field experiencelead Gilbert to identify that the main factors in controlling vertical multiphase flow
1
30
were tubing diameter, oil rate and gas/liquid ratio. His curves were developed usingfield data. However, later curves published by other investigators are based onlaboratory experimental data and flow maps.
Tubing Size, In. : 1.995
Liquid Rate, STBL/D : 500
Water Fraction : 0
0 4 8 12 16 20 24 28 32 36 40 44 48 52 560
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
0
100
200300400
5006008001000
1200
1500
20003000
Pressure, 100 PSIG
Dep
th, 1
000
Ft.
Gas Liquid Ratio
Figure 24 shows that a different gradient curve is required for each tubing size, liquidrate and water fraction. Each graph consists of a series of lines referring to a differentgas/liquid ratio. Figure 25 explains how the gradient curve can be applied to calculatethe flowing bottom hole (P
wf) or, more precisely, the flowing tubing intake pressure
Figure 24
Example "Gradient" or
Pressure Traverse Curve
Institute of Petroleum Engineering, Heriot-Watt University 31
1Well Performance1
for a fixed wellhead pressure. This is then repeated for a series of production rates toderive the Tubing Performance Relationship (TPR).
Pressure
GLR
Pwh Pwf 1
d1
H
q01Dep
th
Req
uire
d Tu
bing
Inta
ke P
ress
ure
Pressure
GLR
Pwh Pwf 2
d2
H
q02Dep
th
Pressure
Oil Flow Rate, q0
GLR
Pwh Pwf 3
d3
H
q03
Dep
th
Pressure
GLR
Pwh Pwf 4
Pwf 1
Pwf 2Pwf 3
Pwf 4
d4
H
q04
q01q02
q03q04
Dep
th
The gradient curve is used as follows:
(i) Select the gradient curve appropriate for the specified oil rate (qo1
), tubing size,gas/liquid ratio and water fraction.
(ii) find the point on the x axis at which the pressure equals the wellhead pressure.Move vertically downwards to find the depth (d
1) on the appropriate gas/liquid ratio
line that corresponds to this wellhead pressure.
(iii) Move downwards by a distance (H), equivalent to the tubing length.
(iv) Moving horizontally and then vertically, identify the pressure on the same gas/liquid ratio line as was used in (ii) corresponding to this new depth (d
1 + H). This
is the required tubing intake pressure (Pwf1
).
The process may now be repeated at other oil rates (qo2
, qo3
and qo4
). Each rate requiresuse of a different gradient curve appropriate to these higher rates (tubing size, waterfraction and gas/liquid ratio are constants!). The tubing intake pressures (P
wf2, P
wf3,
and Pwf4
) may now be plotted as a function of oil rate(qo1
, qo2
, qo3
and qo4
).
This is the OUTFLOW curve specific to the set of conditions that were used togenerate it (wellhead pressure, tubing size, liquid rate). It will be combined in chapter1.12 with the Inflow Performance Relationship to estimate the well production rate.
Figure 25
Contruction of the tubing
performance relation (TPR)
using gradient curves
1
32
However, first we need to amplify our earlier discussion or flow maps (chapter 1.7)then look at how they can be implemented in well performance computer programs(chapter 1.11).
1.7 Flow Maps and CorrelationsMany authors have studied the phenomena of 2- and 3-phase flow in vertical, inclinedand horizontal pipes. They have proposed a large number of flow maps andcorrelations based on the large (field and laboratory generated) databases generatedduring their studies e.g. the Duns and Ros study contained 4000 separate laboratorydata sets. In this case the database analysis procedure consisted of:
• Expressing the data (e.g. gas and liquid velocities, pipe diameter and liquidviscosity) in a dimensionless form.
• Using these dimensionless numbers to draw a flow map indicating the boundariesbetween the various flow regimes (some studies concentrated on only one flowregime) equations were developed to describe these boundary lines.
• Correlations for each flow regime allow the calculation of slippage, hold up andfriction factor and hence pressure drop in the pipe.
10-1
10-1
1
1
10
10
102 103
Region 1
Region 2
Region 3
BubbleFlow
PlugFlow
SlugFlow
FrothFlow
MistFlow
Dimensionless Gas Velocity Number NGV
Dim
ensi
onle
ss L
iqui
d V
eloc
ity N
umbe
r N
LV
Alternatively, the models can be based on a mechanistic description of the underlyingfluid mechanics. Typical examples of two types of flow maps - one for vertical flowbased on experiments (Duns and Ros, “Vertical Flow of Gas and Liquid Mixtures inWells” Proc Sixth World Petroleum Congress, Vol 2, paper 22, 1963) and the otherfrom a theoretical analysis of horizontal flow regimes (Taitel and Dukler, 1976) arepresented as figs 26 and 27 Many of the methods were subsequently modified toextend their range of application - an example being Figure 27b (Taitel, Barnea andDukler, “Modelling Flow Pattern Transitions for Steady Upward Gas-Liquid Flow inVertical Tubes”, AIChE, J, 26, 345-354, May 1980) which extends the horizontal flowwork summarised in Figure 27a.
Figure 26
Duns and Ros flow pattern
map
Institute of Petroleum Engineering, Heriot-Watt University 33
1Well Performance1
We discussed earlier that many different studies have been published and that noneof them are universally applicable - some being more limited than others. Applicationto oil and gas production requires that they should include:
0.1
0.01
0.10
1.00
10.0
75.0
1 10.0 100.0 900.0
Bubbly
Intermittent
Annular
Stratified WavyStratified Smooth
Superficial Gas Velocity (ft/sec)
Sup
erfic
ial L
iqui
d V
eloc
ity (
ft/se
c)
0.010.001
0.03
0.3
3
30
0.3 3.0 30 300
Superficial Gas Velocity (ft/sec)
Sup
erfic
ial L
iqui
d V
eloc
ity (
ft/se
c)
Bubbly
Slug or Churn
Barnea Transition
Dispersed Bubble
Annular
Figure 27 (a)
Taitel-Dukler horizontal
flow map
Figure 27 (b)
Taitel-Barnea-Dukler flow
map
1
34
(i) Phase Slippage and
(ii) Flow Regime under
(iii) Vertical/horizontal and inclined flow condition.
A few of the studies that will be encountered by the practising Production Engineerare listed in table 4:
Reference Data Source Fluids Comments
Gilbert Field data G, O, W Introduced vertical, multiphase gradient curves
Duns and Ros Field and Lab. data G, O, W Vertical flow over wide flow (air, oil & water flow rate range in 11/4 - 31/8 in. pipes)
Griffith and Wallis1 Laboratory data G, W Good slug flow correlation (air & water flow in used by later investigators
narrow pipes)
Hagedoorn and Field experiment G, O, W Forms basis for widely Brown2 (gas, oil & water used correlation
flow in 1 - 4in. pipes) Aziz and Govier3 Field & Lab. data G, W Correlations developed by
(air, oil & water flow mechanistic fluid mechanical in a wide range of pipes) study tested against field data Beggs and Brill4 Laboratory data G, W Correlations useable at all
(air & water flow in inclination angles 1-11/2 in. pipes)
1. Griffith, P. and Wallis, G.B., “Two-Phase Slug Flow,” J. Heat Transfer, Trans.ASME, Ser. C, 83, 307-320, August 1961.
2. Hagedorn, A.R. and Brown, K.E., “Experimental Study of Pressure GradientsOccurring During Continuous Two-Phase Flow in Small-Diameter Vertical Conduits,”JPT, 475-484, April, 1965.
3. Govier, G.W. and Aziz, K., The Flow of Complex Mixtures in Pipes, Robert E.Drieger Publishing Co., Huntington, NY, 1977.
4. Beggs, H.D. and Brill, J.P., “A Study of Two-Phase Flow in Inclined Pipes,” JPT,607-617, May 1973.
Complex calculation procedures are required to calculate the tubing performancerelationships based on the multiphase flow correlation methods developed by thevarious investigators. This will not be discussed in detail here, but can be found in the
Table 4
Flow Correlations
Institute of Petroleum Engineering, Heriot-Watt University 35
1Well Performance1
original papers and text books. The practising petroleum engineer will most oftenobtain access to these techniques in two manners:
(i) Use of gradient curves e.g. Figure 24 was calculated using the Hagedoorn andBrown correlation
(ii) Direct application of the correlations calculation procedure in one of the many(commercial) computerised well performance prediction packages available on theopen market (chapter 1.11).
Hence it is sufficient for our purposes to mention just a few of the key points for fourof the main correlations.
1.7.1 Duns and RosDuns and Ros defined a flow map (Figure 26) together with a series of correlations forcalculating the boundaries between the flow regimes as well as the slip velocity (V
s).
The Friction factor is calculated from the liquid Reynolds Number when flow is in theBubble or Slug regions; while the gas Reynolds number is used in the Mist region.Finally, calculation of the pressure drop is completed by adding an acceleration termfor flow in the Mist region only.
Many well flow simulation computer programs include modifications of the originalDuns and Ros correlation. These include some or all of:
(i) use of a different flow map (by Gould et al)
(ii) addition of the Beggs and Brill correction to modify the hold up correlation toallow for well deviation
(iii) use a modified friction factor (Kleyweg et al, “Gas Lift Optimisation in theClaymore Field,” Offshore Europe Conference, 1983).
1.7.2 Hagedorn and BrownHagedorn and Brown developed a simple flow map and a liquid hold up correction forthe slug flow regime; the Griffith correlation being used for the bubble flow regime.The friction factor is calculated using a two-phase Reynolds number.
This original work has been modified to include points (ii) and (iii) discussed above- viz. the “Beggs and Brill” correction to modify the hold up for all angles of deviationand the Kleyweg friction factor for single phase flow.
The modified Hagedorn and Brown correlation is probably the most widely usedcorrelation for well performance calculations.
1.7.3 Beggs and BrillThe “Beggs and Brill” method is based on a study of the flow regimes that occur inhorizontal pipes. The flow regime and hold up are calculated as though the pipe washorizontal and a correction made to account for the change in hold up due to the angleof inclination when the pipe is not perfectly horizontal. (NB. The flow regime
1
36
calculated is the one that would have occurred if the pipe was horizontal). TheKleyweg single phase friction factor approach can also be used.
Given the above, it is probably not surprising that this is often the preferred correlationfor simulating the flow in the horizontal and highly deviated portions of wells as wellas in flowlines and pipelines. The correlation is particularly suitable for simulatingpipelines in hilly terrain since it can cope with both upward and downward flow.
1.7.4 GrayThe Gray correlation was specifically developed for gas wells producing smallamounts of liquid - either water or condensate. Experience has shown that this isnormally the best correlation for these conditions.
1.8 Temperature Modelling
PAvg,TAvg
Heat Transfer Through Walls
AnglePIn,TInFluid and Heat In
Fluid and Heat Out
POut,TOut
Dep
th
Temperature
Geotherm
al Gradient
Surface
Fluid Temperature Profile (q1)
Fluid Temperature Profile (q2)
q2>q1
The law of conservation of energy dictates that all enthalpy changes e.g. phasechanges driven by pressure changes, work done overcoming frictional forces etc. arereflected by a corresponding temperature change. Further, large scale heat loss fromthe (hot) produced fluid produced from the (hot) reservoir will occur as it flowsupwards to the (cool) surface. Figure 28 and 29 illustrate the calculation procedure and
Figure 28
Temperature modelling /
Calculation proceedure
Figure 29
The average tubing
temperature increases as
the production rate
increases
Institute of Petroleum Engineering, Heriot-Watt University 37
1Well Performance1
compares the formation (geothermal) temperature with the fluid temperature duringproduction. In general, the higher the production rate, the hotter the fluid will be at anygiven depth (since the increase in the (rate of) supply of energy (heat) is proportionalto the production increase while the heat losses from the wellbore by thermalconductivity etc. are a only function of the temperature difference between the welland the surroundings i.e. independent of the production rate.
This temperature change will effect the average fluid properties - which in turn willalter the pressure drop calculation (and hence the temperature change). A fullsimulation of flow in a well thus requires a coupling of the fluid temperatureprediction model with the pressure calculation. This temperature model may rangefrom a simple analytical equation to a rigorous numerical description of heat flows.The coupling of temperature and pressure requires an iterative procedure for theircalculation. Figure 30 charts the pressure (inner) and temperature (outer) loops.
START
Given P1,T1,H1,DL,qEstimate ∆T & ∆P
Calculate Fluid Props.Calculate ∆PEST
| ∆PEST _
∆P |<ε∆P
| ∆HEST _
∆H |< ε∆H
T=T1+∆T/2Calculate ∆HEST
P=P1+∆P/2
P2 = P1+∆PESTT2 = T1+∆T
∆Tnew = ∆Told.∆HEST/∆Hold∆P = ∆PEST
∆P = ∆PEST
STOP
Calculate H2∆H = H2
_H1
YESYES
NO
NO
Outer Loop
Inner Loop
ε∆P and ε∆H are the allowable differences in the calculations between successive iterations
1.9 SURFACE PRESSURE LOSSES
1.9.1 Surface ComponentsThe principal surface system pressure loss is often the surface choke. This is an“optional” pressure loss in the sense that it is designed into the well completion inorder to control the well flow rate and the pressures to which the surface equipmentis exposed. The choke can be eliminated completely when the wellhead pressure hasbeen depleted to such an extent that economic flow rates can only be achieved bylowering the wellhead pressure to its minimum value.
A second source of pressure losses in the surface system is the flow line. It should beremembered that flow line pressure losses are not only related to the length, diameterand wall roughness of this pipe; but that additional pressure losses will occur in pipefittings (T-pieces, elbows, etc) and valves. These additional pressure losses are
Figure 30
Pressure and temperature
calculation
1
38
accounted for as an increase in the effective length of the pipeline. These increasescan be quite substantial e.g. the pressure losses in some types of valve - especiallywhen only partly open - can be up to several hundred times the pipe inner diameter.
Excessive flow line pressure losses can be reduced by installing a parallel, or looped,flow line in order to reduce the fluid’s flowing velocity. (This is an option not availablewithin the well!) Modeling of looped pipelines is relatively simple for single-phaseflow; since the flow will divide itself between the two branches so that there is an equalpressure drop along the two pipelines. The looped pipeline constructed from individualflow lines of diameters d
1 and d
2 will behave in a similar manner to a single pipe with
an effective diameter of (d12.5 + d
22.5)0.4 (this assumes that the same fluid properties and
friction factors apply to both branches).
The above concept is not appropriate for multiphase flow since the liquid and gasphases are unlikely to split equally between the two branches - often most of the liquidwill go into one branch while most of the gas will be diverted to the second one. Themass split ratio between the two lines is difficult to predict - it will depend on the exactarrangement of the T piece and the Reynolds number associated with the flow in eachline. Thus a lower flow velocity and a T junction design where the loop flow lines arenot at the same elevation, will result in most of the denser, liquid flow going into thedownward pointing line.
1.9.2 Flow Through ChokesFlow from a well often has to be controlled for reasons such as:
(i) limitation of the drawdown to prevent water coning, gas cusping or sandproduction;
(ii) dissipation of well energy to meet pressure limitations of the downstream surfaceproduction equipment etc;
(iii) control of well production rates to meet regulatory, reservoir management orproduction equipment constraints.
Chokes differ from other completion equipment in that they are designed to producea pressure loss while other components, such as subsurface safety valves, are designedso that their presence has a minimal effect on the total system pressure losses.Commonly employed chokes (Figure 31) disturb the fluid flow pattern by use of afixed bean, an adjustable rod (which (partly) blocks an orifice) and a rotating disc.Chokes achieve the desired pressure loss by restricting the flow diameter andacceleration of the flowing fluid. The phenomenon of critical flow occurs once thisacceleration in the throat of the choke is sufficient that the flowing fluid’s sonicvelocity is exceeded.
Institute of Petroleum Engineering, Heriot-Watt University 39
1Well Performance1
Direction of Flow
Direction of Flow
Replacable FixedRestriction or Bean
Fixed Choke
Replaceable Orifice
AdjustableRod
Fully Open
Direction of Flow Direction of Flow
Adjustable Choke
Rotating Discs
Fully Opened Throttling Fully Closed
Rotating Disc Choke
Replaceable Orifice
AdjustableRod
Choked Flow
Critical flow prevents a pressure disturbance downstream of the choke from beingpropagated upstream, since a pressure wave can not travel faster than the speed ofsound. The well’s performance (upstream of the choke) can thus be decoupled fromevents occurring in the downstream flow line and separation system. This has obviousadvantages when trying to control the well’s performance.
Figure 31
Example of surface choke
designs
1
40
Sub - Crit
ical F
low
Crit
ical
Flo
w
Pu
Mass Flow Rate of Gas
Pre
ssur
e ra
tio (
Pu
/ Pd)
2.0
1.0
Eddy Currentslead to irreversible
pressure losses
UpstreamPressure
(Pu)
Gas FlowRate (Q)
DownstreamPressure(Pd)
Vena Contracta Abrubt EnlargementLeads to Low Velocity Flow
and (incomplete) Pressure Recovery
Abrubt RestrictionLeads to
High Velocity Flow and Low Pressure
D2 D1
Figure 32 sketches the fluid flow pattern through the choke and the resulting flowbehaviour. The top part of the figure shows how the choke represents an abruptrestriction in the fluid flow in the pipe. This restriction results in an area of highvelocity and decreased pressure in the centre of the choke. This is known as the “VenaContracta”. As shown, it forces the liquid to flow through an even smaller diameterthan that of inner diameter of the choke. The fluid flow expands again to its originaldiameter at the (abrupt) end of the choke. The decrease in velocity results in recoveryof (some) of the pressure that had been lost during passage through the choke. Fullpressure recovery is not normally experienced since there are irreversible pressurelosses due to eddy currents which create disengagement and reattachment of theflowlines to the pipeline wall.
The bottom section of this figure shows how the flow rate through the choke is relatedto the ratio: {(upstream pressure (P
u) / downstream pressure (P
d)}. For sub critical flow
conditions, the flow rate will increase with decreasing downstream pressure {orincreasing (P
u/P
d) ratio} until this ratio is sufficiently large that critical flow occurs.
The ratio (Pu/P
d) normally has a value of the about 2 at this point - the exact figure will
depend on the properties of the flowing fluid, as discussed below - with critical flowcontinues to occur for all higher (P
u/P
d) ratios.
Figure 32
Critical and sub-critical
flow through a choke
Institute of Petroleum Engineering, Heriot-Watt University 41
1Well Performance1
The mass flow rate under critical flow conditions is independent of the upstreampressure for incompressible (liquid) flow. By contrast, it will depend on the upstreampressure when a compressible fluid (gas) is flowing. Single-phase critical flowtypically occurs when the (P
u/P
d) ratio is greater than 1.5. Critical flow in multi-phase
mixtures requires a somewhat greater pressure ratio; (Pu/P
d) typically having a value
of between 2.0 and 3.0.
1.9.2.1 Single Phase Subcritical Liquid FlowSingle phase, liquid flow is described by:
Q C C DP P
DU D= −
* * 2 ρ
where:P
U = Upstream Pressure
PD = Downstream Pressure
D2 = Choke Diameter
CD = the flow discharge coefficient through the choke,C = a constant depending on the units employed andρ = the density.
The choke manufacturer normally supplies a choke performance chart or correlationthat relates the discharge coefficient (CD) to the diameter of the choke (D2) and theReynolds Number.
1.9.2.2 Single Phase Gas FlowIt is relatively simple to derive equations describing the isentropic flow of an ideal gasthrough a choke. These can be found in the standard text books on the subject. It canbe shown that:
PP
U
D c
= +
+
1
12
γγ
γ
where:
{Pu/P
d}
c = the ratio of the up and downstream pressures at which critical flow occurs,
andγ = the ratio of the gas heat capacities at constant pressure and constant volume{or (Cp/Cv)}.
γ has a value of approximately 1.4 for diatomic gasses such as air. Hence the criticalpressure ratio is 1.89, confirming the values quoted above and illustrated in Figure 32.
1.9.2.3 Multiphase Flow Critical Flow RateMultiphase (gas-liquid) flow is not easily described theoretically - empirical correlationshave been developed by a number of investigators which are all of the form:
Pu =
b Q R
DL
c
a
* *
64
1
42
where:Pu = the upstream pressure (psig, except for Ros who uses the unit psia)QL = the liquid critical flow rate (Stb/d)
D64 = the choke diameter (64th of an inch)R = the gas / liquid ratio (scf/STB) and
a,b, & c = are constants given in Table 5.
Correlation a b cRos 2.00 17.40 0.500Gilbert 1.89 10.00 0.546Achong 1.88 3.82 0.650Ausseens 1.97 3.86 0.680Baxendell 1.93 9.56 0.546
The effect of changing the choke size on well production and flow system pressurelosses can be studied using the Nodal analysis technique (see section 1.12). Theupstream side of the choke is normally chosen as the Node. The WellheadPerformance (or combination of the well’s Inflow Performance and TubingPerformance Relationships) is the upstream, inflow component to the node and thechoke, flowlines and separator are the downsteam, outflow component. A typicalresult from such a calculation is shown in Figure 33. This shows an operating pointof 460 b/d for a 16/64 in choke increasing to 1370 b/d for a 40/64 in. choke.
Well headperformance
1200
800
400
00 750 1500
Well production (bfd)
Nod
e pr
essu
re (
psi)
sub critical flow
Choke size16/64 in.
24/64 in.
32/64 in.
40/64 in.
1.9.3 Gathering System LayoutThe layout of the surface facilities and flow lines is important for land fields where,typically (near) vertical wells are drilled at a relatively small interwell spacing. Thewellhead locations thus reflect the subsurface locations - the resulting grid patternbeing illustrated in Figure 34 and 35. Connection of each wellhead directly to a localgathering stations with primary separation facilities (Figure 34) results in each wellbeing connected by a short length of flowline directly to the primary separator. Thisallows lower wellhead pressures than the alternative (Figure 35) where the wells aretied into a common pipeline.
Figure 33
Choke performance curves
Table 5
Flow through chokes -
empirical choke
correlations
Institute of Petroleum Engineering, Heriot-Watt University 43
1Well Performance1
Export pipelineWell head
Gathering station
Main processing facility
Trunk line
Well head to gathering station flow line
Export pipeline
Well head
Main processing facility
Trunk line
Local gathering line totrunk line
g g y y
Flowline pressure drops are thus much larger in the Figure 35 case, unless widediameter pipes are installed e.g. a 50% increase in flow rate can sometimes lead to a300% increase in the frictional pressure loss across a section of pipe. The increase inflow rate in the gathering system as one gets nearer the separator means that wells witha more direct connection to the separator can be produced at a lower wellheadpressure. The performance of each well thus has a much greater impact on itsneighbours compared to the installation of a local gathering station (Figure 34).
1.10 Completions Inflow PerformanceWell performance prediction programs require that the Inflow Performance Relation-ship is specified. This is normally in the form of a Productivity Index value. Forexisting wells this value can be obtained from analysis of a Well test, ProductionLogging Survey etc. A model of the completion, in conjunction with either a “straight
Figure 34
Oil field developed with
local gathering station
Figure 35
Oil field developed with
single processing facility
only
1
44
line PI” or Vogel type inflow relationship, is required in the absence of these fieldmeasurements or when designing a new well.
1.10.1 Perforated Completions
Casing
FormationDamage Zone
PerforationDiameter
Cement
Penetration Length
Crushed Zone wherepermeability of crushedsand grains has lowerpermeability than theundamaged formation
Undamaged formation
The Open Perforation model (Figure 36) can be used to predict well performance. Theinflow performance is affected by the:
(i) Perforation length (L) - longer perforations are more productive
(ii) Perforation diameter (Dperf) - wider perforations will show a reduced frictionalpressure loss
(iii) Perforation density (n) - reducing the distance between perforations willincrease the well productivity
(iv) Perforation phasing - reducing the angle between adjacent perforations willincrease the well productivity
(v) Depth and Permeability reduction caused by Formation Damage - formationDamage has limited effect on well productivity provided it is penetrated by theperforation.
(vi) Permeability and depth of crushed zone around the perforation - perforationclean up procedures should be designed to remove this impaired crushed zone priorto production.
(vii) Formation vertical and horizontal permeability - reduced vertical permeabilityimpedes well production when the perforations are far apart (low shot densities).
Figure 36
Flow through completions -
perforated completion
model
Institute of Petroleum Engineering, Heriot-Watt University 45
1Well Performance1
(viii) Drawdown and properties of the produced fluids - high gas and very high oilflow rates through the perforation lead to extra pressure losses from non-Darcy floweffects.
The relationship between some of the factors discussed above is illustrated in Figures37 - 40 in which the productivity of an example completion is compared with that ofthe equivalent open hole completion.
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
Perforation Penetration Length (in.)
Pro
duct
ivity
Rat
io
0 5 10 15
•Crushed and Formation Damage zones omitted and•Turbulence factor not included
Open Hole
Perforation Density (shots/ft)168
4
90° Phasing
1684
0° Phasing
1684
180° Phasing
Figure 37 compares the effect of perforation density and phasing. For this particularexample, avoiding tortuous flow paths associated with 0º phasing perforations willforce the fluids to flow “round the casing” (Figure 38), has a greater influence on thewell productivity than the perforation density.
Figure 37
Influence of perforation
density and phasing
1
46
y
Zero phased(in - line)perforation
Tortuous fluid flowpath to perforation
Cement
Casing
Penetration Length (in.)
Pro
duct
ivity
Rat
io
0 5 10 15
4 and12 Shots/ft. at 120° phasing
1.0
0.8
0.6
0.4
1
1
KvKh
Open Hole
0.1
0.01
0.10.01
KvKh
Figure 39 illustrates how the effect of an unfavourable vertical permeability can beovercome by placing the perforations closer together. Also, it can be seen how a wellcompleted in a formation with a vertical permeability similar to the horizontalpermeability can have a productivity approaching that of an open hole completion,even when there is a low perforation density.
Figure 38
0° phased perforation
reduce well productivity
Figure 39
Influence of vertical and
horizontal formation
permeability and
perforation density on well
production
Institute of Petroleum Engineering, Heriot-Watt University 47
1Well Performance1
Crushed ZoneRemoved or
Darcy Flow (low rate gas flow)with Non - Darcy Effect (high rate flow)4 Shots/Foot at 0° Phasing
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
00 5 10 15
Pro
duct
ivity
Rat
io
Penetration Length (in.)
Open Hole
K formation = 1K crushed zone
K formation = 10K crushed zone
K formation = 4K crushed zone
Figure 40 illustrates how the cylindrical crushed zone reduces the well performance.Further, high velocity (non-Darcy or turbulent) flow effects often associated withgas wells will be accentuated, further reducing the inflow performance. Thisexplains why so much care needs to be taken when designing completion procedures,
Modern perforating technology has increased the options available to the completionengineer e.g.
(i) perforation charges with a 1.37 m penetration depth for the standard API target,
(ii) ability to perforate > 2600 m of casing in one run by simultaneous detonation of25,000 perforation charges and
(iii) creation of a large inflow area equivalent to 400 cm2/m length for a 9.875 incasing.
1.10.1.1 Perforation Charge PerformanceMore quantitative calculations on the parameters that affect the performance ofperforated completions can be made using the correlations supplied in paper SPE18247, “Semi-analytical Production Models for Perforated Completions” by M.Karaka and S. Tariq. Experiments under realistic downhole conditions have shownthat the (downhole) performance of (shaped) perforating charges depends on the:
(i) weight of explosive,
(ii) whether it is designed to produce a wide diameter or deeply penetrating hole,
(iii) type of charge liner,
(iv) perforation gun design and stand-off from the casing wall,
Figure 40
Crushed zone around
perforation and high rate
gas flow reduce well
productivity
1
48
(v) thickness and type of the casing, rock strength and insitu stresses, formationpressure etc.
Prediction of the Well’s Inflow Performance requires that this downhole perforatorperformance must be estimated through knowledge of the perforation length, widthand properties of the crushed zone (if left in place). Measurements for the abovefactors, preferrably made under simulated downhole conditions, are normally sup-plied by the service company providing the perforating system. These will either bebased on specific physical experiments carried out for the specified well situation orby use of a computer program containing correlations developed from extensiveexperimentation. For example, Schlumberger’s SPAN™ (Schlumberger PerforatingAnalysis) software will predict the:
(i) downhole gun performance,
(ii) optimal under balance required to remove the perforating debris, crushed zone,etc. and
(iii) well productivity.
The impact on well inflow efficiency for different gun types, perforation chargedesigns, perforation phasing can be estimated, compared and an optimum selectionmade.
The standard against which the productivity of the different perforating systems arecompared needs to be evaluated carefully. One option, preferred by the author andused here, is the productivity equivalent to that of the open hole originally drilled inthe absence of formation damage.
1.10.1.2 Perforation Gun SelectionSo far we have discussed perforating system choice in terms of well productivity only.There are many other practical considerations to be borne in mind when selecting aperforating system. These include:
(i) compatibility of the physical dimensions of the perforating gun and the comple-tion.
(ii) the completion technique e.g. perforating prior to or after running tubing.Selection of “Through Tubing”, Casing or Tubing Conveyed Perforating guns.
(iii) casing damage/perforating debris.
(iv) management of sand production. The phasing, orientation and design of theperforating pattern can impact on the severity of sand related production problemswhen the well is placed on production.
(a) A change in the perforation phasing from 60° to 90° in BP’s Magnus Field, whilemaintaining the same perforation density and charge type, gave a substantialreduction in production problems attributed to sand production. This wasascribed to a new gun design that gave a 56% increase in the minimum spacing
Institute of Petroleum Engineering, Heriot-Watt University 49
1Well Performance1
between perforations. This improvement is based on the concept that sandproduction problems are accentuated by the failure of individual perforations tothe extent that they grow together and form a (relatively) large cavity. Maximisingthis separation minimises the chance of perforation collapse and amalgamationwithout compromising the well’s inflow performance.
(b) Sand “run in” into the casing between the perforating operation and placing ofthe gravel pack screen can be an operational problem when installing a gravelpack completion in deviated wells in some unconsolidated formations. Fieldexperience has shown that this “run in” can be minimised by omitting theupward facing perforations from the perforation gun (this requires that theperforation gun has to be orientated before firing). This improvement wasattributed to the ease with which the unconsolidated sand fell vertically down-wards (under the influence of gravity) during (weak) pressure surgescaused by running the completion equipment.
(c) Experience has shown that long term sand production can be minimised by orientating the perforations in the direction of the maximum insitu rock stress.
1.10.2 Gravel Packed CompletionsPerforating strategy in natural flow completions i.e. those which require neither sandcontrol nor hydraulic fracture stimulation, is aimed at delivering sufficient perforationsopen to flow so that the overall well productivity is not reduced by the presence of theperforations. In the previous section we discussed underbalance perforating techniquesused to reduce formation damage by cleaning out the perforation change debris andformation crushed zone. This formation / perforation damage removal process maycontinue when the well is placed on production since there are open perforationsthrough which the debris can flow into the well. However, the installation of a gravelpack traps any remaining debris in the perforation tunnel behind a sheath of gravel.
Gravel packed well completion strategy has the same objective to that for natural flowcompletions - ensuring that the perforations do not limit well production in any way.The restriction on the ability of the flow from the well to remove damage in the longerterm implies that the perforating process has to be designed to minimise the damagecreation. In addition, there is the new factor of minimising formation damage fromthe gravel packing operation itself (Figure 41a).N.B. Production zones to be gravel packed often have a high permeability and areservoir pressure depleted below the hydrostatic value. They are particularly proneto formation damage due to fluid loss and/or exposure to Lost Circulation Material.
1
50
Perforated tubing
Wire wrappedsand screen
Gravel Wel
l cen
ter
line
Casing
Radius Bit
Formation damage from drilling
Perforation Diameter
Tunnellength
Cement
Formation damagefrom the perforatingprocess, perforationgun debris andgravel packing
Formation
Flow
Undamaged formation
Weak sands are often unable to support a defined perforation tunnel. The resultinggravel packed completion is then best represented by Figure 41b. In practice, it is oftenunclear which of the above models is most suited to a particular situation. Further, thedepth and extent of permeability damage due to fluid loss etc. is normally not known.
Perforated tubing
Wire wrappedsand screen
Gravel Wel
l cen
ter
line
Casing
Radius Bit
Formation damage from drilling
Perforation Diameter
Tunnellength
Cement
Flow
Formation damagefrom gravel packing
UndamagedFormation
One practical solution to the above is to use a composite model of the inflow processsimilar to that sketched in Figure 42. This combines a gravel filled tunnel of lengthequivalent to the distance from the wire wrapped screen to the edge of the cementsheath together with the porous media, radial inflow equation.
Figure 41b
Flow through gravel pack
completion in a weak sand
Figure 41 a
Flow through a gravel pack
completion in a semi
competent sand
Institute of Petroleum Engineering, Heriot-Watt University 51
1Well Performance1
Perforated tubing
Wire wrappedsand screen
Wel
l cen
ter
line
Radius Bit
Cement
Skin? Skin?
Tunnellength
Flow
Gravel
Casing
Perforation Diameter
Porous mediaradial inflowequation
UndamagedFormation
Impairment to the gravel pack sand can then be represented:
(i) by reducing the number of perforations that are open to flow,
(ii) as a “skin” value at the wire wrapped screen or formation interface or
(iii) by the gravel permeability being reduced below its unimpaired value e.g. 20/40US Mesh gravel typically has an unimpaired permeability of about 180D.
Formation damage can be simulated by use of a Vogel “Flow Efficiency” factor lessthan 100% or by inclusion of a “skin” between the gravel and the formation.N.B. This skin would have a negative value if it was believed that unimpaired gravelhad been displaced past the cement sheath and that permeability damage to theformation was absent!
This model can be used to design the gravel pack completion by ensuring that thepressure drop across the gravel filled tunnel has a minimal impact on the wellproduction.
(i) As large as possible gravel pack sand size is selected that is capable of controllingthe formation sand and a choice made as how to represent gravel pack sandimpairment (discussed above).
(ii) The perforation density and diameter are varied until an acceptable pressure dropis predicted. The relationship between these two parameters is discussed in chapter1.12.5 on completion design.N.B. Remember that the perforations are now filled with gravel. Also, that gravelfilled perforations will exhibit extra turbulent flow induced pressure losses (non-Darcy flow effects) at much lower flow velocities than open perforations.
1.10.2.1 Non-Darcy Turbulence Pressure LossesThe Darcy contribution to the pressure drop through the perforations can be representedby a gravel-pack skin factor, s
g, and the non-Darcy flow coefficient for the gravel-
filled perforation , Dg. The latter term is normally calculated using the Forcheimer
Figure 42
Flow through completions -
a composite gravel pack
completion models
1
52
equation. Different equations are derived for gas (Dgg
) and oil (Dgo
) wells. Golan andWhitson (see reference 4, section 1.13) quote the following equations for thecalculation of pressure losses across the gravel-filled perforations of an inside-casinggravel packs:
Skh L
k d nandg
g perf= 96
2
* ** *
For gas wells: Dkh L
d nggg g
perf=
−2 45 10 10
4 2
. * * * * *
* *
γ βµ
For oil wells: DB kh L
d ngoo g
perf=
−1 8 10 11
4 2
. * * * * *
* *
βµ
In these equations:kh = the formation permeability-thickness product (md:ft),L = the gravel-packed tunnel length (in),k
g = the permeability of the gravel (md),
dperf = the perforation diameter (in),
γ = the gas relative gravity,µ = viscosity (cp),n = the number of perforations andβ
g = the gravel turbulence factor
Βο = the oil volume factor (bbl/stb).
These terms must be added to other sources of “skin” in the conventional radial inflowequation e.g. for outflow:
Productivity Index = Q
P P
khB In r r S S D Q
reservoir sandface
o e w g g
−
=+ + +141 2
. * * *{ ( / ) * }µ
Where S represents all skin factors apart from that due to the gravel pack and Q is theflow rate. The term D
g*Q is often referred to as the rate dependent skin.
The turbulence factor (βg) is a rock property of the gravel pack sand - its numerical
value is related to the permeability and sand grain size. Formation damage, whichreduces the (Darcy) permeability of the gravel pack sand, will also increase the valueof β
g; leading to even greater pressure losses. The turbulence factor is correlated with
the gravel permeability (Cooke, “Conductivity of Fracture Proppants in Multiplelayers”, J.P.T., 1101-1104, September 1973) as:
g =β bkbkga−
Values for the constants a and b for common gravel sizes are suggested in Table 6.
Institute of Petroleum Engineering, Heriot-Watt University 53
1Well Performance1
U.S. Mesh Average Permeabiliy Turbulence factor Size diameter (md) βg = bkg
(in) a b
40/60 0.014 1.2 x 105 1.6 2.12 x 1012
20/40 0.025 1.8 x 105 1.54 3.37 x 1012
10/20 0.056 5.6 x 105 1.34 8.4 x 1011
-a
1.10.2.2 Restriction of Gravel Pack DrawdownSome operators restrict the pressure drop allowed across the gravel pack completion.This arose because they observed an increased sand control failure rate whenproducing at a high well drawdown - a figure of 300 psi has been used in the USA GulfCoast fields (though other operators have had good experience at much higherdrawdowns). The impact on the production rate of varying this allowable drawdowncan be evaluated by carrying out a nodal analysis calculation which places the nodeat the sand face and omits the presence of the gravel pack i.e.
InflowAverage reservoir pressure - drawdown across formation - sandface pressure
OutflowSandface pressure = pressure drop across (tubing + flowline) + separator pressure
Average reservoir pressure
Production Rate
San
d F
ace
Pre
ssur
e
Inflow
Out flow
∆Pgp=0∆Pgp4
∆Pgp3∆Pgp2
∆Pgp1
Q1 Q2 Q4Q3 Q max
The intersection point (Figure 43) represents the maximum flow rate (Qmax
) when thepressure drop is often across the gravel pack (∆Pgp) is zero. In practice, the pressuredrop across the gravel pack is often greater than zero - it can be estimated from thedifference between the inflow and outflow curves at the production rate actuallyachieved by the well. This value can be compared with that predicted for theunimpaired completion by calculating the pressure losses through the gravel filledtunnels, as described above.
1.11 Computerised Well Performance Prediction ProgramsThe basic concepts required for the calculation of the pressure difference between twopoints that are connected by a pipe have been discussed in the preceding chapters.Section 1.6 introduced the graphical pressure traverse method for estimating pressuredrops; while the subsequent chapters detailed the requirements for a numericaldescription of these calculations. The complicated nature of these calculations
Figure 43
Gravel pack pressure
analysis
Table 6
Gravel pack sand properties
1
54
resulting from a full description of the flow processes require a computer to carry outthe calculations in a reasonable period of time.
The calculation procedure employed consists of splitting the pipe into a number ofsegments and calculating the pressure drop across each segment. Figure 44 schematicallyillustrates the procedure. Steady state flow requires that mass, momentum and energybe conserved between the inlet and outlet of a given volume, or segment (figure 44a).
Steady-State Flow(1) [ Mass flowIn = Mass flowOut ],
(2) [ MomentumIn - MomentumOut ] = Sum of momentum changes ] and
(3) [ EnergyIn + Work + Heat = EnergyOut ]
dL
dX
dZ
θV
In
Out
.
This conservation of mass, momentum and energy is applied to the pressure dropcalculation procedure across a segment of pipe (figure 44b)
VL
ρL
P1 > P2
P1 > P2 = ∆P1-2 = ∫
VG
ρG L
L2
L1
dP(P,T) dLdL
1 2
And is then extended to a series of pipe segments (figure 44c).
1 2 3
i = 1 i = n
∆P1-2 ≈ Σ ∆LdP dL
i = 1 i i
n
Suitable arrangements must be made to ensure that the TOTAL mass, momentum andenergy are conserved when two separate flows join together
Figure 44c
Pressure drop calculation
across a series of segments
Figure 44b
Calculation procedure
Figure 44a
Conservation of mass,
momentum and energy
Institute of Petroleum Engineering, Heriot-Watt University 55
1Well Performance1
Larg
e C
hang
e in
(Normally) Small Changes in
Temperature and Pressure
Tem
pera
ture
and
Pre
ssur
e
Segments
q1
q1-q7 Inflow at Particular Locations
q2 q3q4 q6
q7
Large Diameter Casing
Small Diameter Tubing
Surface
Wellhead
RestrictionSub-surfaceSafety Valve
Hydrocarbon Reserve
Figure 45 illustrates the segmentation process extended to simulate a complete well.It is a “horizontal” well producing with hydrocarbon inflow at 7 specific locations.The pressure (and temperature) changes along the “horizontal” section are muchsmaller than occur in the vertical part of the wellbore. The size of the segment isrelated to the magnitude of the pressure or temperature change that occurs across thesegment:
• Larger segments decrease calculation time;• Small segments maintain accuracy when pressure and temperature are varying rapidly.• Fluid properties are calculated at the average of the inlet and out let conditions (temperature and pressure).
Essentially the same process is followed when the calculations are being performedby hand or by by use of a computerised simulation program (the size of the segmentsis much smaller in the later case!). There are several commercial software packageswhich can be used to estimate well performance, e.g Wellflo TM, Prosper TM, PipesimTM etc.
1.12 Well Performance Sensitivity Study ExerciseEarly chapters in this module studied the Well Inflow (1.4 and 1.5) and Well Outflow.Combinations of these two parameters along with the concept of systems analysis ofproduction systems (nodal analysis) (chapter 1.2) allow us to estimate the wellproductivity under today’s actual or future expected producing conditions. Thesensitivity of the well design (or its robustness) to the many factors which effect wellproduction as the well ages can then be tested.
Figure 45
Pressure and temperature
calculation
1
56
Outflow From Node
InflowTo Node
System OperatingPressure
Pre
ssur
e at
Nod
e
System Capacity
Flow Rate
Figure 46 illustrates the systems analysis concept. The point (or node) at which theanalysis is carried out can be chosen to be anywhere in the producing system - theinflow and outflow being calculated for the complete system up- and down-streamfrom the chosen node. The sensitivity of the production rate to changes in thedimensions of a particular component (situated next to the node) can then beevaluated. This allows the performance of each individual well component to beisolated in turn. Typical examples of node selection are:
• Wellhead: evaluate the effect of flow line size
• Safety Valve: evaluate the effect of the reduced flow caused by the diameter of thesafety valve being smaller than that of the tubing (important in high rate gas wells)
• Sandface: select the optimum tubing size or evaluate well inflow performance (isthere a requirement for reperforation, stimulation {to remove a positive skin(acidisation) or create negative skin (hydraulic fracturing)}?
Other possible node points can be seen in Figure 1 which analyses pressure losses inthe complete production system. Some of the more frequently encountered sensitivityanalyses are described below.
1.12.1 Reservoir Inflow and Tubing Outflow RestrictionsThe impact of (relatively) inadequate reservoir inflow (case 1) with a (larger thannecessary) tubing is illustrated in Figure 47a. The opposite case, productionrestriction by a too small tubing (case 2) is shown in Figure 47b for the same reservoirinflow performance. It comes as no surprise to see that:
q1 >> q
2 and p
reservoir ≈ p
sandface (2) >> p
sandface (1) ≈ p
separator
Figure 46
Systems or nodal analysis
Institute of Petroleum Engineering, Heriot-Watt University 57
1Well Performance1
Pre
ssur
e at
San
dfac
e
Production Rate
Reservoir Inflow
Reservoir Pressure
Separator Pressure
Tubing 1
q1
P1
q2
P2
Pre
ssur
e at
San
dfac
e
Production Rate
Reservoir
Inflow
Reservoir Pressure
Separator Pressure
Tubi
ng2
1.12.2 Tubing Size and Liquid LoadingThe well production will normally increase as the tubing size increases. (The pressuredrop in the tubing decreases so that a greater well drawdown is possible for the samereservoir and separator pressure). However, at a certain point the upward (gas) flowvelocity has decreased so much (due to the tubing diameter increase) that it is no longersufficient to efficiently lift the liquid to the surface i.e. slip phenomena commence andliquid holdup (or liquid loading) begins (figure 48a).
Figure 47a
Reservoir outflow restricts
production
Figure 47b
Small tubing restricts
production
1
58
Pre
ssur
e at
San
dfac
e
Production Rate
Reservoir Inflow
Pro
duct
ion
Rat
e
Reservoir Pressure
Separator Pressure
Maximum Rate
Tubing Diameter
Tubing Outflow
Increasing Tubing Diameter
Unstable Production
Eventually, the increased hydrostatic head (due to the liquid loading) will be greaterthan the reduced friction pressure losses as the tubing diameter increases further. Thisleads to a maximum production rate (figure 48b) at a certain tubing diameter. Unstableflow is encountered with even larger tubing diameters - it is not recommended tooperate in this region since liquid loading will eventually progress to the stage that thewell ceases to flow. The underlying cause for the above is a change in flow regimesas the flow velocity decreases. This allows liquid holdup (slip) to occur, whichbecomes progressively more important as velocities decrease further.
Figure 48a
Liquid loading analysis
Figure 48b
Systems analysis for
increasing tubing diameter
Institute of Petroleum Engineering, Heriot-Watt University 59
1Well Performance1
1.12.3 Effect of Water Cut and Depletion
San
d F
ace
Pre
ssur
e
Production Rate
Reservoir
Inflow
WC
=50%
WC
=100
%
W
C=25%
Tubing Out Flow
Water Cut (WC)=
0%
P1 Reservoir
San
d F
ace
Pre
ssur
e
Production Rate
Reservoir Inflow
(WC)=
0%
q3 q2 q1
TubingInflow
P1 Reservoir Pressure at t1
P2 Reservoir Pressure at t2
P3 Reservoir Pressure at t3
San
d F
ace
Pre
ssur
e
Production Rate
WC=0%
Tubing Outf
low
qt3 qt2 qt1
WC
=50%
WC=2
5%
Reservoir Inflow
Figure 49a
Effect of wate cut on
production
Figure 49b
Effect of depletion on
production rate
Figure 49c
Sensitivity of production
rate to pressure depletion
and water cut development
1
60
An increasing water cut reduces the gas liquid ratio as well as increasing thehydrostatic head between the reservoir and the surface. This is illustrated in Figure49a for a slightly over pressured reservoir. Reservoir simulation can be used to predictthe reservoir pressure depletion with time along with any increase in water cut. Sucha simulation is illustrated in table 7.
Time t1 t2 t3
Reservoir Pressure pres1 pres2 pres3
Water Cut 0% 25% 50%
The effect of this pressure depletion on the production rate is summarised in Figure49b and the two are combined in figure 49c. The production rate at time t
3 is only 25%
of the initial production, while a small further reduction in reservoir pressure orincrease in water cut beyond 50% will cause the well to cease production altogether.
1.12.4 Opportunities for Skin Removal by StimulationWell testing frequently identifies that a positive skin effect is restricting wellproduction. The economic incentive for removing this skin (or even inducing anegative skin) can be evaluated with the help of nodal analysis. Figure 50 shows thecurrent well inflow (skin = +8) together with its partial (skin = +2) and complete (skin= 0) removal. The carrying out of a hydraulic fracture (skin = -3) is also illustrated.
Pre
ssur
e at
San
dfac
e
Production Rate
Reservoir Inflow
Skin=-3Skin=0Skin=+2
Skin=+8
Reservoir Pressure
Separator Pressure
Tubing 2 Outflow
Tubing 1 Outflow
q1+8
q2 ,+8
q2+2q
2 ,-3
q1 ,+2 q
1 , 0
q2 0
q1 ,-3
The flat outflow profile of tubing results in large gains in production that might allowthese treatments to be carried out. Tubing 2 (with a more vertical outflow profile) isalready restricting production with the impaired (skin = +8), while only minor(probably uneconomic) production gains are recorded when the skin is removed - the
Table 7
Reservoir Simulator
Predictions
Figure 50
Opportunities for increased
production by skin removal
Institute of Petroleum Engineering, Heriot-Watt University 61
1Well Performance1
most favourable production (-3) being still less than that achieved with the largertubing and the high (+8) reservoir skin.
Appropriate remedial action can only be taken - and economically justified - when thepressure losses within the complete well system are understood.
1.12.5 Completion Design
N1D1 N2D1
N3D1
N4D1
Pre
ssur
e at
San
dfac
e
Production Rate
Reservoir Pressure
Separator Pressure
Outflow
Increasing Number of Perforations
N1<N2<N3<N4
The high skin discussed in the previous case could have many causes e.g. formationdamage, partial completion etc. One factor under the control of the productionengineer is the number and type of perforations. Figure 51a illustrates an increase inthe numbers of perforations {N
1<N
2<N
3<N
4, all of diameter D
1} while Figure 51b
shows that effect of a restricted number of perforations (N1) can be (partially)
compensated for by an increase in the diameter from D1 to D
3 (i.e. reduction in the
frictional pressure loss in the perforation tunnel itself). Theoretically increasing thenumber of perforations is more beneficial since it improves the inflow from thereservoir as well as decreasing the (average) frictional pressure drop in the perforationtunnel (see Figure 51c for comparison). The cost of the perforation operation willincrease as the number and diameter of the perforations are increased - an economicoptimum will be found when both factors are varied simultaneously.
These theoretical calculations are a useful but not complete guide - it is frequentlyobserved in the field that not all perforations are effective e.g. in gravel packed wellsit is standard practice to assume only 33%–50% of the perforations are effective (i.e.open to flow).
Figure 51a
Effect of number of
perforations on production
rate
1
62
N1D1
N1D2
N1D3
Pre
ssur
e at
San
dfac
e
Production Rate
Reservoir Inflow
Reservoir Pressure
Separator Pressure
Tubing Outflow
Increasing Perforation Diameter
D1<D2<D3
Number of Perforations (N)
Pro
duct
ion
Rat
e
Diameter of Perforations (D)
Perforations all Diameter D1
N1 Perforations
D3
N4N3
N2
N1
D2
D1
1.12.6 Well Head PressureThe separator pressure is often the main component in the surface pressure losses. Itexerts a restrictive “back pressure” on the well production which limits the totalpressure drop available for fluid inflow from the reservoir and onward transportationto the surface. This effect is illustrated in figure 52 - where the wellhead was chosenas the node about which the analysis was carried out. Reducing the separator pressureis often an effective way of increasing the well production.
Figure 51b
Effect of perforation
diameter on production rate
Figure 51c
Optimisation of perforating
schedule
Institute of Petroleum Engineering, Heriot-Watt University 63
1Well Performance1
Pre
ssur
e at
Wel
lhea
d
Production Rate
Reservoir Inflow
Tubing Out Flow
q500 q200 q50
Tubing Out FlowTubing Out Flow
500 psi
200 psi
50 psi
Reservoir PressureSeparator Pressure
This type of “backpressure” on the wells is often encountered in more subtle ways e.g.the gas collecting in the tubing/casing annulus of a well equipped with an artificial liftpump can limit the maximum available drawdown by acting as a back pressure.Venting this casing gas increases the drawdown with a corresponding production rateimprovement.
Figure 52
Effects of separator
pressure on production rate
1
64
1.13 FURTHER READING
(1) Beggs H. D.“Production Optimisation using Nodal Analysis”ISBN 0-930972-14-7published by Oil and Gas Consultants Inc., 1991.
(2) Economides M., Hill A. & Economides C.“Petroleum production Systems”ISBN 0-13-658683-Xpublished by Prentice Hall, 1994.
(3) Economides M. J., Watters L. and Dunn-Norman S.“Petroleum Well Construction”ISBN 0-471-96938-9Published by Wiley, 1998.
(4) Golan M. & Whitson C.“Well Performance” 2nd editionISBN 0-13-946609-6published by the Norwegian University of Science and Technology (NTNU), 1996.
(5) Mian M. A.“Petroleum Engineering Handbook for the Practicing Engineer”, Volume 2ISBN 0-87814-379-3Published by PennWell Books, 1992.
C O N T E N T S
ARTIFICIAL LIFT METHODS:1. INTRODUCTION AND SELECTION CRITERIA2. THE NEED FOR ARTIFICIAL LIFT3. REVIEW OF ARTIFICIAL LIFT TECHNIQUES4. CURRENT STATUS OF THE ROLE OF
ARTIFICIAL LIFT IN FIELD DEVELOPMENT5. SELECTION OF ARTIFICIAL LIFT CRITERIA
5.1. Well and Reservoir Characteristics5.2. Field Location5.3. Operational Problems5.4. Economics5.5. Implementat ion of Ar t i f ic ia l l i f t
Selection Techniques5.6. Long Term Reservoir Performance and
Facility Constants6. ROD PUMPS
6.1. Introduction6.2. The Pumping Unit6.3. The Sucker Rods6.4. The Pump6.5. Rod Pump Operation6.5.1. Pump-Off Control6.5.2. Gas Influx2.6.5.3 Centralisers2.6.5.4 Solids2.6.5.5 Pump Diagnosis2.6.6 Pump Design2.6.6.1 Pump Rate2.6.6.2 Rod Stretch and Over Travel2.6.6.3 Pumping Limit Load Calculations
7. ELECTRIC SUBMERSIBLE PUMPS (ESPs)7.1. Introduction7.2. Well Completion Design with ESP's7.2.1. Applications of ESP's7.2.2. Horizontal Wells7.2.3. “Y” Tool7.3. Basic Pump Selection7.4. Advantages and Disadvantages of
Electric Submersible Pumps7.5. Monitoring the Performance of
Electric Submersible Pumps7.6. New Technology7.6.1. Coiled Tubing Deployed ESP's7.6.2. Auto "Y" Tool7.6.3. Dual Pump Installations7.6.4. Reducing Water Production7.7. Electric Submersible Pump Performance
7.7.1. Simplified Electric Submersible Pump Design8. HYDRAULIC PUMPS
8.1. Advantages of Hydraulic Pumps8.2. Disadvantages of Hydraulic Pumps8.3. New Technology (Weir Pumps)
9. PROGRESSING CAVITY PUMPS9.1. Progressive Cavity Moyon Pump Principle9.2. Progressing Cavity Pump Power Supply9.3. New Technology9.3.1. The Progressing Cavity Electr ic
Submersible Pump (PCESP)9.3.2. Wireline retrievable PCESP
10. HYBRID SYSTEMS11. ARTIFICIAL LIFT TUTORIAL12. FURTHER READING
2Selection of Artificial Lift Types2
2
LEARNING OBJECTIVES:
Having worked through this chapter the student will be able to:
INTRODUCTION• Explain the importance of Artificial Lift (AL) for world oil production.• List the different types of AL.
SELECTION• Select appropriate type AL based on ranking criteria.
ELECTRIC SUBMERSIBLE PUMP• Identify the components of an Electric Submersible Pump.• Describe the preferred applications and the mode of operation of the Electric
Submersible Pump (ESP).• Select well conditions suitable for ESP installation as preferred Artificial Lift option.• Identify the application areas where an ESP is NOT suitable.• Evaluate the advantages of an instrumented ESP completion.
BEAM PUMP• Describe the concept and component parts of a Beam Pump.• Select well conditions suitable for beam pump installation.• Explain the beam pump design methodology• State the background to the use of the Dynanometer card for troubleshooting.
FLUID DRIVEN HYDRAULIC PUMPS• Describe the concept, implications for the well completion and advantages of using
high pressure fluid as a power source.• Explain the mode of operation of the:
(i)Jet pump;(ii)Weir Multiphase pump;(iii)Hydraulic pump.
• Identify their advantages and disadvantages.
PROGRESSIVE CAVITY PUMP• Describe the concept and area of application of the Progressive Cavity Pump (PCP).• Compare rod and electric motor driven PCP’s.• Discuss advantages of wireline retrievable PCESP.
Department of Petroleum Engineering, Heriot-Watt University 3
2Selection of Artificial Lift Types2
ARTIFICIAL LIFT METHODS:
1. INTRODUCTION AND SELECTION CRITERIA
This module will introduce the topic of artificial lift - a production engineering topicof increasing importance in field development. The reasons leading to this increasingimportance in the field development process will be reviewed. The main factorsinfluencing the selection of the most important artificial lift techniques will behighlighted.
A brief description will then be given of all the common artificial lift techniques (rodpumps, electric submersible pumps, progressive cavity pumps and hydraulic pumps)apart from gas lift.
Hydrocarbons will normally flow to the surface under natural flow when the discoverywell is completed in a virgin reservoir. The fluid production resulting from reservoirdevelopment will normally lead to a reduction in the reservoir pressure, increase in thefraction of water being produced together with a corresponding decrease in theproduced gas fraction. All these factors reduce, or may even stop, the flow of fluidsfrom the well. The remedy is to include within the well completion some form ofartificial lift. Artificial lift adds energy to the well fluid which, when added to theavailable energy provided “for free” by the reservoir itself, allows the well to flow ata (hopefully economic) production rate. It has been estimated that in 1994 there wasa world inventory of more than 900,000 producing wells. Only 7% of these flowednaturally while the remaining 93% required some form of artificial lift. The averageproduction per well was less than 70 bpd.
2. THE NEED FOR ARTIFICIAL LIFT
Artificial lift is required when a well will no longer flow or when the production rateis too low to be economic. Figure 1(a) illustrates such a situation - the reservoirpressure is so low that the static fluid level is below the wellhead. Question: Is itpossible for this well to flow naturally under and conditions.
Inflow Performance Relationsh
ip
Static Fliud Pressure Gradient
Dep
thP
rodu
ctio
n R
ate
Pressure
Pr reservoir pressure
Q = 0
Preservoir = Pr
Preservoir
Static Fliud Level
Zero pressure at wellhead
1 (a) The well is unable to start flowing flow naturally due to low reservoir pressure
Figure 1a
The well is unable to
initiatenatural flow.
Figure 1
Artificial lift fundamentals
4
Answer: Yes: If the well productivity Index is sufficiently high and the produced fluidcontains enough gas that the flowing fluid pressure gradient gives a positive wellheadpressure. But, tthe well has to be "kicked off" (started flowing) by swobbing or othertechniques.
Figure 1(b) shows how installation of a pump a small distance below the static fluidlevel allows a limited drawdown (∆p') to be created. The well now starts to flow at rateq. N.B. the static and flowing pressure gradients in figures 1(a) & 1(b) are similar sincefrictional pressure losses in the tubing are small at this low flow rate.
Pr reservoir pressure
Pressure Ps now measured on wellhead pressure gauge
Dep
thP
rodu
ctio
n R
ate
Pressure
Q = q
Fluid Level in Annulus
Flowing Tubing Head Pressure(FTHP)
1 (b) Pump creates a small drawdown and flowrate
Pump
Flow Fluid Pressure Gradients at Rate q
P’flowing
P’
Pr
∆p’
Drawdown
∆ P’pump
qmax
q
P’flowing
or Absolute Open Flow (AOF)
It can be readily seen that the same production rate will occur when the pump isrelocated to the bottom of the tubing, provided the pressure drop across the pump, andhence the drawdown, remains the same. The advantage of placing the pump near theperforations is that the maximum potential production can now be achieved {figure1(c)} by imposing a large drawdown (∆P") on the formation and “pumping the welloff” by producing the well at q
2 is slightly smaller than the AOF.
Figure 1b
Pump creates a small
drawdown and flow rate
Department of Petroleum Engineering, Heriot-Watt University 5
2Selection of Artificial Lift Types2
Dep
thP
rodu
ctio
n R
ate
Pressure
Q = q2
Fliud Level in Annulus
Installation of pump suction belowthe perforations maximises potential drawdown and production rates
Pr reservoir pressure
Flowing Fluid Pressure Gradients at Rate q2
Flowing Tubing Head Pressure Ps
Pr
∆ P"pumpP"flowing
P"flowing
Drawdown ( ∆P" )
q2
P"flowingPump
Artificial lift design requires that the pump to be installed is matched to the well inflowand outflow performance
3. REVIEW OF ARTIFICIAL LIFT TECHNIQUES
The most popular forms of artificial lift are illustrated in figure 2. They are:
Rod
Pump
TubingAnchor
Power Fluid
HighPressure
PowerFluidProduction
Production Production
Production
Electric Motor
Gas FlowMeter
Gas Flow
ElectricPower Cable
PumpGas Lift Valve(unloading)
Operating Valve
Stator
Rod
Rod
Motor
Fluid DrivenPump
(i) Rod Pump (ii) Hydraulic Pump (iii) Submersible ElectricPump
(iv) Gas lift (v) Progressing Cavity Pump(May also be driven by
electric submersible motor)
Production
Gas Flow RateControl Valve
Figure 1c
Installation of pump
suction below the
perforations maximises
potential drawdown and
production rates
Figure 2
The most popular types of
artificial lift
6
(i) Rod Pumps - A downhole plunger is moved up and down by a rod connectedto an engine at the surface. The plunger movement displaces produced fluidinto the tubing via a pump consisting of suitably arranged travelling andstanding valves mounted in a pump barrel.
(ii) Hydraulic Pumps use a high pressure power fluid to:
(a) drive a downhole turbine pump or
(b) flow through a venturi or jet, creating a low pressure area which producesan increased drawdown and inflow from the reservoir.
(iii) Electric Submersible Pump (ESP) employs a downhole centrifugal pumpdriven by a three phase, electric motor supplied with electric power via a cablerun from the surface on the outside of the tubing.
(iv) Gas Lift involves the supply of high pressure gas to the casing/tubing annulusand its injection into the tubing deep in the well. The increased gas contentof the produced fluid reduces the average flowing density of the fluids in thetubing, hence increasing the formation drawdown and the well inflow rate.
(v) Progressing Cavity Pump (PCP) employs a helical, metal rotor rotating insidean elastomeric, double helical stator. The rotating action is supplied bydownhole electric motor or by rotating rods.
In fact, nearly all the major classes of pumps are employed in the various forms ofartificial lift (figure 3).
Pumps
Dynamic
OtherJet or Venturi Pump
Centrifugal
Axial
Electric
Submersible
Pump
Progressing
Cavity
Pump
Rod Pump
?
Rotodynamic
Rotary
Reciprocating
PositiveDisplacement
4. CURRENT STATUS OF THE ROLE OF ARTIFICIAL LIFT INFIELD DEVELOPMENT
Figure 4 shows relative frequency of the different types of artificial lift installed in theUSA in 1992. The predominance of rod pumps indicates the vast majority of wellsare on land locations in mature fields with low well production.
Figure 3
Pump classification
Department of Petroleum Engineering, Heriot-Watt University 7
2Selection of Artificial Lift Types2
Electric Submersible Pump 4%
Others < 1%
Gas Lift 10%Rod Pump – 85%
Progressing Cavity Pump < 1%
Hydraulic Pump < 2%
Figure 5 is a corresponding breakdown for a major international oil company’s morethan two and a half million barrels a day of gross fluid production (which yields morethan one million barrels a day of oil) which is lifted by the various types of artificiallift. The larger contribution from Gas Lift and ESPs reflects the greater contributionof high rate and offshore wells compared to the figures for the USA.
Gas lift 51%
Beam Pumping 17% Electric Submersible Pump 29%
Progressing Cavity Pump 2%
Hydraulic 1%
Venturi 1%
Artificial lift is being more widely applied in field development than ever before dueto:
(i) Field development status - oil producing provinces such as the North Seahave become mature with the consequent reductions in flowing bottom holepressure (depletion) and increasing water cuts.
(ii) Absence of Pressure maintenance. The development plans for many of theearly, giant North Sea fields employed early water injection to maintain thereservoir pressure above the hydrocarbon fluid’s bubble point, even after asignificant fraction of the hydrocarbon reserves had been produced. This meantthat the high water cut wells still continued to flow at high production rates.Many of the current crop of smaller fields currently beingdeveloped do notemploy any form of pressure maintenance, resulting in a early need for artificiallift.
(iii) Satellite or Subsea Wells. These wells are often positioned a considerabledistance from the host platform. The extra pressure drop caused by flowthrough these long, subsea pipelines needs to be overcome by some form ofpressure boosting. This could either be an increased pressure boost from anESP installed downhole or by a multiphase pump mounted on the sea bed.
Figure 4
Relative frequency of
different types of artificial-
lift methods installed in the
USA in 1992 (data from
J.Clegg, S.Buchan and
N.Heln, JPT, December
1993, p1128)
Figure 5
Production from artificial
lift wells for a major
international oil company
8
(iv) Business drivers. Profitable field development requires that the average wellproduction rate exceed a minimum value with higher values being moreprofitable. Well design can increase the well flow rate of return. Recent welldesign innovations include:
(a) Advanced well design: drilling of long (horizontal) exposures to theproducing formation.
(b) Large diameter tubing to decrease the frictional pressure lossese.g. Norske Shell’s Draugan field achieved flow rates of over 76,000bopd with a 9 in. production tubing.
(c) Early installation of artificial lift to increase the flow rate.
Marginal (possibly subsea) field development requires reservoir developmentplans with a minimum number of wells where there is little or no need for wellintervention to repair or modify the downhole installation. High equipmentreliability is thus a “must”. Developments in the application of downholeelectronics and measurement sensors means that monitoring of the performanceof the artificial lift equipment of the operating conditions results in improvedperformance. One operator found that implementation of a real time,Supervisory, Control and Data Analysis (SCADA) system in a large artificiallift project consisting of more than 500 rod pumped wells resulted in a:
(a) 6% production increase (lift conditions optimised and immediate alarmgiven when wells ceased producing).
(b) 50% reduction in well entries (earlier recognition of developing problemsallowed preventive maintenance).
(c) 5% reduction in energy consumption (wasteful “over-lifting” operatingconditions recognised immediately).
(v) Integration of artificial lift software into field development planning andoperation. Design procedures for artificial lift techniques has undergone atremendous development:
(a) 1975: nomograms, slide rules and early calculators allowed single well,“snap shot” optimisation for the current production conditions.
(b) 1985: early, software based, field wide surveillance systems becameavailable.
(c) 1995: “Thinking in systems” became a reality. The first field/productionsystem models became available. These contain (simple) versions ofmodels describing the performance of the reservoir, well (including anyinstalled form of artificial lift), manifolds, facilities and pipeline (Figure 6).This integrated system has the ability to:
Department of Petroleum Engineering, Heriot-Watt University 9
2Selection of Artificial Lift Types2
• Calibrate (or automatically history match) the various modelelements against actual measured data.
• Compare the measured field data with the production plan orforecast.
• Optimise (normally maximise) the production within any(permanent or temporary) constraints.
• Recognise discrepancies between forecast and actual production.The identified wells/production systems are thus listed forengineering investigation and possible remedial activities.
Integrated Production Model Compare to ForecastAuto History Match
Simple
ReservoirModel
Simple
WellModel
Simple
Manifoldand
PipelineModel
Field Datafor Model
calibration and
comparisonwith
production plan/forecast
Results
Remedialactivities
and Archiving
Simple
ProcessModel
Building and operation of such a model requires input from geoscientists, productiontechnologists, pipeline and process engineers as well as production operations staff.
The availability of such an integrated models allows the “total cost of ownership”, or“total lifetime cost” of a particular form of artificial lift, to be evaluated when selectingthe preferred artificial lift technique for a particular field development. For example,two or more lift methods may be technically capable of producing a well at thedesignated production rate. It is relatively easy to obtain figures for the initial, capitalcosts of installation. However, the equally (or even more important) costs due to(un)reliability, energy consumption, maintenance, manpower etc. figures have to beobtained from field operational data.
(vi) Technical innovation has increased the scope of artificial lift. One example isthe development of multiphase pumps which are now available for both subseaand downhole application. For example, the B area of Captain field in the NorthSea contains a viscous (50-150 cp) crude oil bearing zone overlain by a gas cap.Production of these reserves was only possible due to development of apump capable of operating with free gas fractions in excess of the 30% vol (thenormal operating limit of a conventional ESP). Innovation has also resulted in:
(a) the development of hybrid technologies such as downhole separationand pressure boosting for oil production and water injection.
(b) a step change in artificial lift reliability (see Table 1) through:
Figure 6
An integrated production
model
10
• improved engineering design
• the ability to monitor downhole conditions from the surface.
• better materials selection and, most importantly,
• better training of wellsite personnel who install and operate theequipment.
Date 1970 1990 1983 1990 1997 1990 1993 1997
Data Source Amoco* Thums♦, California North Sea†
Lift Type Mean Time Before Failure (months)
Sucker Rods 20 75
Rod Pump 20 40
ESP 15 48 10 20 37 12 30 56
* Private communication J. Lea, Amoco.♦ T. Lutz, presented at Artificial Lift 1997, Dubai.† Presented at 1997 SPE Artificial Lift Workshop.
5. SELECTION OF ARTIFICIAL LIFT CRITERIA
There are many factors that influence which is the preferred form of artificial lift.Some of the factors to be considered are:
5.1. Well and Reservoir Characteristics(i) Production casing size.
(ii) Maximum size of production tubing and required (gross) production rates.
(iii) Annular and tubing safety systems.
(iv) Producing formation depth and deviation (including doglegs, both planned andunplanned).
(v) Nature of the produced fluids (gas fraction and sand/wax/asphaltene production).
(vi) Well inflow characteristics. A “straight line” inflow performance relationshipassociated with a dead oil is more favourable than the curved “Vogel” relationshipfound when well inflow takes place below the fluid’s bubble point. Figure 7shows that reducing the flowing bottomhole pressure from 2500 to 500 psiincreases the well production rate by 125% for the dead oil. This is more thandouble the 60% increase expected for the same reduction in bottom hole pressureif a “Vogel” type inflow relationship is followed with a well producing belowthe bubble point.
Table 1
Artificial Lift Reliability
Department of Petroleum Engineering, Heriot-Watt University 11
2Selection of Artificial Lift Types2
Figure 7
Influence of fluid in flow
performance on production
increase achieved when
well drawdown is increased
Vogel Curve(Below Bubble Point)
~ 60% Increase in Production
125% Increase in Production
"Straight Line" IPR
5000
4500
4000
3500
3000
2500
2000
1500
1000
500
00 60 120 180 240 300 360 420 480 540 600 660 720 780
Total Liquid Production, (BFPD)
Flo
win
g B
otto
mho
le P
ress
ure
(PS
I)
Reductionin FBHP due to artificial lift installation
5.2. Field Location(i) Offshore platform design dictates the maximum physical size and weight of
artificial lift equipment that can be installed.
(ii) The on-shore environment can also strongly influence the artificial lift selectionmade. For example:
(a) an urban location requiring a maximum of visual and acoustic impact or
(b) a remote location with minimal availability of support infrastructure
can lead to different artificial lift types being selected for wells of similar design andproducing characteristics.
(iii) Climatic extremes e.g. arctic operations will also limit the practical choices.
(iv) The distance from the wellhead to the processing facilities will determine theminimum wellhead flowing pressure (required for a give production rate). Thismay, for example, make the choice of an ESP more attractive than Gas Lift. Thisis because the extra pressure drop in the flowline, due to the injected gas, makesGas Lift an unsuitable option for producing satellite hydrocarbon accumulationsisolated from the main field.
(v) The power source (natural gas, mains electricity, diesel, etc) available for theprime mover will impact the detailed equipment design and may effect reliabilitye.g. the voltage spikes often associated with local electrical power generationhave been frequently shown to reduce the lifetime of the electrical motors for ESP’s.
5.3. Operational Problems(i) Some forms of artificial lift e.g. gas lift are intrinsically more tolerant to solids
production (sand and/or formation fines) than other forms e.g. centrifugal pumps.
12
(ii) The formation of massive organic and inorganic deposits - paraffins, asphaltenes,inorganic scales and hydrates - are often preventable by treatment with suitableinhibitors. However, additional equipment and a more complicated downholecompletion are required unless, for example, the inhibitor can be carried in thepower fluid for a hydraulic pump or can be dispersed in the lift gas.
N.B. The physics and chemistry of these processes was discussed in Chapter 4entitled “Formation Damage”.
(iii) The choice of materials used to manufacture the equipment installed within thewell will depend on the:
(a) Bottom Hole Temperatures.
(b) Corrosive Conditions e.g. partial pressure of any hydrogen sulphide andcarbon dioxide, composition of the formation water etc.
(c) Extent of Solids Production (erosion).
(d) Producing Velocities (erosion/corrosion).
5.4. Economics(i) A lot of attention is often paid to the initial capital investment required to install
artificial lift. However, the operating costs are normally much more importantthan the capital cost when a full life cycle economic analysis is carried out. Thiswas illustrated by J Clegg et al (JPT, December 1993, p 1128). His data has beenused to prepare Figure 8. It can be seen that, for this well, the capital costrepresents a small proportion of the total project costs. Thus it is often viable toinvest extra to ensure the best equipment is installed in the well if this will resultin increased revenue (production) and/or reduced operating costs.
Nominal Decline = 25% / Year
Economic Limit
200
100987
6
4
3
3.53
0 5 10 15 20
Energy 6%
Maintainence and Operators 6%
Capital Cost 1%Net Revenue 87%
Time - (Years)
Net
Pro
duct
ion
Rat
e (B
OP
D)
Land Well Major Contributing Factors to Project Net Value
Reserves290 x 103 bbl oil
(ii) Good operating cost data for the different artificial lift methods in differentlocations is difficult to find. Reliability (discussed earlier) is one key issuewhile the second is energy efficiency (and hence energy costs). This latter is
Figure 8
Example full life cycle
eceonomics
Department of Petroleum Engineering, Heriot-Watt University 13
2Selection of Artificial Lift Types2
more tractable since it can be calculated from first principles. There is a widevariation - see Figure 9. Only rod pumps, ESP’s and PCP’s show values >50%while gas lift, particularly of the intermittent variety, is inefficient in energyterms. Changing energy costs can alter the ranking order of the various artificiallift methods.
80
70
60
50
40
30
20
10
0Rod
PumpPCP ESP Turbine
HydraulicPump
VenturiHydraulic
Pump
GasLift
IntermittentGas Lift
Ene
rgy
Effi
cien
cy (
%)
(iii) Maintenance costs will vary between operating locations depending on thestate of the local, service company infrastructure. It can be costly in remotelocations.
(iv) The number of wells in the field with that particular form of artificial lift(economy of scale) will influence the operating costs.
(v) Similarly, the desirability and/or need for automation (how many operators areto be employed) and the decision as to whether or not to install centralisedfacilities will also influence the operating costs.
(vi) The speed with which the “learning curve” is climbed for the more sophisticatedforms of artificial lift will depend on the training provided and the skill base ofthe operations staff.
5.5. Implementation of Artificial lift Selection TechniquesAs discussed the artificial lift design engineer is faced with matching facilityconstraints, artificial lift capabilities and the well productivity so that an efficient liftinstallation results. Frequently, the type of lift has already been determined and theengineer has the problem of applying that system to the particular well. A morefundamental question is how to determine the optimum type of artificial lift to applyin a given field.
There are certain environmental and geographical considerations that may be overriding.For example, sucker rod pumping is by far the most widely used artificial lift methodin North America. However, sucker rod pumping may be eliminated as a suitable form
Figure 9
Comparison of the energy
efficiency of the major
artificial-lift methods
adapted from J.Clegg et al
(JPT, December 1993,
p1128)
14
of artificial lift if production is required from the middle of a densely populated cityor on an offshore platform with it’s limited deck area. There are also practicallimitations - deep wells producing several thousands of barrels per day cannot be liftedby rod pumps. Thus, geographic and environmental considerations may make thedecision. However, there are many considerations that need to be taken into accountwhen such conditions are not controlling.
Some types of artificial lift are able to reduce the sand face producing pressure to alower value than others. The characteristics of the reservoir fluids must also beconsidered. Wax & formation solids present greater difficulties to some forms ofartificial lift than others. The producing gas-liquid ratio is key parameter to beconsidered by the artificial lift designer. Gas represents a significant problem to allof the pumping methods; while gas lift, on the other hand, utilizes the energy containedin the produced gas and supplements this with injected gas as a source of energy.The “Advantages and Disadvantages of the Major Artificial Lift Methods” are listedand compared in Tables 2 & 3.
Rod Pumps
Simple, basic design
Unit easily changed
Simple to operate
Can achieve low BHFP
Can lift high temperature,viscous oils
Pump off control
ElectricSubmersible Pump
Extermely high volume lift using up to1,000 kw motors
Unobtrusive surface location
Downhole telemetryavailable
Tolerant high wellelevation / doglegs
Corrosion / scaletreatments possible
Venturi Hydraulic Pump
High volumes
Can use water aspower fluid
Remote power source
Tolerant high welldeviation / doglegs
Gas Lift
Solids tolerant
Large volumes in highPI wells
Simple maintenance
Unobtrusive surface location / remote powersource
Tolerant high welldeviation / doglegs
Tolerant high GORreservoir fluids
Wirleine maintenance
Progressing Cavity Pump
Solids and viscous crude tolerant
Energy efficient
Unobtrusive surface location with downhole motor
Rod Pumps
Friction in crooked / holes
Pump wear with solidsproduction (sand, wax etc.)
Free gas reduces pumpefficiency
Obtrusive in urban areas
Downhole corrosion inhibition difficult
Heavy equipment foroffshore use
ElectricSubmersible Pump
Not suitable for shallow, low volume wells
Full workover requiredto change pump
Cable susceptible todamage during installation with tubing
Cable deteriorates at high temperatures
Gas and solids intolerant
Increased productioncasing size oftenrequired
Venturi Hydraulic Pump
High surface pressures
Sensitive to change in surface flowlinepressure
Free gas reduces pumpefficency
Power oil systemshazardous
High minimum FBHP.Abandonment pressuremay not be reached
Gas Lift
Lift gas may not be available
Not suitable for viscous crude oil or emulsions
Susceptible to gas freezing / hydrates atlow temperatures
High minimum FBHP.Abandonment pressuremay not be reached
Casing must withstand lift gas pressure
Progressing Cavity Pump
Elastanes swell in some crude oils
Pump off control difficult
Problems with rotating rods (windup and after spin) increase with depth
Table 3
Disadvantages of major
artificial lift methods
Table 2
Advantages of major
artificial lift methods
Department of Petroleum Engineering, Heriot-Watt University 15
2Selection of Artificial Lift Types2
5.6. Long Term Reservoir Performance and Facility ConstraintsAnother factor that needs to be considered is long term reservoir performance. Someyears ago Neely indicated that two approaches, both of which have disadvantages, arefrequently used to solve the problem of artificial lift selection and sizing.
(i) A prediction of long term reservoir performance is made and artificial liftequipment installed that can handle the well’s production and producingconditions over its entire life. This frequently leads to the installation of oversizedequipment in the anticipation of ultimately producing large quantities of water.As a result, the equipment may have operated at poor efficiency due to under-loading overa significant portion of its total life.
(ii) The other extreme is to design for what the well is producing today and notworry about tomorrow. This can lead to many changes in the type of liftequipment installed during the well’s producing life. Low cost operations mayresult in the short term, but large sums of money will have to be spent later onto change the artificial lift equipment and/or the completion.
Likewise, in a new field development, the fluid handling requirement from someartificial lift types can significantly increase the size and cost of the facilities required.Only the produced fluid is handled through the facilities with rod pumps and ESPs.However, gas lift requires injection gas compression and distribution facilities and theadditional, produced gas increases the size of the production facilities required.Similarly, the use of Hydraulic pumps can result in the additional power fluid volumesbeing many times that of the produced oil volume. This results in high fluid handlingcosts as well as difficulties in accounting for the oil produced (when oil is used as apower fluid).
The selection of the artificial lift for a particular well must meet the physicalconstraints of the well. Once a particular type of lift is selected for use, considerationshould be given to the size of the well bore required to obtain the desired productionrate. It can happen that the desired production can not be obtained because the casingprogramme was designed to minimize well cost, resulting in a size limitation on theartificial lift equipment that can be installed. Even if production rates can be achieved,smaller casing sizes can lead to higher, long term production costs due to wellservicing problems, gas separation problems etc.
Figure 10 is offered as a screening selection tool in which areas where particularartificial lift methods have been frequently applied are compared as a function ofdepth and well rate. It must be realised that there are many proven applications wherea particular form of artificial lift has been installed in a well at greater depths orproduced at higher rates than is indicated in this figure.
16
1 10 100 1000 10000 100000
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
0
Gross Production Rate (bpd)
Dep
th, F
t.
1 10 100 1000 10000 100000
2000
4000
6000
8000
10000
12000
14000
16000
18000
0
Gross Production Rate (bpd)
Dep
th, F
t.
IntermittentPlungerGas Lift
SuckerRod
ESP
HydraulicVenturi
HydraulicReciprocating
PCP
ContinuousGas Lift
6. ROD PUMPS
6.1. Introduction(Sucker) rod or beam pump was the first type of artificial lift to be introduced to theoil field. It is also the most widely used in terms of the number of installations worldwide. In 1993, some 85% of the USA population of artificially lifted wells wasproduced by rod pumps and more than 70% of these produced less than 10 barrels ofoil per day. The low cost, mechanical simplicity and the ease with which efficientoperation can be achieved makes rod pumps suitable for such low volume operations.
Rod pumps can lift moderate volumes (1,000 bfpd) from shallow depths (7,000 ft) orsmall volumes (200 bfpd) from greater depths (14,000 ft). They are normallymanufactured to standards set by the American Petroleum Institute (API). This meansthat, unlike other artificial lift methods, the equipment manufactured by the varioussupplies is fully interchangeable.
6.2. The Pumping UnitThe surface equipment for a rod pump is illustrated in Figure 11. The prime mover,normally an electric motor or gas engine, drives a speed reducing set of gears so thatits fast rotation, of say 600 revolutions per minute, is reduced to as low as 20 strokes
Figure 10
Typical application areas of
artificial lift techniques
Department of Petroleum Engineering, Heriot-Watt University 17
2Selection of Artificial Lift Types2
per minute or less. The connection between the surface pumping unit and thedownhole pump is the polished rod and the sucker rods. The polished rod moves upand down through a stuffing box mounted on top of the wellhead. This stuffing boxseals against the polished rod and prevents surface leaks of the liquids and gassesbeing produced by the well.
Gear Reducer
Horse Head
Bridle
V BeltPrime Mover
Clamp
Polished Rod
Stuffing Box
Tubing Head Flow line
Tubing
Casing Head
Sucker Rod
Casings PUMPING UNIT
Beam
Counter Weight
6.3. The Sucker RodsThe sucker rods, typically 25 ft long, are circular steel rods with diameters between0.5 in and 1.125 in, in increments of 0.125 in. A threaded male connection or pin ismachined at each end of the rod. The two rods can be joined together by use of a doublebox coupling (Figure 12). Square flats are machined near the pins and at the centreof the coupling to provide a grip for a wrench to allow the rods and couplings to bescrewed together. The sucker rods are subjected to continuous fatigue when the pumpis in operation. The weight of the rod string is one component of this fatigue load -it can be minimised by using a tapered sucker rod string. This involves installinglighter, smaller diameter rods lower down in the well where the load they have tosupport (weight of rods and fluid in the tubing string) is less than at the top of the well.
Double Box Coupling
Square Flat for Wrench
PinPin
Circular Sucker RodCircular Sucker Rod
6.4. The PumpThe pump is located near the perforations at the bottom of the string of sucker rods.Figure 13 shows that it consists of a hollow plunger with circular sealing ringsmounted on the outside circumference moving inside a pump barrel which is eitherinserted into the tubing or is part of the tubing itself. A standing valve is mounted at
Figure 12
Sucker rods are joined
together by a coupling
Figure 11
The surface equipment for
a rod pump
18
the bottom of the pump barrel while the travelling valve is installed at the top of theplunger. The standing and travelling valves consist of a ball which seats (closes off)an opening.
Tubing
PolishedPumpBarrel
TravelingValveOff Seat
Plunger
TravelingValve
On Seat
Fluidbeing
Lifted toSurface
StandingValveOn Seat
StandingValve
Off Seat
StandingValve
On Seat
TravellingValve
Off Seat
Fluid Inflow FromPerforations
Fluid Flowfrom Pump
Barrel toTubing
Sucker Rods
FliudFlow
SealingRings
AroundPlunger
Circumference
Up-wardRod Movement
DownwardRod Movement
The “UP” and “DOWN” movement of the pump barrel allows the fluid flow to openand shut these valves as shown in Figure 13. The left hand schematic shows theplunger status at the end of the “DOWN” stroke. The "Upward" rod movementreduces the pressure within the pump barrel and the upward flow of fluid from belowthe pump lifts the standing valve’s ball off its seat. The pressure due to the fluidcolumn above the plunger keeps the travelling valve ball on its seat. The situation isreversed during the “DOWN” stroke - compression of fluid within the pump barrelforces it to flow through the hollow plunger and to lift the travelling valve off its seat;while ensuring that the standing valve remains closed.
6.5. Rod Pump OperationThis chapter describing of rod pump operation deals with some of their moreimportant operational aspects.
6.5.1. Pump-Off ControlThe well will produce the maximum gross volume of fluid when the drawdown ismaximised i.e. the fluid level in the well is maintained at a limited distance above thepump. The pump capacity will often be greater than the well inflow capacity - thepump motor must be stopped at regular intervals when the fluid level is reduced to aspecified, minimum safety level above the pump. This monitoring is often performedwith an “Echometer”. Figure 13a. This tool is attached to the wellhead in a pressuretight housing. It consists of a firing mechanism, a microphone and an amplifierrecorder. Typically, a gas actuated device generates an accoustic pulse at the wellhead.
Figure 13
Rod pump operation
Department of Petroleum Engineering, Heriot-Watt University 19
2Selection of Artificial Lift Types2
Fluid levelFluid level
Pump
Reflection
Accoustic pulse
Casing collars
Reflection fromtubing collars
RodsTubing
Reflected wave amplifiedand arrival time recorded.
Charge ignited
Sonologequipment
time
0
This is reflected back by subsurface items such as tubing collars, but the mainreflection is from the fluid level at the bottom of the casing. The depth of the fluid levelcan now be found by multiplying this time by half the velocity of sound in the casing/tubing annulus.
The pump can now be restarted once the casing fluid level has risen sufficiently dueto inflow from the reservoir.
Once calibrated, the pump unit can be put on timer control, i.e. the “Echometer” neednot be used continuously since well inflow performance normally shows a steady,predictable decline rate with time. Further, the performance of the pump itself can bechecked by measuring the dynanometer card at regular intervals (see section 2.6.5.5on Pump Diagnosis).
6.5.2. Gas Influx(Free) gas sucked into the pump will reduce the pump efficiency due to its compressiblenature. Placing the pump below the perforations maximises the use of the (limited)gas separation capacity of the casing. Minimising the volume between the travellingvalve at the bottom and the downstroke and the standing valve helps ensure the gasis pushed out of the pump during each down stroke.
Placing the pump below the perforations is advantageous since it increases themaximum possible drawdown. This is not always possible in practice. Many types of“gas anchors” have been tried in order to overcome the resulting difficulties. They allaim to separate the “free” gas from the liquid prior to the liquid entering the pump. Thegas flows into the tubing/casing annulus where it is vented or gathered at as low apressure as possible (to allow a minimum bottom hole pressure to be reached). Anexample of an effective gas anchor is shown in Figure 14. Here a packer and crossoverdirect the multiphase flow above the pump, ensuring that only liquid is sucked into thepump barrel. (The extra packer also allows the tubing to be anchored - minimising thefatigue loads to which the rod string is exposed.
Figure 13a
"Echometer" or "Sonolog"
fluid level survey.
20
Casing
Rods
Tubing
Liquid Flow
Crossover
Packer
Pump Barrel and Valves
Gas
Produced Gasand Liquid Pump Intake
(Perforated Pipe)
BypassTube
Producing Zone
6.5.3. CentralisersThe sucker rods may require centralisers or protectors in deviated wells to reduce wearon the tubing and rods (Figure 15). This requirement becomes more extreme incrooked or highly deviated wells, sometimes to the extent that rod pumps can not beused.
Extra centraliser / protector required hereto prevent rod / tubing wear
Centraliser / Protector
Centraliser / Protector
Tubing
Sucker Rod
Rod Movement
Figure 15
Centralising rod reduces
friction and wear
Figure 14
Packer type gas anchor
Department of Petroleum Engineering, Heriot-Watt University 21
2Selection of Artificial Lift Types2
6.5.4. SolidsRod pumps also have a very limited ability to lift sand due to the low fluid velocityin the production tubing plus wear on the pump valves, seats and plunger. The lattercan be overcome by suitable pump design and choice of construction materials.
Wax and inorganic scale deposition also interfere with efficient rod pump operation.Continuous injection of an inhibitor below the pump to ensure protection of thecomplete downhole equipment is complicated unless there is a (normally unacceptable)complication of the completion design. Removal of wax by hot oil/solvent circulationor injection of a scale inhibitor into the formation are also possible (see Chapter 4).Further, recovery of the pump and rods using a well pulling hoist is often a relativelylow cost, simple operation.
6.5.5. Pump DiagnosisThe condition of the pump can be evaluated by measuring the load at the top of thepolished rod as a function of its position i.e. as it moves up and down during the strokelength. This is recorded in the form of a dynamometer card. Examples of theoreticaldynanometer cards are shown in Figure 16.
"Up" Stroke
"Down" Stroke
Polished Rod Up
Polished Rod Down
Stroke Position
Load
Maximum Load
Maximum Load
Walking Beam Decelerating
Walking Beam Decelerating
Standing ValveCloses
Travelling ValveCloses
Minimum Load
Minimum Load
(a) Inelastic Rods
"Up" Stroke
"Down" Stroke
Polished Rod Up
Polished Rod Up(Maximum Stroke Position)
Polished Rod down (Minimum Stroke Position)
Recoil
Rods and Fluid Being lifted
Rods and Plunger Falling Through Fluid
Rod
Stre
tche
s
Rod
Rel
axes
Polished Rod Down
Stroke Position
Load
Load
Maximum Load
Minimum Load
(b) Elastic Rods
Stroke Position
(c) Full Simulation
Figure 16
(Theoretical) dynamometer
cards for:
(a) inelastic rods
(b) elastic rods
(c) elastic rods with rod,
fluid and surface pump unit
22
(i) Figure 16 (a) records the variation in load for inelastic rods. The load is eitherhigh or low depending on whether the polished rod is moving up or down.
(ii) Figure 16 (b) adds the elasticity of the rods - the full increase in load is no longerinstantaneous when the polished rod starts moving in a particular direction.
(iii) Figure 16 (c) adds the further dimension of rod - fluid and surface pump unitdynamics. The times at which the various processes become controlling duringthe pump cycle are indicated.
These theoretical calculations have been made for a perfectly operating pump unitpumping liquid only. Practical problems such as:
(i) excessive rod or pump friction
(ii) restriction in the flow-path
(iii) vibrations
(iv) sticking plunger
(v) gas lock etc.
will all alter the shape of the dynamometer trace in a distinctive manner allowing thesource of the problem to be diagnosed and then rectified.
6.6. Pump DesignAPI Recommended Practice 11L, published by the American Petroleum Institute,describes a field proven method for designing all elements of a Rod Pump. We willonly discuss selected elements here.
6.6.1. Pump RateThe pump rate (Q) is related to the volume displaced (V) by each pump stroke and thespeed rate or number of strokes per minute (N). Thus:
Q = K*V*N* φ = K*A*S*N* φ
Where:
A is the area of the pump barrelS is the length of the pump strokeφ is the efficiency factorK is a constant to convert the above units to barrels per day
The maximum speed (N) of the pump unit is determined by the speed at which thesucker rods fall downward in the “DOWN” stroke. (Early pump rod failure occurs dueto metal fatigue if they are placed under compression due to them being forceddownwards by the pump unit exceeding this maximum speed.) As would be expected,this maximum speed decreases as the length of the pump stroke increases. Typicalmaximum values are quoted in Table 4.
Department of Petroleum Engineering, Heriot-Watt University 23
2Selection of Artificial Lift Types2
Maximum Allowable Pump Speed†
For a Conventional Pump Unit* Strokes Per Minute
Stroke Length (in.) 30 60 90 120 180 240 300
Maximum Pump Speed (SPM*) 34 24 19 17 14.5 11.5 10.5
N.B. Low pump speeds and large diameter plunges lead to the greatest energyefficiency, but also the largest equipment loads. It is common practice to put a fewlarger rods capable of carrying any compression loads due to buckling at the bottomof the rod string. Also, the addition of sinker bars will increase the rate of rod fall, butalso increase the load on the rods.
Two factors which reduce the efficiency of the pump are gas influx (see section2.6.5.2) and rod stretch.
6.6.2. Rod Stretch and Over TravelDuring the “UP” stroke the rods support their own weight and that of the fluid in thetubing. Hooke’s law of elasticity dictates that the rods will increase in length inresponse to this load. This decreases the effective travel of the plunger downholecompared to the distance moved at the surface. For example, 0.875 in sucker rodsdriving a 2 .25 in plunger pump set at 6,000 ft will show a stretch of 29 in when liftingfresh water.
The load imposed on the rods by the fluid is related to the area of the plunger timesthe hydrostatic head. Rod stretch thus increases when a larger diameter pump isselected. There is a similar effect of tubing stretch. This also reduces the effectivestroke length, if the tubing is not anchored.
Plunger Overtravel”, by contrast, increases the effective stroke length. This effect iscaused by elongation of the rods due to dynamic forces generated by the weight of therods reversing direction during the pumping cycle. At this point, the weight of a 6,000ft string of 0.875 in rods (weight 2.25 lb/ft or 13,500 lb total weight) will be broughtto a halt and reverse direction over a time period of less than one second. The rod stringvelocity will be approximately 6 ft/second when the plunger nears the bottom of itsstroke if the pump unit is operating with a 64 in stroke length at a pump speed of 15strokes per minute. The resulting increase in length of the rods generated by haltingand reversing this momentum is called “Plunger Overtravel” (approximately 14"according to Marsh and Coberly’s 1931 method for this example).
We can now estimate that:
Effective stroke length = surface stroke length - rod stretch + plunger overtravel≈ 64" - 29" + 14" ≈ 49"
6.6.3. Pumping Limit Load CalculationsAPI RP 11L can be used to calculate the maximum and minimum polished rod loads,the peak torque and the theoretical horsepower required once the pump speed, strokelength, plunger diameter and rod sizes have been chosen. Typically, a motor powerof twice this theoretical value should be installed to allow for surface and downholeenergy losses.
Table 4
Maximum allowable pump
speed
24
It was mentioned previously that the rod string was continuously subjected to fatigue.Goodman showed how the “Maximum Allowable Stress” on the sucker rods was afunction of the grade of the sucker rods and the minimum polished rod load/crosssectional area of the top rod. This “Maximum Allowable Stress” has to be decreasedby a service factor related to the operating conditions e.g. the presence of corrosive saltwater or, more importantly, hydrogen sulphide. This explains why these aggressivefluids limit the application (either maximum depth of a pump installation or maximumfluid volume which can be lifted) of rod pumps unless more expensive, specialitygrade sucker rods are used.
7. ELECTRIC SUBMERSIBLE PUMPS (ESPs)
7.1. IntroductionElectric Submersible Pumps (ESP’s) are a versatile form of artificial lift with pumpsranging from 150 to 60,000 bfpd in operation. A typical low pressure well that is beingartificially lifted using an ESP system is illustrated in figure 17. The functions of thevarious components are summarised as follows:
Switchboardand Motor
Controls
High Voltage Electrical Supply
Vent Box
Cable PenetratorThrough Wellhead
Surface Cable
Casing
Centraliser
Flat Cable
Cable
Cable Banded to Tubing
Cable Protector
Tubing
Centrifugal Pump
Pump Intake withOptional Gas Separator
Seal or Protector
Pothead ConnectsCable to Motor
Downhole SensorPackage
Electric Motor
Fluid
Pump Discharge Head
Figure 17
A well completed with
artificial lift using an
electric submersible
centrifugal pump
Department of Petroleum Engineering, Heriot-Watt University 25
2Selection of Artificial Lift Types2
Frequency Drive (VFD). A VFD allows the speed of the electric motor to bealtered e.g. starting the pump using the “nameplate” design frequency of 50Hz(Europe) or 60Hz (North America) results in high instantaneous electric motorcurrents since the power developed by the pump is proportional to thefrequency. These can be reduced by supplying the electric power at lowerfrequencies. It also allows the pump flow rate to be adjusted to the well inflowconditions since flow rate is also proportional to frequency.
Practical experience shows that a 60Hz motor can be operated between35Hz and 80Hz. VFD installation increases the surface energy losses fromsome 3% to 5-15% of total power supplied.
(ii) The vent box separates the surface cable from the downhole cable. This ensuresthat any gas, which travels up the downhole cable, does not reach the electricalswitchgear.
(iii) The downhole cable penetrates the wellhead. It is banded to the tubing atregular intervals. Additional protection is supplied by cable protectors whichare installed at critical points to prevent damage while the completion is beingrun into the hole. A “flat pack” cable shape is employed across the largerdiameter completion components to minimise total width. The cable enters theelectric motor housing at the Pothead. It not only carries the electrical powersupply for the motor (up to 750HP motors are being routinely installed), butalso carries the measurement signal from the downhole sensor packageinstalled underneath the motor.
(iv) The pump unit consists of a stacked series of rotating centrifugal impellersrunning on a central drive shaft inside a stack of stationary diffusers, i.e. it isessentially a series of small turbines. The pressure increase is proportional tothe number of stages while the pump capacity (volume) increases as thediameter of the impeller increases. Rotation of the impeller accelerates theliquid to be pumped which is then discharged into the diffuser where this kineticenergy is transformed into potential energy i.e. a pressure increase. Theimpeller/diffuser pairs are arranged in series with the discharge of one unitbeing the suction of the next one. The number of pump stages (impeller/diffuser) pairs may range between 10 and more than 100, depending on thepressure increase required. Abrasion resistance to produced solids is verydependent on the detailed design and materials selection employed duringpump design.
However, as discussed at the beginning of this chapter, ESP’s with their rapidlyrotating internals are not really compatible with large quantities of produced sandeven when hardened, wear resistant materials are used. The option of using otherforms of artificial lift - gas lift and PCPs - should be considered.
Standard pump impellers are very sensitive to gas fractions greater than 20% vol in theproduced fluid. Alternatively, changes in pump design such as altering the designof the impeller from pure radial flow to mixed (i.e. a combination of both radial andaxial) flow can double the gas/fluid ratio to 40% vol.
26
Motor Housing
Drive Shaft
Impeller
A tapered pump design using mixed flow impellers in the lower pressure stages(with the higher gas volume fractions) and radial flow impellers in the upper stagescan prove to be effective.
(v) The pump intake may include a rotary gas separator if gas fractions higher than20%. This consists of a centrifugal device, which separates the lower density,gaseous phase from the denser liquid phase. The latter is concentrated at thecentre of the device and enters the pump suction while the lighter, gas phase isdirected towards the casing/tubing annulus where gas is vented/gathered atsurface.
A reduction in casing/tubing annulus pressure increases the maximum achievabledrawdown at the formation face. A single rotary gas separator can increase theESP’s gas handling capabilities up to 80% vol. Two separators, arranged intandem, are even more efficient, increasing the pumpable gas fraction to >90%vol. However, the addition of extra equipment always comes with the cost ofgreater operational problems e.g. produced formation solids can damage therotary separator, scale formation can unbalance rapidly rotating equipment.Some operators will not use them due to these problems which have resultedin rotary separators having a poor reputation for reliability .
Figure 17(a)
Cutaway schematic
drawing of Electric
Submersible pump
Department of Petroleum Engineering, Heriot-Watt University 27
2Selection of Artificial Lift Types2
Alternative completion strategies where the pump is placed below theperforations or some forms of gas anchor (see also section 2.4) may beemployed to limit the gas influx. Figure 18(a) illustrates the use of a shroud tomake use of the casings ability to separate produced (free) gas from the liquid.Note that, if it is decided to mount the pump below the perforations {figure18(b)}, then a shroud is also required to initially direct the production flowbelow the electric motor where it provides the necessary cooling. The normalcompletion design of mounting the motor at the bottom of the ESP system,which is then placed above the perforation, provides this cooling automatically.
Pum
p
Gas
Liqu
id
Liqu
id
Liqu
id
Liqu
id
Liqu
id Liqu
id
Gas
Pro
tect
orM
otor
Pum
pP
rote
ctor
Mot
or
Shroud
DownholeSensor
Perforations
Perforations
Gas G
as
(a) (b)
It should be noted that:
(a) Liquid/gas separation in the casing will only work when the upwardvelocity of the gas bubbles is greater than downward fluid velocity, i.e.it will work better for lower rate wells with larger annular clearancesand where the gas is produced as large bubbles e.g. from a separatezone or different perforations to those producing the liquid.
(b) The fitting of a shroud increases the maximum ESP diameter (requiringa large casing for installation of the same motor/pump combination orrequires that a smaller diameter pump/motor be chosen. Practicalexperience has shown that, for a given power requirement, smallerdiameter equipment is often less reliable Further, high power equipmentis not available in the smaller sizes.
Figure 18
Two ESP completion
designs to aid gas
separation in the casing
28
(vi) The Protector or Seal unit connects the drive shaft of the electric motor to thepump or gas separator shaft. It also performs as:
(a) an isolation barrier between the clean motor oil and the well fluids;
(b) an expansion buffer for the motor oil when it reaches operatingtemperature;
(c) equalises internal motor pressure with the well annular pressureand
(d) absorbs any thrust generated by the pump.
(vii) The electric motor is powered by three phase alternating current supplied by thecable connected to the motor at the pothead. They are available in sizesbetween 15 and 900 HP in the manufacturer’s catalogue. Two or even threemotors may be placed in series if high pump power requirements exist.
The motor is filled with oil which insulates the electrical winding. {Ingress ofreservoir fluids (water) is a common cause of motor failure}. A second, lessobvious cause of failure is power surges/voltage spikes/harmonics on thepower supply. These are more prevalent when the power is generated locallyrather than supplied by the (electrical) utility grid. The requirements for the useof a completion design which directs the fluid flow along the motor toprovide sufficient cooling was discussed above in section (v) on pump intakedesign.
When the motor is switched off the head of fluid present in the tubing willreverse the flow direction through the pump as it flows back into the reservoir.This will cause the motor to spin backwards. Trying to restart the motor whileit is rotating backwards will lead to the motor burning out very quickly. Thiscan be avoided by
(a) installing a check valve in the tubing to prevent fluid backflow(however this results in a “wet string” when the tubing/ESP isrecovered
(b) electronically preventing motor restart for a specified time after ithas been shut down or
(c) using a sensor to detect backspin and preventing motor restart(see the next section on downhole sensors).
(viii) A downhole sensor package may be mounted underneath the motor. Measurements can include:
(a) Pump suction and discharge pressures and temperatures.
(b) Fluid intake temperature.
Department of Petroleum Engineering, Heriot-Watt University 29
2Selection of Artificial Lift Types2
(c) Electric motor temperature.
(d) Vibration.
(e) Current leakage.
A downhole flow meter and/or phase cut can be added to the above and all the abovedata transmitted to surface via the power cable. The above can be combined withmeasurement of the power supply frequency and surface current/voltage as well aswellhead temperature, pressure and surface flow rate so as to be able to present acomplete picture of well performance. The data can be:
(a) stored at the well site and downloaded to a (hand held) data logat regular intervals for later analysis
(b) used to trigger on-site alarms which shut the ESP unit down e.g. if thepump suction pressure falls below a preset value indicating that thefluid level in the well is reducing and the well is being “pumped off”
(c) transmitted continuously to the operations office where moresophisticated monitoring analysis can be carried out
(d) replace non-routine well surveillance operations e.g. the sensors may besufficiently accurate to obviate the need for running memory gaugesinto the well when preforming flowing bottom hole pressure surveys,build up tests or reservoir pressure monitoring.
Such sensor packages are now being applied to other forms of artificial lift e.g.directly to EPCPs (see Section 2.9.3) but also to PCP (Section 2.9) rod pumps(Section 2.6) and gas lift (Chapter 3), where similar measurements will enablecorrective action to be taken to maximise the efficiency of the lift operation.
At the beginning of this section we stated that Figure 17 was for a low pressure wellsince a packer had not been installed in the well, i.e. the well will probably not flowwithout artificial lift. Inclusion of a packer in the completion design, as is oftenrequired by the regulatory authorities in live and many offshore wells, precludesventing the gas to the surface via the casing/tubing annulus unless a dual packerarrangement is employed with a safety valve installed on at least the main productiontubing and (possibly) on the gas vent line as well. One example of a possiblecompletion design is illustrated in Figure 19.
30
Gas Vent
Penetrator
Wellhead
1/4" HydraulicControl Lines
Surface Controlled,Subsurface
Safety Valves
Dual Packer
Tubing
9 5/8" CasingShoe
Cable, Banded to Tubing
Pump
Seal Section
Motor
Sensor Package
Gravel Pack Screen
Gravel Pack
Production Casing
Pump Intake
Sliding Circulation Sleeve for Well Killing
Wellhead Penetrations
7.2. Well Completion Employing Electric Submersible Pumps (ESPs)
7.2.1. Typical ESP ApplicationsThe ESP application illustrated in figure 17 is the standard application where it is usedas to lift production from a single zone through a single tubing. Many other applicationconfigurations are possible. Examples are given in Figure 20:
(i) Figure 20 (a) shows aquifer water being lifted from supply zone and pumpeddirectly to an injection well.
(ii) Figure 20 (b) illustrates a dump flood powered by an ESP where the watersupply well and injection well are combined into one. Note that the ESP isinverted with the pump at the bottom. The ESP is being used here to replacethe conventional surface mounted transfer pump. Measurement of theinjection flow rate and pressure can be made by inclusion of a sensor package(not shown) in the well completion design.
Figure 19
ESP completion
incorporating packer and
surface controlled sub
surface safety valve
Department of Petroleum Engineering, Heriot-Watt University 31
2Selection of Artificial Lift Types2
(iii) Figure 20 (c) shows an ESP placed in a shallow well being used to boostpressure in a surface flow line (note the shroud installed to ensure adequatemotor cooling).
Cable
Aquifer SupplyZone
InjectionZone
Aquifer SupplyZone
InjectionZone
Shroud
Pum
pP
rote
ctor
Mot
or
Mot
orP
ump
Pro
tect
or
Motor PumpProtector
Pum
pP
rote
ctor
Mot
or
Low PressureFlow Line
Low Pressure Line
High Pressure Line
High PressureFlow Line
(a)
(b)
(c)
(d)
The pump and motor may also be mounted at the surface - either horizontally (alongthe ground) or placed vertically (reduces required platform area for offshore application).In these cases a conventional, air cooled electric motor is used. Pressure increases ofup to 3000 psi at 7000 bfpd flow rates have been achieved. Higher volumeapplications can be catered for by manifolding a series of pump units in parallel.
7.2.2. Horizontal WellsThe ability of ESP to pump large volumes of produced fluid coupled with theflexibility of pump design and operation makes them very suitable for the largevolume production associated with horizontal wells. Experience has shown that thepump can be placed anywhere within the well at angles up to 80° providing the doglegseverity is not too great (<6º/100ft). Placing the pump near the bottom of the well notonly maximises the potential drawdown which can be created at the formation whilethe (near) horizontal section will enhance the separation of the gas to the upper portionof the wellbore due to its lower density.
Figure 20
ESP applications
(a) Direct water injection
(b) Powered dumpflood
with ESP
(c) Pressure boosting
surface pipelines with ESP
(d) Horizontally mounted
ESP surface pump
32
7.2.3. “Y” ToolThe “Y” tool is a device to allow wireline or coiled tubing access below the ESP. Itis illustrated in Figure 21. The bypass tubing should be at least 2.375” OD (allowing1.6875"/16" logging tools to pass), but 2.875” OD tubing is preferable since thisallows the larger sizes of coiled tubing to pass. However, the larger diameter tubingdoes reduce the maximum diameter of ESP that can be installed. Larger diametermotors are more efficient, tend to have longer run lives and are shorter for a givenpower requirement.
Wellhead Penetrator
Power Cable
Centraliser
TRSSSV Control Lines
Packer Penetrators
PumpPump Intake
Production Casing
Seal / Protector
Motor
(Optional) 1/4" ChemicalInjection Tubing
Production Liner
Downhole Sensor Package
Packer
Packer
Tubing Retrievable Subsurface Safety Valve (TRSSV)
Producing Formation
Wireline / CoiledTubing Retrievable
Plug in Nipple
Bypass Tubing
Fishing Neck for Plug
Installation of “Y” tools allows all the normal wireline and coiled tubing conveyedoperations to be carried out below the ESP. These include:
(i) cased hole logging
(ii) well stimulation
(iii) perforating
(iv) setting bridge plugs for water shut off
(v) installation and recovery of pressure memory gauges
(vi) running and retrieval of plugs
(vii) downhole sampling.
Omission of the “Y” tool from the downhole completion design implies that theseoperations are only possible when tubing and ESP are recovered to the surface.
Figure 21
The "Y"tool"
Department of Petroleum Engineering, Heriot-Watt University 33
2Selection of Artificial Lift Types2
7.3. Basic Pump SelectionThe pressure increase that the pump is required to deliver, also called the “TotalDynamic Head (TDH)” or difference between the pump discharge and suctionpressure, is the sum of three components. Figure 22 pictures the various components:
(i) Friction loss of production tubing (∆Pfric)
(Minimum) Total Dynamic Head (TDH) to be supplied by pump is the sum of (i), (ii) and (iii) (Assuming flowing wellbore pressure at pump inlet is zero)
(ii) Hydrostatic head due to fluid column from ESP depth to surface (p * g * h)
(iii) Surface flow lineback pressure Psurf
Multi Well Manifold
Separator
(i) The hydrostatic head from the ESP pump to the surface. This is equal to the(average) density of the produced fluid in the tubing (ρ) multiplied by the TrueVertical Depth at which the ESP is installed (h) and the acceleration due togravity (g).
(ii) Friction pressure loss in the tubing (∆Pfric
)
(iii) The surface pressure (Psurf
) required to overcome flowline back pressure andflow the produced fluid to the separator at the required production rate. Thiscan have a high value if the completion is a satellite well situated some distance(up to 50 miles) from the host platform.
Thus: TDH = p*g*h + ∆Pfric
+ Psurf
Assuming the flowing wellbore pressure at the pump inlet is essentially zero i.e. wellis "pumped off" and is producing with a maximum drawdown.
Figure 22
Pump duty requirements
34
The data describing the performance (see Section 2.7.A) of ESP’s provided by themanufacture is normally measured with water. They also supply a correction factor,based on the actual density and viscosity, when other fluids are being pumped. Furthercorrection is required if significant volumes of free gas are being injected by the pump- not only will its volume decrease as the pressure increases, but it may also dissolvecompletely in the oil. One popular application area of ESP’s is the production ofviscous crude oils at high water cuts. The design process can be simplified here sincethe density of the crude oil is similar to that of water, there is little gas and the producedfluid stream has an external water phase, i.e. the manufacturer’s performance curvesbased on pumping water can only be applied directly.
Once the pump has been chosen, optimum motor and seal section can be identified,along with the electric cable, variable speed drive, etc. It also needs to be checked thatthe chosen combination will operate efficiently for a variety of well conditions(higher/lower well PI, greater water cut, lower reservoir pressure etc).
The choice of correct ESP design, along with the actual, operational installation ofESP’s, is thus a complex task. This results in the average run lifetime, or “mean timebefore failure (MTBF)” often being very low initially when EPS's are first introducedinto a field / producing area (Especially when experienced staff are not available). TheMTBF then increases as the “learning curve” is climbed. This is illustrated in Figure23 prepared from data presented at IIR’s conference on Artificial Lift Equipment,Dubai, 1997, and IBC’s Artificial Lift Workshop, Aberdeen, 1997. This figure showshow the THUMS project at Long Beach, California, has been employing ESP’s since1965 and how the average run lifetime gradually increased throughout the 16 yearhistory. During this period, the numbers of operational ESP’s remained constant atabout 600. ESP’s were only introduced into the North Sea in the late 1980’s. Theinitial run times were low compared to those achieved at THUMS, but a steep learningcurve developed and by 1996 the average run lives had become similar.
MTBF North Sea, Days
MT
BF
Day
s
YEAR
MTBF at THUMS, Days
Long term trends
1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 19970
100
300
200
500
400
600
700
900
800
1000
1100
1200
Thums
Nor
th S
ea
Figure 23
Average electric
submersible pump lifetimes
Department of Petroleum Engineering, Heriot-Watt University 35
2Selection of Artificial Lift Types2
It need to be mentioned here that average statistics need to be treated with caution -the actual run times will be very dependent on the aggressiveness of the producingconditions e.g. temperature, concentration of sand produced, corrosivity of theproduced fluid (H
2S, CO
2, etc), skill of the rig crew, manufacturing quality control etc.
This is illustrated by a more detailed study of the THUMS data:
(i) Oil is produced from three horizons where the average run lifetimes are 300,700 and 800 days - indicating the importance of the role played by the actualproducing conditions.
(ii) 10% of the ESP’s fail within 30 days and 32% within 180 days, confirming thepotential for damage during installation and the need for the highest manufacturingquality control standards.
7.4. Advantages and Disadvantages of Electric Submersible PumpsTables 5 and 6 respectively (Advantages and Disadvantages of ESP’s) have been puttogether based on the discussion in Sections 2.7.1 and 2.7.2. The points discussed inthese tables should be self-explanatory when read in conjunction with previoussections.
Can be installed in deviated wells (<80…)
High production rates
Suitable for high water cut wells
Controllable production rate
Efficient Energy usage (>50% possible)
Access below ESP via "Y" tool
Comprehensive downhole measurements available
Can pump against high Flowing-Tubing Head Pressure
No extra flow lines required
Minimum surface footprint - 6ft well spacing
Low surface profile for Urban and offshore environments
Quick restart after shut down
Concurrent drilling and production safer compared to gas lift
(high pressure gas not present in annulus)
Long run pump life possible
Advantages of Electric Submersible Pumps
Disadvantages of Electric Submersible Pumps
Susceptible to damage during completion installation
Tubing has to be pulled to replace pump
Not suitable for low volume wells (<150bpd)
Pump susceptible to damage by produced solids (sand / scale / asphaltene)
High GOR’s presents gas handling problems
Power cable requires penetration of well head and packer integrity
Viscous crude reduces pump efficiency
(Viscous) emulsions form over a range of water / oil ratios
High temperatures can degrade the electrical motors
Table 5
Advantages of electric
submersible pumps
Table 6
Disadvantages of electric
submersible pumps
36
7.5. Monitoring the Performance of Electric Submersible PumpsPrior to the implementation of automated SCADA (Supervisory Control and DataAcquisition) systems, the monitoring of ESP performance was limited to a surfacemeasurement of the current supplied to the pump along with an infrequent (possiblymonthly) well test. However, considerable information can be derived on wellperformance - Figure 24(a) is a schematic example of a 24 hour chart recording ofelectrical current consumed by an electric submersible pump during normal operation.The current taken is very constant.
MID
NIG
HT
1AM
2AM
3AM
4AM
5AM6AM7AM
8AM
9AM
10A
M11
AM
MID
DAY
1PM
2PM
3PM
4PM5PM 6PM
7PM
8PM
9PM
10PM
1
70
60
5040
3020
10
70
60
50
40
30
2010
70
60
5040
3020 10
70
60 50 40 30 20 10
70
60
5040
3020
10
70
60
50
40
30
2010
70
60
5040
302010
70
60
50
40
302010
RETEMMAGNIDROCER
DATE ON Day one 7.30
Day two 7.30DATE ON
TIME
TIME
AM
AM
By contrast, the current taken when the well is being pumped off shows a much moreerratic behaviour (Figure 24b).
Figure 24(a)
Ammeter chart monitors
electric submersible pump
performance. Normal
operation.
Department of Petroleum Engineering, Heriot-Watt University 37
2Selection of Artificial Lift Types2
MID
NIG
HT
1AM
2AM
3AM
4AM
5AM6AM7AM
8AM
9AM
10A
M11
AM
MID
DA
Y1P
M
2PM
3PM
4PM5PM 6PM
7PM
8PM
9PM
10PM
1
70
60
5040
3020
10
70
60
50
40
30
2010
70
60
5040
3020 10
70
60 50 40 30 20 10
70
60
5040
3020
10
70
60
50
40
30
2010
70
60
5040
302010
70
60
50
40
302010
RETEMMAGNIDROCER
DATE ON Day one 7.30
Day two 7.30DATE ON
TIME
TIME
AM
AM
The chart was installed at 07.30 am and the pump started at 08.15 am - note the large,initial surge in current while the motor is getting “up to speed”. A steady current isthen drawn for the next 3 hours - decreasing slightly as the fluid head above the pumpdecreases. At 11.10 the current begins to oscillate rapidly - the size of theseoscillations increases until 1.15 pm when the pump was shut down. It was suspectedthat the problem was due to gas being formed when the flowing bottom hole pressurewas reduced to below the bubble point, leading to gas locking and the pump ceasingto pump. This was confirmed by leaving the fluid level in the well to build up for 100minutes and restarting the pump at 3.05 pm. The same cycle repeats itself, howeverthis time the problems appear after some 2.5 hours steady production. The pump wasshut down a second time at 6.15 pm. A third cycle was started at 8.20 pm after a second100 minute shut in - current oscillation starting again after 2 hours production. Thewell was shut in just after midnight.
The basic problem is that the pump is pumping faster than fluid is flowing into the well.Continual stopping and restarting the ESP motor is not recommended due to excessivewear and tear followed by motor burnout. The options are to:
• Install a lower capacity (smaller) pump section.
• Operate the pump at a lower speed.
• Stimulate the well to improve the inflow.
Figure 24(b)
Ammeter chart monitors
electric submersible pump
performance. Well pumped
off.
38
Modern SCADA systems allow a much more complete picture to be built up.Installation of a downhole monitoring package (as shown in Figure 17) allow themotor/pump conditions to be monitored closely. Figure 25 is an example of a normalstart up. It has been analysed as follows:
Motor Temperature
Vibration Current Leakage to Earth
Pump Suction Pressure
Fluid Temperature at Pump Intake
Pump Discharge Pressure
Normal Steady OperationSurface Choke Adjustment
Pump ShutDown
PumpStart Up
Time
A B
• Initially, prior to energising the pump, the pump intake and discharge pressureshave the same value; as does the motor and pump intake pressures. The pumpstarts up at point A, as shown by:
(i) the pump discharge pressure increasing,
(ii) the motor temperature becoming warmer than the fluid entering thepump,
(iii) a limited amount of vibration,
• There then follows a period of surface choke adjustment, as shown byfluctuations in the pump discharge pressure and increased vibration. However,after point B, steady operating conditions are achieved and a slow decline inpump suction and discharge pressure are observed as the well is pumped “off”.
Protection of the ESP can now be achieved by monitoring the pump’s condition andshutting it down when problems develop before physical damage to the pump results.Thus “pump off” control can be implemented by stopping the pump when the intakepressure drops below a preset value. The pump is then restarted once the well pressurebuilds up to a second, higher predetermined value. This type of monitoring is verydiagnostic when problems develop, but can have much more “added value” whencombined with surface flow measurements (gross flow rates and water cut). However,continually restarting pump motors reduces their operational life. Installing a correctlysized pump unit is the preferred solution.
Figure 25
Electric submersible pump
and motor condition
monitoring
Department of Petroleum Engineering, Heriot-Watt University 39
2Selection of Artificial Lift Types2
7.6. New Technology
7.6.1. Coiled Tubing Deployed ESP’sThe completion designs discussed so far within this section on ESP’s all employ thepump installed as part of a conventional completion string with the power cableattached to the outside of the tubing. Replacement of any part of the ESP followinga failure requires a workover. The process of ESP installation and recovery can bespeeded up and made more efficient by installing the ESP at the end of a coiled tubing(see Figure 26). The set up is “conventional” in the sense that the cable is mountedon the outside of the coiled tubing while the produced fluids flow to the surface viathe inside of the coiled tubing. A dual packer arrangement is required with ESParrangement shown. The produced fluid is pumped into the casing via the annulus,flows passed the seal and electric motor sections (cooling) and then back into thecoiled tubing via a cross-over mounted below the upper packer.
Power Cable(Mounted on Reel)
Coil Tubing Reel
Power Cable Clampedto Coiled Tubing
(Cable Cannot Support it’s Own Weight)
Upper Packer
Production Via Coiled Tubing
Crossover
Sensor Package
Pump Intake
Fluid Inflow
Motor
Seal
Pump Discharge
Lower Packer
Replacement of an ESP producing a shallow, depleted horizon in a land well requiresa light workover hoist. A coiled tubing deployed ESP can speed up the process, butthe advantages become much greater when access to the well site is limited e.g. whenfor an offshore well located in a small platform. The (limited) weight requirementsof a coiled tubing package often allow its installation on the platform using theplatform crane; while making a conventional (jackup) rig available is a much moretime consuming, expensive operation.
An alternative, simpler arrangement is illustrated in figure 27 in which the ESP cableis installed within the coiled tubing and the production travels to surface via theannulus. This arrangement has the advantages that:
Figure 26
Schematic view showing
installation of a coiled
tubing deployed ESP
40
(i) Reduced frictional pressure losses lead to higher flow rate or reduced powerrequirements.
(ii) Faster running can be achieved with the cable inside the protected environmentof the coiled tubing.
(iii) It opens up the possibility of installing the ESP in a live well (well killing is a major source of production loss due to formation impairment).
Coiled Tubing
Internal Power Cable
Shear Sub
Electrical Penetrator
Electric Motor
Seal or Protector
Pump Discharge(Production to SurfaceVia Annulus)
Pump
Pump Intake Stinger
Packer
Production Casing / lineror Production Tubing
Fluid Flow
The disadvantage to the system is that the production takes place via the ProductionCasing/Coiled Tubing Annulus, which raises a number of safety issues concerningbarrier policy and corrosion. Alternatively, a wider diameter well can be drilled anda (large diameter) production tubing installed.
7.6.2. Auto “Y” ToolThe advantage of installing a “Y” tool to allow access below the pump was describedearlier. However, wireline/coiled tubing recovery and replacement of the plug in thebypass tube are required each time the conventional “Y” tool is used. The cost and riskassociated with these two wireline operations can be avoided by use of the Auto “Y”tool developed by Phoenix Petroleum Services of Aberdeen. It’s operation isillustrated by figure 28:
Figure 27
Alternative design for
coiled tubing deployed ESP
featuring annular
production
Department of Petroleum Engineering, Heriot-Watt University 41
2Selection of Artificial Lift Types2
(a) A spring holds the diverter plate across the pump leg when the ESP is switchedoff - full access is now provided to the bypass.
(b) Flow generated by the ESP starting up moves the diverter plate and opens thepump leg.
(c) The diverter plate is seated onto the bypass leg by the high pressure in theproduction tubing during normal ESP operation.
(d) The diverter plate can be held in the mid position if it is desired to operate theESP during a logging run.
Flow
HighPressure
Coiled Tubingor wireline
FlowFlow
Lower pressurein bypass leg
Byp
ass
Leg
Pum
pLe
g
(a) Diverter blockspump leg whenESP shut downfull accessto bypass
(b) Diverter movesacross due tofluid circulationvia bypass when ESP starts up
(c) Diverter continues toclose bypass leg when pump operational due to greater pressure in production tubing
(d) Diverter held openduring CT / wirelinelogging run with pumpoperating
7.6.3. Dual Pump InstallationsMore than one Electric Submersible Pump can be installed for a number of reasons:
(i) greater power installed downhole than can be (economically) achieved with asingle ESP. The ESPs are placed such that the discharge of the lower ESP formsthe suction of the upper ESP.
(ii) Dual Zone Completion. Figure 29 shows a dual completion with each zonehaving its own ESP and production tubing. The production tubing for the upperzone is installed concentrically within that for the Lower Zone.
Figure 28
Auto "Y" tool removes need
for pulling / rerunning
plugs
42
Polished Bore
Sensor Package
Sensor Package
Pump Discharge
Pump Discharge
Crossover
Dual Cable Clamp
Cable penetrator
Lower ESP
Stinger
Production Tubing
Production Tubing
Production Tubing
Production Tubing(Lower Zone)
Production Tubing(Upper Zone)
Pump Support
Bypass Tubing
Install lowerzone downholeflow meter here
Install lowerzone downholeflow meter here
Packer
Casing or Liner
Lower Production Zone
Upper Production Zone
Upper ESP
Perforations
(iii) the completion could have been simplified by use of a single production tubingand allocating production between the Upper and Lower zones by installingdownhole flow meters above the pumps (see figure 29).
(iv) remote, offshore locations have very high cost associated with the frequentreplacement of ESPs if the run lives are short e.g. due to excessive sandproduction. The concentration of sand in the produced fluid is often particularlyhigh after a workover. Completion designs have been developed with a“sacrificial” lower ESP - see figure 30. The operational concept is as follows:
Figure 29
Dual completion with
separate ESP's for each
production string. ESP's
discharge into dual
concentric production
tubing
Department of Petroleum Engineering, Heriot-Watt University 43
2Selection of Artificial Lift Types2
Lower ESP
Upper ESP
Sensor Package
Sensor Package
Auto "Y" Tool
Dual Cable Clamp
Auto "Y" Tool
ESP Support
(a) The lower ESP is started and the well produced until the sand productiondecreases to an acceptable level. The upper ESP may now be started to furtherincrease the well production. The upper ESP can be operated independently ofthe lower ESP in the case that the lower unit fails due to erosion by producedsand or other reasons.
(b) The auto “Y” tool allows production to be switched between the two ESPswithout wireline or coiled tubing intervention. However, it must be checkedthat fluid is not being circulated around the pump, since this will lead to rapidpump failure (motor burnout due to overheating).
7.6.4. Reducing Water ProductionThe level of water production is an ever increasing problem as reservoirs mature.
Figure 30
Dual pump installation with
"sacrificial" lower ESP
44
Hydrocyclones (see section 9.25) have become the preferred technique to separate theproduced oil and water. Their dimensions and lack of moving parts make them highlysuitable for installing downhole (see also chapter 10). Figure 31 illustrates oneequipment design. This shows a single electric motor powering an upper and lowerpump unit. The equipment is installed below the producing zone. The lower pumpunit supplies sufficient power to operate the hydrocyclone and to inject the underflow(water containing approximately 100ppm oil) into the water injection zone. Thehydrocyclone overflow (typically 50% oil) is transferred via a bypass tube to the upperpump which pumps it to surface. The maximum production rate depends on the casingsize and the number of hydrocyclones installed in parallel (maximum capacity of ahydrocyclone is 2,500 bfpd). Rates up to 20,000 bfd for a 9.675 in. casing are feasible.
Oil Concentrate to Surface (50% Water)
Lower Pump Unit
Hydrocyclone Separator
Water (± 1000 ppm oil)
Electric Motor
Oil
Con
cent
rate
Byp
ass
Pump Inlet
Producing Zone(Water Cut > 85%)
Water InjectionZone
Upper Pump Unit
This technology has been extended to meet the challenges of:
(i) Installing two separators in series. This 2 stage separation allows downholewater separation to be started at the lower water cut from the production zoneof 65%. A minimum produced water cut of 80% is required to achieve efficientoperation (acceptably low oil concentration in the rejected water stream) witha single hydrocyclone stage.
(ii) Produced Sand. Any sand particles produced will be separated with the waterflow to be injected due to its greater density. Blockage of the injection zone willoccur rapidly if significant volumes of sand are being produced. This can beavoided by treating the injection water stream with a hydrocyclone designed toconcentrate the produced sand particles in the underflow (see figure 32). This(small) underflow stream is added to the oil concentrate and produced tosurface while the bulk of the water stream is injected as normal.
Figure 31
Downhole water separation
driven by an electric
submersible pump
Department of Petroleum Engineering, Heriot-Watt University 45
2Selection of Artificial Lift Types2
High WaterCut Production
From Production Zone
Lower Pump Unit
Upper Pump Unit
OVERFLOW:Oil Concentrate
Oil Concentrateand Sand Particles
to Surface
To Injection Zone
UNDERFLOW:
Water, ±1000 ppm Oiland Sand Particles
UNDERFLOW:Some Water and
Sand Particles
OVERFLOW:Water, 100 ppm Oil
Hydrocyclone(Oil
Concentrator)
Hydrocyclone(Sand
Rejection)
(iii) Coning Suppression. This concept is illustrated in figure 33. Production fromthe oil zone perforations results in a water cone being formed once the criticaloil production rate has been exceeded. This critical rate can be increased byproducing water from below the oil/water contact via “coning suppressionperforations” and injecting the water into an injection zone using an inverted ESP.
"Normal" Water ConeDue to high water production at "Oil Zone" perforations
Oil production to surface
Water Injection Zone
Electric SubmersiblePump
Supressed Water ConeDue to water production at "Anti Coning" perforations
"Anti Coning"perforations
Oil Zone Perforations
Oil / Water Contact
This concept has been shown to work - but alternative technologies such ashorizontal production wells will often be more attractive economically.
(iv) Managed Water Injection. The water injection zones illustrated above weresimple perforated completions. Advanced well concepts in which the water isinjected into a long (near) horizontal lateral split into a number of zones, wherethe volume of injected water will be regulated by an electrically adjustablechoke, will become feasible in the next few years.
Figure 32
Flow diagram for downhole
separation with sand
rejection hydrocyclone
Figure 33
Suppresed coning and in
situ water disposal
46
7.7. Electric Submersible Pump PerformanceThe centrifugal pump unit employed in ESP’s is a dynamic-displacement pump inwhich the pump rate depends in the pressure head generated - the pump rate is lowwhen the pressure head is high and vice versa. This is different from the positivedisplacement pumps discussed earlier in which the pump rate and discharge pressureare independent of one another.
The relationship between pump rate and pressure generated for dynamic displacementpumps is called the pump characteristic (see Figure 34). It is measured by the pumpmanufacturer in laboratory tests using a standard fluid (water) with the pump runningat 3500 rpm (60 Hz electrical supply) or 2915 rpm (50 Hz supply).
7084
72
60
48
36
24
12
300 600 900 1200 1500 1800 2100 2400 2700
60
.60
.40
.20
50
40
30
20
10
Pump Rate (B/day)Measured at 50 Hz, motor speed of 2915 rpm and a fluid viscosity of 1 cp and density 1g/cm3
Hyd
raul
ic H
orse
Pow
er /
Sta
ge
Pum
p O
nly
Effi
cien
cy
Pum
p H
ead
(ft.
wat
er)
Maximum Efficiency
Pump Head
Pump Efficiency
Motor Power
Recommended PumpOperating Range
The “pump head”, or increase in pressure per stage (∆P), is expressed in terms of thepressure generated by an equivalent column of water (H
water). It decreases as the pump
rate increases:
∆P = ρ*g*h = 0.433*γ* Hwater
The discharge pressure is proportional to the specific gravity (γ) for other liquids withthe same viscosity, i.e. the pump head - pump rate relationship can be used for allliquids, but only requiring correction for changes on viscosity.
Pump power is the workdone per unit time which equals the pump rate multiplied bythe pump head (q*∆P).
Power is normally expressed in terms of Kilowatt (KW) or Horse Power (HP); where1 HP = 0.746 KW. The pump or hydraulic power is the (useful) work done by the pump
Figure 34
A typical pump
characteristic curve for a
centrifugal pump
Department of Petroleum Engineering, Heriot-Watt University 47
2Selection of Artificial Lift Types2
while the mechanical power is the work done by the electrical motor which is requiredto drive the pump. The pump efficiency (E) is thus:
E = hydraulic power/mechanical power
Pump Efficiency is also recorded in the pump characteristic curve and a recommendedpump operating range indicated based on ±10% of the maximum efficiency point.
7.7.1. Simplified Electric Submersible Pump DesignESP design is available as an option in many of the commercially available wellperformance programs, e.g. WellfloTM. The simplified, manual procedure outlinedbelow to evaluate the installation of an ESP into the vertical well Edinburgh-1 followsthe same basic steps as the more complex, computerised, design procedures.
Table 7 summarises the Edinburgh-1 well conditions.
Well Edinburgh 1
Depth (h) 7000ft
Reservoir pressure (Pr) 1700psi
Well productivity Index (PI) 2 STB / day / psi
Tubing Internal Diameter (d) 2.26 in or 0.188ft
Surface manifold Pressure (Pm) 50 psi
Design Well Production (Q) 1400 STB / day
Produced Fluid properties Water
Fluid Density 0.433 psi / ft
Viscosity 1cp
Pump set at same depth as perforations
(i) The pipe friction loss (∆Pf) at the desired well production is given by:
∆P fLd
vgf f= ( ) ( )
* * *ρ2
2
where fm is the moody friction factor, v is the fluid velocity and g the
acceleration due to gravity {32.173 (ft/s2) (lbm/lb
f)}. Now:
vQ STB day ft bbl
s day d ft
ft s
=( ) ( )
( ) ( )( )= =
/ * . /
, / * /
* . *, * . * .
. /
5 615
86 400 4
1400 5 615 486 400 3 14 0 188
3 28
3
2 2
2
π
As discussed previously, the value of fm, a function of Reynolds Number, pipe
roughness and the fluids’ properties, can be found from a Moody Diagram. It hasa value of 0.03 for this calculation.
Table 7
Data for Well Edinburgh 1
48
∆P psif = ( ) ( )
=0 03 0 43370000 188
3 28
2*81
2
. * . *.
*.
32.2
(ii) Pd = P
s + ∆P
f + ∆P
HH
where Pd is the required pump discharge pressure and ∆P
HH is the hydrostatic
head due to the 7000 ft column of fluid. Ps is the wellhead pressure required to
transfer the fluid to the surface facilities (50 psi).
Pd = 50 + 81 + (0.433 psi/ft) * (7000 ft) = 3162 psi
(iii) The Flowing Bottom Hole Pressure and the pump intake pressure (PIn) are the
same and can be calculated from:
PIn
= Pr - Q/PI = 1700 - 1400/2 = 1000 psi
N.B. It is essential that {P - PIN
} > 50 psi to ensure select there is a minimum heightof fluid above the pump section so that it doesn't "run dry".
(iv) Using the pump performance chart shown in Figure 34, the head per stage (H)at 1400 b/d is 58 ft and the hydraulic horsepower per stage (HHP) is 0.52.
The number of pump stages (N) and the minimum electric motor power (HHP)required can now be calculated for a pump running at 2915 rpm.
NP P psi
H ft γ psi ft γstagesd in=
−( ) ( )( ) ( )= −( ) =
* * . / * * .0 4333162 1000
58 0 433 86
and HHP = 86 (stages)*0.52 (HHP/stage)*(γf) = 45 HP
where γf is the specific gravity of the fluid (unity in our case)
(v) An electric motor to power the pump may now be chosen (minimum 205 HPand 50Hz).
N.B. Choosing a pump speed other than 2915 rpm introduces extra complicationssince the pump rate of an ESP is proportional to the speed
i.e.pump rate pump rate
pump speedpump speed
2
1
2
1
=
where (1) denotes the initial rate (2915 rpm) and (2) refers to the new speed ofthe motor (and pump, since ESP’s do not have a gearbox). Further, the motorspeed also controls the hydrostatic head produced.
hydrostatic headhydrostatic head
pump speed pump speed
2
1
2
1
2
=
Department of Petroleum Engineering, Heriot-Watt University 49
2Selection of Artificial Lift Types2
The power required may now be calculated
motor power motor power
pump speedpump speed
2
1
2
1
3
=
Variable Frequency Drive (VFD) provides the ability to change the pump andelectric motor speed by altering the frequency of the electricity supply. The pumpcharacteristic performance curves are also measured by the manufacturer for a rangeof conditions and are reported in their data books - Figure 35 shows the format of atypical example (not the same one as discussed above).
21
18
15
12
9
6
3
15 30 45 60 75 90 105 120 135
Pump Rate (m3/day)
Pum
p H
ead
(m. o
f wat
er)
Operating Range
(Hydraulic Power) / Stage at Maximum Efficiency80 Hertz
70 Hertz
60 Hertz
50 Hertz
40 Hertz
30 Hertz
90 Hertz
3.07
2.16
1.45
.91
.53
.27
.11
(vi) The final stage in this simplified design procedure is to evaluate the robustnessof the design for a series of well inflow conditions i.e. changes in well productivityindex or reservoir pressure. These are performed by carrying out a nodalanalysis on the ESP pump. Figure 36 is a typical example of such an analysis(also not the same one as discussed above).
Figure 35
Typical changes in pump /
motor characteristic
performance as a function
of electric supply frequency
50
0
250
500 750 1000
Well Production Rate (STB / day)
Pre
ssur
e at
ES
P In
let (
psi)
1250 1500 1750 2000
500
750
1000
1250
1500
1700
evruC
ekatnIp
muP
PI = 4 STB/psi
PI = 2 STB/psi
PI = 1.0 STB/psi
PI = 0.5 STB/psi
This figure shows that:
(a) The well fluid level above the pump (1000 psi or 2310 ft TVD) is high if thewell’s PI was 2 STB/d/psi and 1400 STB/day were being produced i.e. the wellis not being “pumped off” and a larger pump could have been installed. Theproduction rises to 1540 STB/d (and the fluid level to 1315 psi or 3035 ftTVD) if the well productivity index increased to 4 STB/d/psi.
(b) The well production reduced to 1190 STB/d for the lower well productivityindex of 1 STB/day. The lowest well PI plotted (0.5 STB/d/psi) results in anegative well inflow pressure. The well has now been “pumped off” - anunacceptable situation which would be corrected by restricting the tubingoutflow with a choke. The minimum well inflow pressure will be dictated bythe minimum pump charging pressure required (depends on pump design), gasinterference e.g. bubble point etc.
(vii) Cable selection - which depends on pump power, voltage selected and downholetemperature, may now be made.
(viii)Further details can be found in API RP 1154 - “Recommended Practice forSizing and Selection of Electric Submersible Pump Installations”.
The ESP manufacturers can supply software to carry out a more sophisticated designanalysis than that described here. Further, many of the well design of nodal analysispackages included data from the pump manufacturers so that the well analysis andselection process can be automated.
Figure 36
Well flow / pump out flow
(nodal) performance curve
Department of Petroleum Engineering, Heriot-Watt University 51
2Selection of Artificial Lift Types2
8. HYDRAULIC PUMPS
Hydraulic pumps use a high pressure power fluid pumped from the surface (Figure 37)which:
Controls power fluid supply rate to each downhole engine.
Manifold Values
To Well, 2, 3 etc.
Commingled Exhaust Power Fluid andProduced Fluid
Power Fluid
Pump Discharge
Produced Fluid
Downhole Engine(Turbine, Positive
Displacement Typeor Venturi)
Pump
Commingled Power Fluidand Produced Fluid to
Separation
Power Fluid(Usually Produced
Water)
DesandingHydrocyclone
PressurisingPump
Sand
Filter
(i) drives a downhole, positive displacement pump. Figure 38 shown how the flowof power fluid through the upper engine unit is translated into a flow of highpressure produced fluid during both the “UP” and “DOWN” strokes.
Figure 37
Principals of hydraulic lift
operation. Turbine pump is
illustrated.
52
Engine Piston
Piston ConnectingRod
Pump Piston
Engine and pump pistonsmove downwards during
Down Stroke
Power Fluid Power Fluid
Engine and pump pistonsmove upwards during
Up Stroke
Engine Exhaust
High PressureProduced Fluid
High PressureProduced Fluid
High PressurePower Fluid
ExhaustPower Fluid
ProducedFluid
Valve Open
Valve Open(Produced Fluid
Intake)
Valve Closed
Valve Closed
Engine Exhaust
Engine Unit
Pump Unit
High PressureProduced Fluid
Valve open
Produced Fluid Produced Fluid
(ii) powers a centrifugal or turbine pump (see Section 2.8.3).
(iii) creates a reduced pressure by passage through a venturi or nozzle (Figure 39)where pressure energy is converted into velocity. This high velocity/lowpressure flow of the power fluid commingles with the production flow in thethroat of the pump. A diffuser then reduces the velocity, increasing the fluidpressure and allowing the combined fluids to flow to surface.
Figure 38
Operation of positive
displacement hydraulic
pump
Department of Petroleum Engineering, Heriot-Watt University 53
2Selection of Artificial Lift Types2
Fishing Neckfor Pump Recovery
Inner Tubing
Spring HoldsNozzle in Place
Produced Fluid
Vented Gas
Produced Fluid
Produced Fluid High pressurePower Fluid
Well Casing
Outer Tubing
Standing Valve(Ball and Seat Valve)
Tubing Packer
Vented Gas
Diffuser
Throat
Nozzle
Venturi
Power FluidSuppliedby Inner
ConcentricTubing
CommingledProduced Fluid and Exhaust Power Fluid
CommingledPower and
Produced Fluid
Vented Gas
The power fluid consists of oil or production water (the large oil inventory in thesurface power fluid system makes oil accounting difficult once high water cuts arebeing produced). The power fluid is supplied to the downhole equipment via aseparate injection tubing. The majority of installations commingle the exhaust fluidwith the production fluid {an “open system” Figure 40(a)}. If difficulties or high costsare encountered in preparing power fluid of the required quality from the productionfluid, then a “closed system” may be installed in which the power fluid returns to thesurface via a (third) separate tubing {Figure 40(b)}. This option is not available witha venturi pump. The completion design may also allow gas to be vented to surface viathe casing/tubing annulus.
Figure 39
Jet or Venturi pump
operation
54
Produced Fluid
High PressurePower Fluid
CommingledPower and
Produced Fluid
High PressureProduced Fluid
Engine
ProducedFluid Level
Pump
StandingValve
Power Fluid
Commingled Produced
andPower Fluid
Power Fluid
Produced Fluid
Gas
Gas Vent to Annulus
Produced FluidProduced Fluid
(a) Open Power Fluid System (b) Closed Power Fluid System
ExhaustPower Fluid
A typical power fluid supply pressure of between 1,500 and 4,000 psi. is provided bya pressurising pump (Figure 37). This may be a reciprocating plunger (triplex) pumpor a multi-stage, centrifugal pump. This pressure determines the pressure increaseachievable by the downhole (positive displacement or centrifugal) pump. The pumprate (and the rate at which power fluid has to be supplied) is determined by the diameterand speed of the downhole pump.
“Clean” power fluid is required to avoid erosion of the downhole pump components.The power fluid is often drawn from a settling tank where the larger solids areremoved. It is then pumped via a desanding hydrocyclone and a guard filter beforehaving its pressure raised to the operating pressure by the charge pump. The powerfluid from the pressurising pump may supply one or more wells (Figure 37).
8.1. Advantages of Hydraulic PumpsHydraulic pumps have the following advantages:
(i) Suitable for crooked and deviated wells.
(ii) Reciprocating and turbine pumps can work at great depths (up to 17,000 ft),while jet pumps are restricted to about half that value.
Figure 40
Types of hydraulic pump
installation
Department of Petroleum Engineering, Heriot-Watt University 55
2Selection of Artificial Lift Types2
(iii) Very flexible speed control by the (surface) supply of power fluid. Turndownto <20% of design maximum speed can be achieved.
(iv) Jet pumps, with no moving parts can handle solids. Weir pumps areclaimed to be manufactured from erosion resistant materials which aim togive a 5 year lifetime with “reasonable” solids production associated witha prepacked screen completion for installation in a soft formation.
(v) The power source is remote from the wellhead giving a low wellhead profile,attractive for offshore and urban locations.
(vi) The power fluid can carry corrosion or other inhibitors downhole, providingcontinuous inhibition when the well is producing.
(vii) The pump unit can be designed as a “free” pump; the pump unit having thecapability of being pumped through the power fluid tubing from the surface toits downhole location (figure 41). It can then be recovered by reversing theflow direction. The ability to recover the pump without the need to move a rig/workover hoist to the wellsite is attractive for offshore platforms as well asremote and urban locations.
Standing Valve Closed
Standing Valve Open
Standing Valve Closed
Install Pump Production
ProductionFlow Line
PowerFluid
PowerFluid
PowerFluid
CommingledPower and
Produced Fluid
Pump Recovery
"Free" Hydraulic Pump
Figure 41
Installation and recovery of
a "free" hydraulic pump
56
8.2. Disadvantages of Hydraulic Pumps(i) Pumps with moving parts have a short run life when supplied with poor quality
(solids containing) power fluid. Jet pumps can have a long run life under similarconditions.
(ii) Positive displacement and centrifugal pumps can achieve very low flowingbottom hole pressures in the absence of a gas effect. The lowest pressureachievable by jet pumps are much higher, being comparable to gas lift.
8.3. New Technology (Weir Pumps)A recent innovation that has been field tested over the last few years is the Weir Pump- a hydraulically driven engine coupled to a turbine pump with improved gas handlingabilities. It uses an open power fluid circuit - the power fluid is returned to the surfacecommingled with the production. Its performance characteristics are very similar tothat of an ESP - but it also has the ability to achieve stable operation over a wide rangeof flow rates with gas fractions of at least 80%. In addition, large gas slugs associatedwith surging flow can be handled without mechanical damage to the pump or its motor.
Hydraulic (and Weir) pumps have several intrinsic advantages over ESP’s:
(i) The continuous supply of cool power fluid increases the maximum allowablebottom hole formation temperature at which the pump can be installed.
(ii) Solids free power fluid lubricates the pump bearings, enhancing their producedsolids handling capabilities (typical power fluid specifications are 100 ppmsolids with a maximum diameter of 0.1 mm). The need (and cost) of continualcleaning of the power fluid in a (multi-well) subsea development could beminimised by use of a “closed” power loop.
(iii) There is no requirement for a mechanical seal to be installed between the motorand the pump, as is the case for an ESP.
(iv) Variation in the power fluid flow rate provides pump speed control as well asa “soft start” capability.
(v) The pump set operates at high speeds. This results in a short (3-4 m) pump unitsince:
Number of pump stages required α {1 / (pump speed)2}
This short length (and weights typically less than 500 kg) allow wireline retrieval ofthe pump unit in wells at deviation angles of up to 55º. Coiled tubing installation isrequired for higher deviation angles (up to 80º).
Figure 42 illustrates the completion design used in the Texaco field trial in their NorthSea “Captain” field. This included:
Department of Petroleum Engineering, Heriot-Watt University 57
2Selection of Artificial Lift Types2
Power Fluid Temperature and Pressure Measurements
Flow Rate (Venturi), and Outlet Temperature and PressureMeasurement
Pump Unit Speed and Vibration Measurement
Inlet Temperature and Pressure Measurement
Packer With Polished Bore
6 5/8" Prepacked Wirewrapped Screens
9 5/8" Casing Shoe
Fluid Loss Control Valve5 1/2" Surface Controlled Subsurface Safety Valve
Produced Fluid
Power Fluid
Actuated Choke
Actuated Choke
2 3/8" By-pass Tube
Lower "Y" Tool
Upper "Y" Tool
10,000 BPD Hydraulic Submersible Pump Assembly
7" Tubing
(i) Installation of an upper and lower “Y” tool connected by a 2.375" by-passtubing. This tubing allows access (e.g. for production logging purposes) to theproducing interval without having to recover the pump. A Surface Controlled,Sub Surface Safety Valve has been installed below the lower Y tool. Also, a“Fluid Loss Control” valve is installed at the bottom of the tubing to preventinjection of workover fluid into the completion during pump recovery or otheroperations being carried out above the packer.
(ii) Comprehensive flow rate, temperature and pressure measurement as well aspump performance monitoring, such as pump speed and vibration measurements,has also be installed.
(iii) As discussed previously, the development of the viscous oil rim in Texaco’s“Captain” field depended on developing a reliable artificial lift system havinga combination of effective gas handling and viscous fluid pumping capabilities.It was concluded after a one year’s field trial that:
(a) The viscous crude oil, and any resulting high viscosity emulsions, waspumped at flow rates ranging from 20% - 120% of design with watercuts varying from 0 - 100%.
(b) Fluid gas fractions of 30% to 75% at the pump suction were routinelypumped while exceptional, slug flow conditions of greater than 90%gas volume fraction, were managed by the fluid driven turbine pump.
(c) The pump and turbine design was robust. It withstood the imposed loadsand was resistant to the typical solids production encountered incompletions requiring sand control.
Figure 42
A hydraulic (Weir) pump
completion with down hole
monitoring
58
9. Progressing Cavity PUMPS
Progressing Cavity (or Moyno) Pumps are becoming increasingly popular for theproduction of viscous crude oils. Figure 10 summarises the application area (wellrates & depths) where Progressing Cavity Pumps (PCP) are typically employed. Atypical completion is illustrated in Figure 43 where a prime mover (in this case anelectric motor) is shown rotating a sucker rod string and driving the PCP. Thissection will describe the principle on which the pump operates, the resultingadvantages and disadvantages and, finally, takes a look at new technology.
WellheadCasing Vent
Tubing
Electric Motor
Belt DrivenSpeed Reducer
Torque Anchor
Coupling connectsrods to drive shaft
Rods
Centraliser
Casing
Progressing Cavity Pump
Production Zone
Figure 43
A well completed with
artificial lift using a
progressing cavity pump
Department of Petroleum Engineering, Heriot-Watt University 59
2Selection of Artificial Lift Types2
9.1. Progressing Cavity (or Moyno) Pump PrincipleFigure 44 illustrates the main components of a PCP. A steel shaft rotor of diameterd has been formed into a helix {Figure 44(a)}. The rotor is rotated inside anelastomeric pump body or stator, which has been molded in the form of a double helixwith a pitch of the same diameter and exactly twice the length of the pitch given to therotor {Figure 44(b)}. Figure 44(c) shows that, when assembled, the centre line of therotor and the stator are slightly offset, creating a series if fluid filled cavities along thelength of the pump. Figure 45 is a perspective view of Figure 44(c), which helpsexplain how the interference fit between the rotor and stator creates two chains ofspiral (fluid filled) cavities.
(a) Assembled Pump (b) Rotor Geometry (c) Stator Geometry
Centreline of Stator
Centreline ofStator
Rotor contactsstator herecreating a
sealed cavity
Elastomeric Stator
Centreline of Rotor
Centreline ofRotor
Pitc
h R
otor
Pitc
h S
tato
r
Steel Rotor
ecc d+2ecc
d
d+4ecc
FluidFilledCavities
d is minor diameter of rotor and slator, ecc is rotor eccentricity
Figure 44
Cross section progressing
cavity pump and its
components.
60
d
d+4ecc
Steel Rotor
Rotor contacts stator here, creating sealed cavity
ecc
Pitc
h of
Sta
tor
Pitc
h of
Rot
or
Elastometric Stator
Fluid Filled Cavities
The rotor within the stator operates as a pump. This causes the fluid, trapped in thesealed cavities, to progress along the length of the pump from the suction to the pumpdischarge. These cavities change neither size nor shape during this progression.Figure 46 (a-e) shows how, as one cavity diminishes, the next one increases at exactlythe same rate; giving a constant, non-pulsating flow. It acts as a positive displacementpump. The pressure increase that can be achieved by the pump depends on the numberof “seal-lines” formed along the pump body by the rotor and stator. Typically, this isfound to be 300-200 kPa pressure increase per stage. It is found that fluid will “slip”backwards if a greater pressure increase is demanded from the pump. This can beavoided by increasing the number of pump stages. Wear of either the stator of rotorwill decrease this value since the the maximum pressure increase depends on thisinterference fit. However, the construction of the stator body from an elastomer makesthis pump design relatively tolerant to produced solids - particularly since they areoften used to pump viscous oils which provides a lubrication film to protect the rotorand stator from wear.
Figure 45
Perspective view of
operating Progressing
Cavity Pump
Department of Petroleum Engineering, Heriot-Watt University 61
2Selection of Artificial Lift Types2
Rotor at top side of cavity
Centre Line Cross Section atCentre Line
90… Rotation
Stator Body
270… Rotation
180… RotationRotor is at bottom of cavity
360… Rotation
Rotor returnedto topside of cavity
The flow rate achieved by a liquid filled (no gas) PCP pump is directly proportionalto the speed of rotation of the rotor (N):
Flow rate = k*Pitch of Stator*4*eccentricity*Stator minor diameter*N
Where k is a constant.
The presence of gas reduces the efficiency and a gas anchor is frequently included inPCP completions (see also Section 2.6.5.2 for discussion of gas anchors). Figure 14illustrates one of the many available forms of gas anchors that can be used.
The advantages and disadvantages of a PCP are summarised in Tables 8 and 9. It canbe seen that the pump’s characteristics make it very suitable for artificial lifting wellsproducing medium to high viscosity crude oil reserves. These crude oils often havea tendency to form highly viscous emulsions when mixed under high shear with theproduced water (as occurs in a centrifugal pump). They are often found in (relatively)shallow, young (geologically speaking), soft formations where the inclusion of sandcontrol in the completion design is a necessity. Finally, elastomer selection problemsare minimised because these crude oils tend to have a low GOR as well as a lowaromatic content and their shallow location results in a cool Bottom Hole Temperature.
Figure 46
Operating principle of
Progressing Cavity Pump
62
Advantages of PCP Comment
Simple design Quick pump unit repaired by replacing
rotor and stator as a complete unit
High volumetric efficiency In the absence of gas
Efficient design for gas Tolerant of produced solids at reasonable
anchors available levels
High energy efficiency PCP is a Positive Displacement Pump
Emulsions not formed due to ESPs and Weir pumps promote emulsion
low shear pumping action formation due to high pump speeds
Capable of pumping viscous (1) Diluent mixed as required with crude
crude oils oil if extreme viscosities to be pumped
(2) "Water-like" behaviour observed at high
water cuts when oil becomes the
internal phase
Disadvantages of PCP Comment
High Starting Torque
Fluid compatibility problems with Carry out tests prior to producing a new
elastomers in direct contact with crude oil
aromatic crude oils
Gas dissolves in the elastomers, Avoid rapid depressurisation of the pump
at high bottom hole pressure destructive bubbles formed when
pressure is lowered rapidly
9.2. Progressing Cavity Pump Power SupplyTraditionally, PCP’s have been powered by an electric motor and gearbox mountedabove the wellhead and turning a string of sucker rods connected to the PCP pump;i.e. the rods are rotated rather than reciprocated (Figure 43). As discussed in section2.6 on Rod Pumps, this string of sucker rods is susceptible to failure - especially incrooked or deviated wells or when formation sand is being produced. There is asimilar tendency to a higher frequency of tubing failures, since the rods are rotatinginside the tubing. Installation of centralisers on the sucker rod string can mitigate thisproblem (Figure 15). This rod/tubing frictional contact, even when reduced bycentralising the sucker rod string, leads to a large loss of starting torque as well aswastage of power when the pump is operating.
Further, the tubing has to be pulled and then rerun when the pump unit requires repair.This can normally be done by a light workover hoist, since the wells are not normallycapable of natural flow.
Table 8
Advantages of a PCP
Table 9
Disadvantages of a PCP
Department of Petroleum Engineering, Heriot-Watt University 63
2Selection of Artificial Lift Types2
9.3. New TechnologyThe resulting high torque and friction losses, as well as the tubing and rod failurediscussed above, can be reduced by placing the motor downhole - this is known as aProgressing Cavity Electric Submersible Pump. Secondly, low cost replacement ofthe PCP unit can be achieved by making it wireline retrievable. These newdevelopments will be discussed in the following two sections.
9.3.1. The Progressing Cavity Electric Submersible Pump (PCESP)The PCESP share the same electric motor, seal, cable and control technology as theconventional Electric Submersible Pumps (ESP) discussed in Section 2.7. The majordifference is that a gearbox is required to reduce the speed of rotation since thecentrifugal pump employed with a conventional ESP is a high-speed device; while aPCP is a low speed device. The layout of the two pump types is compared in Figure47. Some typical dimensions have been included to illustrate the length of thecomplete PCESP unit. In the example shown, which can be run inside a 7" casing, itranges from 34 ft for a low power unit to 81 ft for a medium power device.
Tubing
Cable Cable
Centrifugal Pump
Pump Intake orGas Separator
Seal or Protector
Electric Motor
ProgressingCavityPump
Gearbox
*9 ft.- 40 ft.
*10 HP - 5 ft. 60 HP- 21 ft.
9 ft.
11 ft.
Electric Submersible Pump (ESP)Progressing Cavity
Electric Submersible Pump (PCESP)
*Depends on power required
Figure 47
Comparison between layout
of an ESP and a PCESP
64
Practical experience has shown that the expected gains in reduced pump powerrequirements and tubing failures are achieved on changing from a rod driven PCP toa PCESP. Electric submersible motors have become very reliable - providing properdesign and operational procedures are in place. The most frequently replaced item inmany PCESP projects is the PCP unit itself. These costs could be reduced byredesigning the wellhead, completion and the PCP unit so that it becomes retrievableby wireline. This is discussed in the next section.
9.3.2. Wireline Retrievable PCESPThe wellhead and the tubing are enlarged so that they are both large enough to passthe PCP motor and stator unit, which is modified as follows (Figure 48):
4 1/2" Tubing(for 7" Casing)
2 7/8" Tubing
Cable
Wireline RetievableProgressing Cavity
Pump
Progressing CavityPump
Pump Pack Off
Seal Unit
Electric Motor
Gearbox
Fishing Neck
Seating nippleon Top of Seal Unit
Re-latchable ShaftConnection
Drive Shaft
Pump Body
Rotor
Stator
(i) A (re)latchable drive shaft connection is made at the bottom of the unit so thatit can be connected and disconnected to the motor unit and gearbox.
(ii) The PCP Unit is seated on a nipple placed on top of the seal unit.
(iii) The PCP unit body is sealed against a pump pack-off element.
(iv) A fishing neck is provided for wireline recovery.
Figure 48
Modification of a PCESP to
allow wireline retrieval
Department of Petroleum Engineering, Heriot-Watt University 65
2Selection of Artificial Lift Types2
The motor, gearbox and seal unit are all installed as per the conventional PCESPdesign at the bottom of the production tubing.
Extensive testing of this concept at the Thums field, Long Beach, California, providedthe operational data summarised in Table 10. This shows that, despite the considerableincrease in capital cost, the much reduced pump charge out costs resulted in a verysignificant saving in production lifting costs. (This cost saving is very dependent onthe MTBF or “Mean Time Before Failure” of the pump and other downholecomponents). This is a good example of the concept discussed earlier (Section 2.5.4)when extra up-front capital expenditure can result in reduced “Total cost of ownership”through more efficient operations. Also, that these decisions can only be made whena comprehensive cost database is available.
Economics of wire line retrievable PCESP*
Initial Installation Cost Pump Change Cost Savings on "total cost of ownership"
180% of ESP 18% of ESP
125% of ESPCP 13% of ESPCP 15%
*Data provided by Thomas Lutz to the conference on Artificial Lift Equipment '97,
Dubai, December 1997
10. HYBRID SYSTEMS
The combination of two forms of artificial lift has been used in a few cases. Mosthybrid systems combine gas lift with one of the other types of artificial lift. Forexample it has been shown that injection of gas above an ESP or PCP can reduce thehydrostatic head against which the positive displacement pump (e.g. ESP, PCP etc.)is pumping against by upto 40%. This can lead to a significant increase in pumpperformance. It may also allow the well to continue to produce even when the pumpis experiencing mechanical problems e.g. ESPs are susceptible to high levels of sandproduction and may show frequent failure. Gas lift is tolerant of sand production andwill often be capable of drawing the well down sufficiently that production continues,although at a lower rate compared to when the pump is operational.
• Increased volumetric efficiency - higher liquid volumes.• Decreased injection gas requirements compared to gas lift alone.• Increased reservoir drawdawn and production.• Increase pump installation depth - allows greater reservoir drawdown.• Reduction in pump and motor power requirements.• Lower electrical energy consumption compared to pump alone.• Reduces electrical conduit requirements.• Gas lift provides backup in case of pump failure.
Table 10
Economics of wireline
retrievable PCESP
Table 11
Benefits of combining gas
lift with a positive
displacement artificial lift
method e.g. ESP, PCP
sucker rod etc.
66
11. ARTIFICIAL LIFT METHODS TUTORIAL
Question 1.
Hardly any artificial lift equipment was installed during the first 10 years of OilProduction in the North Sea. Suggest four factors that could explain this.
Answer 1.
Artificial lift was not required due to properties of developments typical in that period:
• Light oil being produced from high permeability reservoirs i.e. (relatively lowdrawdowns),
• Reasonable high GOR aided natural flow,• Water Injection supported reservoir pressure above the bubble point pressure atnear hyrostatic or greater pressures, allows the production wells to continue to flowunder natural flow
• Large reservoirs and distant well spacing delayed water breakthrough• High water cut wells were shut-in. Other, lower watercut wells produced inpreference to watered out wells since production was facility or pipeline constrained.
Question 2.
Two types of artificial lift were installed once it became apparent that productionpressure boosting would be required. Which types were these and what were thereasons that they were chosen?
Answer 2.The two artificial lift types were:
Gas Lift• High GOR water drive reservoirs• High reservoirs permeabilities• No need for very low FBHPs (Water Injection, drawdown was relatively low).• High Pressure gas availability, often surplus to power or export requirements• (reasonably) deviated holes• Wireline equipment maintenance - no extra demand placed on drilling rig whichcould continue drilling development wells.
• Tolerate some sand production as was experienced in completions without sandcontrol
• No extra space required at production wellhead• Full-access to oil producing formation
ESPs• High production rates even at high water cuts• Suitable for highly deviated wells• Logging / coiled tubing access to formation via Y tool• Flexible - flow rate controllable over a wide range
- downhole flow measurement and pump condition monitoring available
Department of Petroleum Engineering, Heriot-Watt University 67
2Selection of Artificial Lift Types2
• Efficient use of Energy• Can pump against high wellhead pressures e.g. for satellite wells• No extra space required at production wellhead
Question 3.
New completion technology has contributed to reduction of the Operating costs ofArtificial Lift Equipment. Name two significant developments.
Answer 3.
Examples of new technology:• Coiled tubing for:
- Insert strings- Conveyed pumps
• Wireline maintainable pumps.• Variable speed drives for ESPs• Downhole measurement & control for ESPs
Question 4.
Which IPR Curve (AL1 or AL2) is more beneficial for Artificial Lift?
0 70 140 210 280 350 420 480 560 630 700
2500225020001750150012501000750500250
0
0 60 120 180 240 300 360 420 480 540 600
2500225020001750150012501000750500250
0
Total Liquid Production bbl/d
Inflow Performance Relationship (AL2)
Inflow Performance Relationship (AL1)
Total Liquid Production bbl/day
Bot
tom
Hol
e P
ress
ure
(psi
)
xx
x
x
x
x
x
x
x
x
Bot
tom
Hol
e P
ress
ure
(psi
)
68
Answer 4.
A “straight line” inflow performance relationship associated with a dead oil is morefavourable than the curved “Vogel” relationship found when well inflow takes placebelow the fluid’s bubble point. This is because for a “straight line” inflow performancerelationship the % Increase in production is directly related to the % increase indrawdown achieved by the introduction of the form of artificial lift. The increase inproduction is considerably less for a curved “Vogel” relationship.
Question 5.
i) Why are these curves different?ii) What impact might this have on the selection of the Artificial Lift type?
Answer 5.
i) A straight-line IPR assumes that oil is undersaturated, that is, only slightlycompressible. This condition does not apply to gases or saturated oil wells, both ofwhich are highly compressible. The effect of compressible gas and two-phase flowon IPR results in “larger-than-linear” pressure drops being required to increase theproduction rate i.e. a curved IPR is observed in this case. The rate - pressure relationtends to show a more pronounced curvature at higher production rates.
ii) By applying the same drawdown, i.e. producing under similar flowing bottomholepressures, wells with the ‘Straight line” IPR (dead or undersaturated oils) wouldyield higher production rate than wells with the curved “Vogel” IPR. The increasingproduction of associated gas due to producing below the bubble point pressure in thelatter case would tend to favour the installation of Gas lift while, for example, RodPumps can be applied to the dead oil or undersaturated oil wells.
Question 6.
List up to 6 key features for both Rod Pumps and Gas Lift that form the basis of thefollowing statement:
“Worldwide, 85% of Artificial Lift equipment installed is rod pumps. This is mainlyin low production rate wells while gas lift is the most popular artificial lift techniquefor medium rate wells”.
Answer 6.
Rod Pump - Main features• The vast majority of wells produce at low rates (generally less than 100 bpd) andmoderate depths
• Relatively cheap, so their use can be justified on such low rate wells• Rod pumps are mechanically simple to operate and easy to repair/maintain/replace.Can be operated by inexperienced personnel
Department of Petroleum Engineering, Heriot-Watt University 69
2Selection of Artificial Lift Types2
• Sensitive to gas and solids (wax/scale/sand) - Solids can cause wear as well asdamage moving parts which then need to be replaced
• Not suitable for (highly) deviated wells (most land wells are near vertical)• Obtrusive in urban locations. Equipment too heavy for offshore use• Pump can be easily changed and performance monitored using relatively simpleand inexpensive techniques
• Viscous oil can be pumped
Gas Lift - Main features• Suitable for medium to high rates• Suitable for water drive reservoirs with a high bottomhole pressure• High well PIs and high permeabilities mean FBHP can be excessively high, limitingproduction
• High GOR => advantage rather than a drawback• Gas has to be available• Wireline serviceable at deviation up to 65˚. Coiled tubing can service more highlydeviated / horizontal wells
• Limited surface requirements once gas available - can be used off-shore or in urban locations
• Fully open tubing giving access for production logging• Subsurface tubing, and annular, safety valve• Flexible - gas lift string design can be adjusted as well conditions change• Forgiving of poor design & operation, but difficult to run efficiently• Can handle (tolerate) produced solids
Question 7.
What considerations are important when choosing an Artificial Lift Method forsubsea wells for a satellite development at a distance 30 Km from a host platform?
Answer 7.
ESPs are normally the preferred Artificial lift method for the following reasons:• Can generate high pump pressures to overcome extra friction from 30 kmpipeline
• Does NOT use power fluid (gas for gas lift, liquid for hydraulic pumps) whichwould lead to extra friction in pipeline.
- Hence electricity probably preferred source of power- Remote control capability at long distances
• Advantageous to place pump so as to minimise length of flowline with multi phaseflow
• System design (artificial lift type/tubing/flowline/reception facilities) should besuitable for complete life of oil field.
• Long pump lifetime (reliability to complement lifetime design)
Multiphase pumping means Subsea pressure boosting can also be considered as anoption as well as ESPs
70
FURTHER READING
(1) CHOLET, CProgressing Cavity PumpsParis: Edition Technip 1997ISBN 2-7108-0724-6
(2) BROWN, K EThe Technology of Artificial Lift Methods. Volumes 1-4Pen Well Books, 1980
(3) PERRIN, D.Well Completion and Servicing - Oil and Gas Field Development TechniquesParis: Editions Technip, 1999ISBN 2710807653
Department of Petroleum Engineering, Heriot-Watt University 71
2Selection of Artificial Lift Types2
C O N T E N T S
1. INTRODUCTION2. GAS LIFT INTRODUCTION3. GAS LIFT APPLICATION
3.1. Gas Lift Advantages and Limitations3.2. Review Example Gas Lift Completion
Designs4. GAS LIFT DESIGN OBJECTIVES
4.1. Gas Lift Design Constraints4.2. Gas Lift Design Parameters4.3. The Surface Gas Network
5. THE UNLOADING PROCESS DESCRIBED5.1. Safety Factors5.2. Gaslift Valve Spacing Criteria
Summarised6. SIDE POCKET MANDRELS
3.6.1 Other Uses of Side Pocket Mandrels7. GAS LIFT VALVE MECHANICS
7.1. Casing or Inflow Pressure Operated (IPO)Valves
7.2. Dome Pressure Calibration7.2.1. Temperature Correction7.3. Valve Performance7.3.1. Dynamic Valve Performance7.3.2. Valve Performance Flow Model7.4. Proportional Response Valves7.5. Dynamic Valve Response and Gas Lift
Completion Modeling7.6. Well Stability
8. GAS LIFT DESIGN PROCEDURES8.1. An Example Design - Optimising The
Performance of a Gas Lifted Well8.2. An Example Design - Gas Lift Unloading
Calculations8.3. Further Gas Lift System Considerations8.4. Further Gas Lift System Calculations
9. OPERATIONAL PROBLEMS9.1. Gas Quality9.2. Solids9.3. Changes in Reservoir Performance9.4. Gas Supply Problems9.5. Well Start - Up (Unloading)9.6. Well Stability9.7. Dual Gas Lift9.8. Trouble Shooting9.9. Trouble Shooting Techniques9.10. Some Field Examples of Operational
Problems
33Gas Lift
10. FIELD PRODUCTION OPTIMISATION11. NEW TECHNOLOGY FOR CONTINUOUS
FLOW GAS LIFT12. INTERMITTENT GAS LIFT13. GRAPHICAL GAS LIFT DESIGN EXERCISE
FOR WELL EDINBURGH - 213.1. Introduction13.2. Initial Condition - The "Dead" Well13.3. Construction of The "Equilibrium Curve"13.4. The Unloading Process13.5. Gas Lift Optimisation Exercise
14. FURTHER READING
2
LEARNING OBJECTIVES:
Having worked through this chapter the Student will be able to:
• Describe the gas lift process.
• Explain the impact of the key gas lift process variables.
• Identify application areas/advantages for gas lift.
• Discuss the limitations of the gas lift process.
• Describe the well unloading process.
• Identify and explain the action of gas lift hardware components.
• Design a gas lift completion.
• Identify reasons why efficient gas lift depends on availability of high qualitydata.
• Construct a methodology for revenue optimisation with limited gas availability.
• Describe the intermittent gas lift and plunger lift processes.
Department of Petroleum Engineering, Heriot-Watt University 3
33Gas Lift
1. INTRODUCTION
Chapter 2 introduced the concept of artificial lift and discussed the different types ofequipment that a Production Technologist can choose from. It was complete apartfrom gas lift, the subject of this chapter. The objective of installing gas lift in acompletion is to increase the drawdown on the producing formation by injecting gasinto the lower part of the tubing string and consequently reducing the flowing gradientin the production string. The concepts of multiphase flow and well performancediscussed in Chapter 1 are obviously very important here.
We will first introduce the basics of gas lift and discuss its advantages and disadvantages.The design, operation and maintenance of the gas lift valves, which control the gasinjection from the annulus into the tubing, will then be described. The procedure todesign a gas lift completion string using one of the commercially available computerprograms will be discussed. A manual design exercise will illustrate the designprocess. Typical gas lift operational problems and their solution will then be dealt withand the need for continual optimisation of the gas lift reviewed. Finally, some of themost recent developments in gas lift technology will be discussed.
2. GAS LIFT INTRODUCTION
A continuous flow gas lifted well completion has been sketched in figure 1. Thecompletion differs from the natural flow completions discussed earlier in that:
Injected Gas(Control and Metering)
Produced Fluid and Injected Gas to Separator
Gas Lift Valves
(c) Large Gas Bubble Displaces Liquid Slug
(b) Gas BubbleExpands as
the Hydrostatic Pressure Reduces
(a) Injected Gas ReducesAverage Fluid Density
Gas Injected at"Operating Valve"
Producing Formation
Perforations
Liquid
Gas
(a) Reduction ofFluid Density
(b) Expansion of Gas
Bubbles
Liquid
Gas
(c) Displacement of
Liquid Slugsby Gas Bubbles
Liquid
Gas
Figure 1
Gaslifted Well Completion
4
(i) Gas, at a controlled volume and pressure, is injected into the tubing/casingannulus.
(ii) The tubing string has been fitted with a number of gas lift valves. These valvesare installed at carefully spaced intervals so that any liquid present above them in thecasing/tubing annulus (e.g. due to killing of the well) can be removed by injectionof gas at the top of the well annulus leading to the liquid U-tubing into the tubing andits subsequent ejection from the well. The gas injection point into the tubing is thentransferred to successively deeper gas lift valves (see section 3.5 for details).
(iii) The gas is injected into the tubing through the “operating valve”. The injectedgas enables the well to resume production by :
(a) the injected gas reducing the average fluid density above the injectionpoint.
(b) some of the injected gas dissolving into in the produced fluids, providingthey are undersaturated with respect to the gas solubility. The remainder,in the form of bubbles, will expand due to reductions in the hydrostaticpressure as the fluids rise up the tubing.
(c) the coalescence of these gas bubbles into larger bubbles occupying thefull width of the tubing. These bubbles are separated by liquid slugs,which the gas bubbles displace to surface. This is called slug flow.
The design of a gas lift completion thus consists of two separate distinct parts:
(i) Choice of the installation depth, type and design of the gas lift valves placedabove the operating valve so that any liquid in the tubing and casing/tubing annuluscan be unloaded via the wellhead (see section 3.5).
(ii) Optimisation of the flowing gas lifted well. The well essentially behaves as aconventional flowing well, except that the gas/liquid ratio (GLR) suddenly increasesat the operating valve depth (see section 3.8).
The wellbore opposite the perforations is treated as the node pressure when the systemis analysed using the “nodal analysis” process discussed in chapter 1.12. The analysisequates the following at any given flow rate:
Inflow to Node (the perforations):
Preservoir
- Pdrawdown
= Pperforations
Outflow from Node (the perforations):
Pseparator
+ ∆Pflowline
+ ∆Pchoke
+ ∆P(tubing above operating valve)
+ ∆Ptubing below operating valve
= Pperforations
The pressure drop across the tubing below the gas injection valve is estimated withusing multiphase flow correlations (chapter 1.1.7) or pressure traverse curves (chapter
Department of Petroleum Engineering, Heriot-Watt University 5
33Gas Lift
1.1.6) using the “natural” gas liquid ratio. The pressure drop between the gas injectionvalve and the surface is calculated using the “enhanced” gas liquid ratio calculatedfrom the sum of the {lift + produced} gas rate divided by the liquid production rate.
Figure 2 illustrates a pressure traverse across the well when it has reached steady stateoperation. The gas is being injected at the wellhead at a pressure of 1100 psi. Thepressure of the gas in the annulus increases with depth due to its density (typically atthe rate of 30 psi/1000 ft). The gas is initially being injected at the valve 4 at 3800 ft.The well is producing with a 500 psi drawdown. The flowing pressure gradient fromthe producing perforations to the operating gas lift valve is equal to 0.44 psi/ft. Thereis a 250 psi pressure drop across the gas lift valve and the average fluid gradient abovethe injection valve has been reduced 0.27 psi/ft by the injected gas. The situation fordeeper gas injection is also sketched in which the gas is being injected through valve7 at 5000 ft. The gas lift pressure is now just sufficient to allow injection to occur ifthe pressure drop across the gas lift valve is restricted to 50 psi. It can also be seen thatthe deeper injection allows the drawdown to increase to 850 psi.
Producing Formation
Perforations
0 500 1000 1500 2000 2500 3000
1000
2000
3000
4000
5000
6000
7000
Wellhead Annular Gas Injection Pressure (psi)
Dep
th (
Ft.
TV
D)
Casing (Gas) Pressure Gradient
Pressure Drop Across Valves
Gas Injectionat Valve 4
Gas Injection at Valve 7
Flowing Tubing P
ressure Gradient
Above P
oint of Injection
Flowing Tubing Pressure Gradients
Below Point of Injection
Drawdown, Valve 4 gas injection
ReservoirPressure
Drawdown, Valve 7 gas injection
Flowing Bottom HolePressure, Valve 7
Flowing Bottom HolePressure, Valve 4
Injected Gas(Control and Metering)
Produced Fluid and Injected Gas to Separator
1
2
3
4
5
67
Operating Gas Lift Valve
It can be appreciated from this diagram that the gas injection pressure is the maincontrol on the depth of gas injection while the gas injection rate also contributes to theextent of the reduction in the flowing pressure gradient. These parameters can beadjusted as required on a day-to-day basis. The pressure settings of the gas lift valves
Figure 2
Pressure Traverse Through
a Well
6
(which control the pressure levels at which the valve opens and closes - see section7.3) can be adjusted when required using wireline techniques - see section 6). Thedepths at which the valves are set can only be altered by pulling the tubing andrecompleting the well with a tubing string in which the spacing between the sidepocket mandrels has been altered.
Increases in the gas injection rate through a gas lift valve set at a given depth willincrease the fluid production rate until a maximum is reached (figure 3). At this pointthe “reduction in average fluid density in the tubing due to a slight increase in the gasinjection rate” is being exactly counterbalanced by the “increased frictional pressurelosses due to the greater mass of fluid flowing in the tubing”. Further increases in thegas flow rate will result in the friction term increasing relatively faster than thehydrostatic head reduction term. This is the “technical optimum gas injection rate”at which the well production is maximised.
Gas Injection Rate
Pro
duct
ion
Rat
e
Unstable flow below thisrate due to too low
gas injection rate
Maximum liquid productionor technical optimum gasinjection rate
Economic optimum gas injection ratewhere marginal extra gas injection
cost balances marginal extraproduction revenue.
Some wells flow"naturally" withoutgas lift.
Others require"Kick off" gas toinitiate production
“The maximum economic gas injection rate” will be somewhat lower - this is the gasinjection rate at which the marginal cost of providing extra injection gas is equal to themarginal revenue from the extra well production.
Figure 3 also illustrates that gas lift may be applied to increase the production fromwells in which will flow naturally at a low(er) rate. The second case illustrated is fora well which is “dead” and does not produce without some form of artificial lift. Gasthen has to be injected at a certain rate (“kick-off” gas) before any well production ispossible.
Figure 3
Effect of gas rate on well
production
Department of Petroleum Engineering, Heriot-Watt University 7
33Gas Lift
An efficient gas lift system depends on a continuous supply of gas at the specifiedpressure. A considerable infrastructure is required for gas lift. This is normally onlyinstalled when there are a number of wells in the area using gas lift as the preferredform of artificial lift. A typical gas lift system arrangement is shown in figure 4. Thisfigure shows several wells producing into a production manifold. The gas is thenseparated, compressed and dried in a dehydration unit. Any excess gas may be soldor make up gas imported, as required by the demand of the gas lift system. The liftgas is supplied to the gas lift manifold, after which the injection gas flow rate andcasing head pressure are adjusted before injection into the individual wells.
Dehydrationunit
Compressor
3 PhaseSeparator
Oil tostorage
Water todisposal
Production manifold
Production pressure and flow ratemeasurement
Injection gaspresssure and flowrate measurement
Importmake upgas
Surplussales gas
Gas
P F P F
Injection Gas m
anifold
The metering and control equipment for a gas lifted well that is being individuallytested is illustrated in figure 5. Both manual and automatic lift gas control areillustrated.
Figure 4
Gas lift system
8
Producing Formation
Perforations
Wing Valve Safety Valve
Inlet Valve
Master Valve
Choke Box
Echometer(Measures fluidlevel in annulus)
Gas Lift Pipe Line
Casing
Tubing
Operating Gas Lift Valve
(Orifice) GasFlow Meter
(Orifice)Flow Meter
MainValve
Needle Valveto Control Lift Gas
Gas Lift Manifold
Gas
Oil
Water
Packer
Oil Pipe Line
Data Logger
UnloadingGas Lift Valves
Data Logger
Production test Separator
MainValve
Gas Lift Manifold
Flow meter thatautomatically adjustschoke setting
Manual control
Automatic control
OR
3. GAS LIFT APPLICATIONS
The process described above is called “continuous flow gas lift”. “Intermittent gaslift” is used in low rate production wells. This approach involves switching off theinjection gas at regular intervals so as to allow the fluid level in the well to build up.The gas injection is recommenced, and the fluid in the tubing lifted to surface, whena sufficient depth of produced fluid is present in the well. The cycle is then repeated.Intermittent gas lift is thus used for cases when the outflow capacity of the gas liftedtubing is greater than the formation’s capacity to produce fluid into the well.
The module on multiphase flow in vertical tubing explained how the flowing gas willby-pass some of the liquid in the tubing (the slip phenomenon). This liquid will fallback down the well each time the gas lift is switched off. Fall back can be avoided byinstalling a plunger at the bottom of the well. Gas injection now occurs underneaththis plunger, which rises upwards, displacing the liquid above it to the surface. Theplunger falls to the bottom of the well when the gas is switched off. The downholecompletion is arranged so that inflowing fluid can collect above the plunger while acheck valve ensures that the injected gas can not be injected into the formation. Thecycle can now be repeated at a regular time interval. This will depending on the well
Figure 5
Metering and control of a
gas lifted well.
Department of Petroleum Engineering, Heriot-Watt University 9
33Gas Lift
productivity and the volume of liquid displaced to the surface by the plunger. Thismethod is described in greater detail in chapter 3.12.
Gas lift has been applied to a wide range of production scenarios - as can be seen fromTable 1. In fact, gas lift is the only artificial lift method that actually works better ina well that is producing at a significant gas/liquid ratio. Gas lift is often the preferredartificial lift method for wells with a:
(i) high gas-oil ratio;
(ii) high productivity index;
(iii) (relatively) high bottom hole pressure due to reservoir pressure support beingprovided by a natural or artificial water drive.
1 Production wells which will not flow naturally.
2 Increase production rate in flowing wells.
3 Unload liquid from wells that will flow naturally once on production.
4 Unload liquid in wet gas wells which would otherwise cease to flow.
5 Back flow injection wells.
6 Lift aquifer wells.
The key point of gas lift is that a reliable, adequate (in terms of pressure and flow rate)gas supply has to be available at all times. The proviso at the end of the sentence isthe key one. When the field/wells are operating normally the (lift) gas system (figure3) will be fully charged with gas. This gas will be recovered and recirculated manytimes. Extra volumes of “make-up” gas associated with the current oil production willonly be required to make good any losses from the system, as well as any gas used forcompression or other power requirements. When planning a gas lift installation fora field one should specifically allow for the:
(i) decrease in (fresh or make-up) gas supply as the field reserves are depleted andthe well water cut increases. This can result in gas being imported during the lateproject life, particularly for offshore developments when the produced gas is alsoused to generate the platform’s electrical power.
(ii) case when none of the wells flow naturally. An external gas source is thenrequired to bring the (first) well(s) onto production after a facility shutdown.(Vapourised, liquid) Nitrogen can be used for this purpose if there is no provision toimport natural gas.
(iii) fact that, if only a low rate gas supply is available, it will take a long time toreturn all the wells to production after a shutdown.
(iv) choice of lift gas injection pressure has to be made at an early stage in the projectlifetime when the gas compressor specifications are drawn up and little informationmay be available about actual well performance.
Table 1
Continuous flow gas lift
applications
10
3.1 Gas Lift Advantages and LimitationsThese are summarised in tables 2 and 3. They are self explanatory if read inconjunction with the above discussion.
Operation of gas lift valves is unaffected by produced solids (sand etc.)
Gas lift operation is unaffected by deviated or crooked holes.
Use of side pocket mandrels allows easy wireline replacements of (inexpensive)
gas lift valves when deviation <60 .
Provides full bore tubing access for coiled tubing or other well service work.
High fluid gas oil ratio improves lift performance rather than presenting problems
as with other artificial lift methods.
Flexible - can produce from a wide range depths & flow rates
- uses the same well equipment from 100-10,000bpd production rates
- copes with uncertainties and changes in reservoir performance,
reservoir pressure, water cut & production index over the well life.
Low surface profile important for offshore & urban locations.
Tubing & annular subsurface safety valves available when required by
safety regulations.
Gas lift tolerates "bad" design - though "good" design is more difficult.
Gas lift has a low initial (downhole) equipment cost.
Gas lift has a low operational and maintenance costs. Major workovers are
infrequent when wireline servicing is possible.
Well completions are relatively simple. This can be important in remote areas.
Gas lift operation independent of bottom hole temperature.
High back pressure on sandface due to fluid in the tubing restricting production.
- e.g. lifting a well with a Productivity Index of 1 bpd/psi from 10,000 ft with a
static bottom hole pressure of 1000 psi is difficult.
- Flowing bottom hole pressure is greater than with e.g. Electric Submersible
Pumps. This leads to potential loss of reserves.
Gas lift is inefficient in energy terms (typically 15-20%).
Gas compressors have a high capital cost. They require expensive maintenance &
require skilled operations staff. However, they may already be required for gas sales.
Annulus full of high pressure gas represents a safety hazard.
High installation cost can result from top sides modifications to existing platforms
e.g. Compressor installation.
Adequate gas supply required throughout project life
- Decreasing BHP, increasing water cut etc.
- Sufficient gas to start up FIRST well
- Slow start up after facility shut down
- Increased gas handling requirements in facilities.
Gas lifting of viscous crude (<15 API) is difficult and less efficient.
Wax precipitation problems may increase due to cooling from (cold) gas injection &
subsequent expansion.
Hydrate blocking of surface gas injection lines can occur during cold weather if gas
inadequately dried.
Lifting of low fluid volumes is inefficient due to gas slippage.
Good data management and complete network modelling required for efficient /
maximum profitability operation.
Table 2
Gas lift advantages
Table 3
Gas lift limitations
Department of Petroleum Engineering, Heriot-Watt University 11
33Gas Lift
3.2 Review Example Gas Lift Completion DesignsQuestionFigure 6 shows types of completion designs for gas lifted wells. You should:
(i) identify the type of completion;
(ii) describe the completion’s advantages and disadvantages.
Each completion is discussed in turn below:
Production
Production
Production
Gas
Gas
Gas
SCSSSVSCSSSV
(a) (b) (c)
AnswersFigure 6 (a) Single String Continuous Gas Lift CompletionThis is the standard completion design. Gas is injected in the annulus and the producedfluids are lifted to the surface through the tubing. The well may be completed on asingle or on multiple formation zones. In the latter case, the separate zones may be:
(i) produced together (commingled) or
(ii) isolated from one another by packers. The required zone can then be producedselectively by opening and closing the appropriate sliding side doors.
Figure 6 (b) Annular Flow Gas Lift CompletionThe gas is injected down the tubing and the production flows up the annulus. This well
Figure 6 (a) to (c)
Gas lift completions designs
12
design can be found onshore in the Middle East. Higher production rates are achievedcompared to the conventional production configuration where the produced fluidflows up the tubing. This is due to the reduced (frictional) pressure drop in the annuluscompared to the tubing due to its annulus’s larger flow area. The disadvantages arethat corrosion of the casing by the produced fluids will lead to a loss in well integrity(see section 3.9.1). Also, a decline in the well production rate will lead to severeslugging earlier than for tubing flow.
Figure 6 (c)Continuous Gas Lift without the surface section of the Casing/TubingAnnulus being Live (filled with Gas)This completion features the gas being injected into a separate injection string with itsown Surface Controlled Sub Surface Safety Valve (SCSSSV) installed below a dualpacker. The gas is then injected into a single tubing designed for conventional,continuous gas lift. This production string also has a SCSSSSV installed below theupper, (dual) packer. This type of well design has been installed in the North Sea.
Plunger
Gas
(d) (e) (f)
One way orcheck valve
Gas
Production Production
Gas
Short stringproduction
Long stringproduction
Figure 6 (d)Dual, Gas Lifted CompletionThis completion allows two zones to be independently produced by gas lift throughseparate production strings. It is quite difficult to achieve optimum lift on both stringssince the action of the gas lift valves on the different strings interfere with each other- see section 3.9.7.
Figure 6 (d) to (f)
Gas lift completions designs
Department of Petroleum Engineering, Heriot-Watt University 13
33Gas Lift
Figure 6 (e)Intermittent, Plunger LiftThis is installed in low rate wells, particularly in the USA, where the inflow rate fromthe formation is low and smaller than the outflow capacity of the gas lifted tubing. Theplunger prevents fallback of the liquid when the gas is switched off. This liquid wouldnormally have been bypassed by the gas flowing up the tubing (slip) on its way tosurface.
Figure 6 (f) Single Valve (Subsea) Completion
This completion is used when intervention (change of gas lift valve settings etc) isdifficult and/or expensive. The single (orifice) operating valve minimises operationalproblems. However:
(i) the depth of lift gas injection is restricted since unloading valves are not used.
(ii) this injection depth is often maximised by increasing the gas injection pressureabove the normal 1,000-1,200 psi. A compressor capable of delivering gas at sucha higher pressure can only be provided at a substantial extra cost. Compressorpressures over 3,000 psi have been used during the unloading process so that asubstantially greater depth of injection can be achieved. Conventional, much lowerpressures will be required once gas lift has been initiated and the well is flowingsteadily.
4. GAS LIFT DESIGN OBJECTIVES
The gas lift system designed for installation in a specific well should meet thefollowing objectives:
(i) Maximise the (net) value of oil produced. This normally implies that the:
(a) operating valve, through which the gas will be continuously injected,should be situated as deep as possible and
(b) gas injection rate should equal the economic limit at which the marginalvalue of the extra oil produced equals the marginal cost of providing thisextra gas (figure 3).
Further optimisation is required when more than one well is being produced andthere is insufficient lift gas available to meet this economic criteria in all wells (seesection 3.10)
(ii) Maximise design flexibility. The gas lift design should be capable of copingwith the expected changes in the well producing conditions during its lifetime, aswell as the “unplanned” uncertainties in reservoir properties and performance.These changes normally involve deterioration, from a well productivity point ofview, due to decreases in the Reservoir Pressure and Well productivity Index andincreases in the Water Cut.
14
(iii) Minimise well intervention. This is particularly important in subsea or otherwells where wireline access is difficult or impossible.
Well completions with a “dry” tree and deviations less than 60o allow the option toreplace the gas lift valve by a relatively quick, wireline operation. The operatingparameters (or valve performance) of the gas lift valves installed in the side pocketmandrels can thus be adjusted at any time in the well’s life i.e. the tubing productionconditions can be adapted to take into account changes in the reservoir conditions andthe well performance. These operating parameters include the:
(a) tubing or casing pressures (depending on the type of side pocketmandrel installed) at which gas flow through the gas lift valve startsand stops and
(b) port (or choke) size, which controls the maximum volume of gas thatcan be injected as well as the associated pressure drop due to the gas flowthrough the valve.
This ability to modify the valve performance when required leads to great flexibilityin the choice of gas lift operating parameters, despite the fact that the installationdepth of the gas lift valves is fixed. {The (side pocket) gas lift mandrels within whichthe valve is placed are permanent fixtures in the completion string, having beeninstalled during the well completion process}.
The flexibility of a particular gas lift design is further increased by installing one ortwo extra gas lift valves as possible both above and below the chosen depth of theoperating valve. They should be placed as close together as possible, but sufficientlyfar apart that they do not interfere with each other’s operation (typically 150 mvertical depth apart). This is known as the “bracketing envelope”. The inclusion ofthe bracketing envelope mandrels will allow the operating valve to be moved to aslightly higher or lower depth, as dictated by the well & reservoir performancechanges during the well life. This procedure maximises the (liquid) production by,for a given gas injection rate, allowing the well to be lifted from as deep as possiblecommensurate with the current producing conditions.
(iv) Stable Well Operation. Well “heading”, in which the Tubing Head or CasingHead Pressure Shows regular changes (see figure 40 and section 9.6) should beavoided. Stable operation - with a constant value for the casing and tubing headpressures - should be aimed for. This is because stable well operation will alwaysproduce more oil and, often, require less lift gas than unstable gas lifted welloperation.
Casing Head pressure excursions of as little as 5 psi can indicate valve multipointing{the gas injection point changing from one valve to another or a second valve cycling(opening and closing) in addition to the operating valve}.
4.1 Gas lift design constraintsThere are three different sets of circumstances in which a gas lift design has to be madeand the above gas lift design objectives need to be met:
Department of Petroleum Engineering, Heriot-Watt University 15
33Gas Lift
(i) The valves are to be installed as an integral part of the tubing i.e. side pocketmandrels and retrievable gas lift valves are not used. The valve spacing andoperating parameters is then fixed until the tubing is pulled and the well recompleted.Such completions are usually used in shallow, relatively depleted land wells wherea low cost hoist can quickly carry out the operation.
(ii) Side pocket mandrels are included in the completion string. Dummy valves willbe installed in these mandrels initially if the well is to be produced for a period undernatural flow. Gas lift valves will only be installed at a later date when they arerequired to maintain the production rate. Producing the well under natural flow willprovide the information to remove much of the uncertainty in the well and reservoirperformance. This production experience can then be used to choose the valvesettings when the time comes to replace the dummy valves with real valves.
(iii) A gas lift design is to be installed in a well that was completed sometime ago.The valves are to be run into the existing side pocket mandrel locations. The wellconditions (Productivity Index, Water Cut, Reservoir Pressure etc.) may havechanged considerably compared to those when the gas lift was first designed and /or first installed. Further, these conditions may have changed in a manner which wasnot anticipated when the original design calculations were made.
The constraints, which limit the design options, increase from (i) to (iii).
4.2 Gas lift design parametersThe gas lift design process has to answer the following questions to meet the aboveobjectives:
(i) How many unloading valves are required and at what depths should they beplaced?
(ii) What are the required settings for the Unloading Valves?
(iii) What is the depth of the operating valve where the gas is continuously injected?
(iv) What is the gas injection (or casing head) pressure?
(v) How much lift gas should be injected?
(vi) What is the tubing head pressure for the target flow rate?
This is translated into practice by ensuring that the gas lift valve spacing and pressuresetting are such that:
(i) the operating valve should have adequate flow capacity and be placed as deepas possible,
(ii) the available lift gas pressure must be able to displace the fluid in the casing tothe operating valve depth,
16
(iii) all valves can be opened by the appropriate producing pressure gradient, whilethe other valves above it are closed.
4.3 The Surface Gas NetworkFigure 4 illustrated the complete gas lift system. It will have become apparent fromthe above and the following sections that the performance of the gas lift system willdepend on the pressure and flow capacity of the lift gas available at the well head. Thesurface piping network should thus be designed to:
(i) have minimal (< 100psi) pressure loss between the compressor and the mostdistant wellhead,
(ii) prevent one well from interfering with a second well by having sufficient pipevolume to dampen pressure surges and
(iii) provide individual gas measurement and flow control for each well
Large diameter piping encourages all the above - typically 4 in OD piping is used forthe main backbone of the system with individual 2 in OD flow lines installed to eachwell. A ring main system is an option for large systems employing more than onecompressor - the gas lift manifold for a group of wells and the compressors beingattached to the gas supply ring main as appropriate.
5. THE UNLOADING PROCESS DESCRIBED
0 500 1000 1500 2000 2500 3000
1000
2000
3000
4000
5000
6000
7000
Pressure (psi)
Tubing Pressure
Casing Pressure
True
Ver
tical
Dep
th (
Ft.
TV
D)
Producing Formation
Perforations
Injection Gas
To Separator / Storage Tank
Top Valve Open
Second Valve Open
Third Valve Open
Fourth Valve Open
Reservoir pressure
Casing Pressure
and
Tubing Pressure
Figure 7
The "dead" well
Department of Petroleum Engineering, Heriot-Watt University 17
33Gas Lift
Figure 7 shows the situation when a well planned for gas lift has just been (re)completed.The fluid level in the casing and the tubing is just below the surface and balances thereservoir pressure. The well is dead - no fluids are being produced. The hydrostatichead of the fluid column will equal the reservoir pressure, the actual fluid height willdepend on the liquid density - a column of water will have a lower height than an oilcolumn. No gas is being injected into the casing - both the tubing and casing have beendepressurised at surface to atmospheric pressure. All the gas lift valves are open dueto the hydrostatic head of the fluid.
0 500 1000 1500 2000 2500 3000
1000
2000
3000
4000
5000
6000
7000
Pressure (psi)
Tubing Pressure
Casing Pressure
True
Ver
tical
Dep
th (
Ft.
TV
D)
Injection pressure
Producing Formation
Perforations
Injection GasChoke Partially
Open
Top Valve Open
Second Valve Open
Third Valve Open
Fourth Valve Open
Reservoir pressure
To Separator / Storage Tank
Gas injection into the casing / tubing annulus has been started in Figure 8. The fluidis being U-tubed from the casing into the tubing through all the open gas lift valves.The gas lift pressure is sufficient to increase the fluid level in the tubing to the surfaceso that it flows via the surface flowlines into the separator. The pressure in thewellbore at perforation depth is greater than the reservoir pressure i.e. some of theliquid originally present in the well is being injected into the formation. This injectionof contaminated, potentially formation damaging, fluid can be prevented by installinga one way flow valve or check valve at the bottom of the tubing. It is important thatthe unloading process should occur at a controlled rate - the gas injection rate iscarefully controlled through the partially opened injection gas choke (see section3.9.5). This will prevent damage to the gas lift valves as the fluid flows from the casingand into the tubing via the open gas lift valves.
Figure 8
Gas lifted well unloading,
stage 1
18
0 500 1000 1500 2000 2500 3000
1000
2000
3000
4000
5000
6000
7000
Pressure (psi)
Tubing Pressure
Casing Pressure
True
Ver
tical
Dep
th (
Ft.
TV
D)
Producing Formation
Perforations
Injection GasChoke Partially
Open
Top Valve Open
Second Valve Open
Third Valve Open
Fourth Valve Open
Reservoir pressure
To Separator / Storage Tank
Figure 9 shows the situation when the unloading process has lowered the fluid levelin the casing annulus to the top gas lift valve. Gas injection into the tubing has nowcommenced. The injected gas partially evacuates the liquid in the tubing above thetop gas lift valve into the separator under multi-phase flow conditions. This partialevacuation reduces the fluid density in the tubing above the top gas lift valve andensures that further casing fluid to be unloaded through valves No. 2, 3 and 4; sincethe pressure in the tubing at these points is lower than the pressure in the casing. Thewell will also start to produce formation fluid if this reduction in pressure is sufficientto give a drawdown at the perforations.
N.B. Any fluid lost to the formation earlier on in the unloading process will beproduced back first. This will be “dead” i.e. not “live” formation fluid whose intrinsicgas content would reduce the hydrostatic head in the tubing and help bring the well intoproduction.
Figure 9
Gas lifted well unloading,
stage 2
Department of Petroleum Engineering, Heriot-Watt University 19
33Gas Lift
0 500 1000 1500 2000 2500 3000
1000
2000
3000
4000
5000
6000
7000
Pressure (psi)
Tubing Pressure
Casing Pressure
True
Ver
tical
Dep
th (
Ft.
TV
D)
Producing Formation
Perforations
Top Valve Open
Second Valve Open
Third Valve Open
Fourth Valve Open
Injection GasChoke Partially
Open
To Separator / Storage Tank
Reservoir pressureFlowing BottomHole Pressure
Drawdown
In Figure 10 the fluid level in the casing has now been lowered sufficiently to exposegas lift valve No. 2. The top two gas lift valves are open and gas is being injectedthrough both valves. All valves below also remain open and continue to pass casingfluid into the tubing. The tubing has now been unloaded sufficiently to reduce thebottom hole pressure below that of the reservoir pressure (SIBHP). This pressuredifference, or drawdown, induces flow of formation fluid from the reservoir into thewellbore i.e. the well is starting to produce.
Figure 10
Gas lifted well unloading,
stage 3
20
0 500 1000 1500 2000 2500 3000
1000
2000
3000
4000
5000
6000
7000
Pressure (psi)
Tubing Pressure
Casing Pressure
True
Ver
tical
Dep
th (
Ft.
TV
D)
Producing Formation
Perforations
Top Valve Closed
Second Valve Open
Third Valve Open
Fourth Valve Open
Injection GasChoke Partially
Open
To Separator / Storage Tank
Reservoir pressure
Drawdown
Flowing BottomHole Pressure
The process continues in Figure 11. The top gas lift valve has now closed due to thereduced pressure at this point. All the gas is being injected through valve No. 2.Unloading the well continues with valves 2, 3 and 4 open and casing liquid flowinginto the tubing via valves 3 and 4.
There are two basic types of gas lift valves - the valve open and closing action beingin response to either the tubing or the casing pressure. Thus the closure of the top gaslift valve was triggered by the reduction in the casing pressure (for casing pressureoperated valves) or tubing pressure (for fluid operated and proportional responsevalves) after gas lift had been established through valve number two. {See section 7on “Gas Lift Valve Mechanics” for a description of the construction and operation ofthese two types of valve.}
Figure 11
Gas lifted well unloading,
stage 4
Department of Petroleum Engineering, Heriot-Watt University 21
33Gas Lift
0 500 1000 1500 2000 2500 3000
1000
2000
3000
4000
5000
6000
7000
Pressure (psi)
Tubing Pressure
Casing Pressure
True
Ver
tical
Dep
th (
Ft.
TV
D)
Producing Formation
Perforations
Top Valve Closed
Second Valve Open
Third Valve Open
Fourth Valve Open
Reservoir pressure
Drawdown
Injection GasChoke Partially
Open
To Separator / Storage Tank
Flowing BottomHole Pressure
Figure 12 shows valve No. 3 having just been uncovered so that both the No. 2 and3 valves are passing gas. The bottom valve below the liquid level is also open andliquid unloading from the casing / tubing annulus into the tubing continues.NB. a deeper point of injection lowers the Flowing Bottom Hole Pressure; creating agreater drawdown and hence increasing the production rate.
Figure 12
Gas lifted well unloading,
stage 5
22
0 500 1000 1500 2000 2500 3000
1000
2000
3000
4000
5000
6000
7000
Pressure (psi)
Tubing Pressure
Casing Pressure
True
Ver
tical
Dep
th (
Ft.
TV
D)
Producing Formation
Perforations
Top Valve Closed
Second Valve Closed
Third Valve Open
Fourth Valve Open
Flowing BottomHole Pressure
Reservoir pressure
Drawdown
Injection GasChoke Partially
Open
To Separator / Storage Tank
Figure 13 shows that, similar to the chain of events that lead to the closure of valve No.1, the reduction in casing or tubing pressure once valve No. 3 starts to flow gas willresult in valve No. 2 closing. All the gas is now being injected through valve No. 3.
Figure 13
Gas lifted well unloading,
stage 6
Department of Petroleum Engineering, Heriot-Watt University 23
33Gas Lift
0 500 1000 1500 2000 2500 3000
1000
2000
3000
4000
5000
6000
7000
Pressure (psi)
Tubing Pressure
Casing Pressure
True
Ver
tical
Dep
th (
Ft.
TV
D)
Producing Formation
Perforations
Top Valve Closed
Second Valve Closed
Third Valve Closed
Fourth Valve Open
Injection GasChoke Partially
Open
To Separator / Storage Tank
Reservoir pressure
Drawdown
Flowing BottomHole Pressure
The process has continued to its logical conclusion in Figure 14. Valve No. 4 has beenexposed to gas flow and valve No. 3 has shut. All the gas is being injected throughvalve No. 4 - this is the operating valve. An operating valve can be either a gas liftvalve or a simple orifice.
Figure 15 is an ideal illustration of the development of the tubing head and casing head(or gas lift) pressure with time during unloading process described in Figures 7 to 14.The sequential reduction in casing head pressure as gas is successively injectedthrough the lower gas lift valves is shown. This is not always so clearly observed inpractice. The annulus pressure is not only controlled be the settings of the gas liftvalve(s) referred to above, but also by the balance between the casing head (surface)gas injection rate and the (total) gas passage rate through the valves. Further, it canbe seen that the erratic behavior of the tubing head pressure as the well is beingunloaded is replaced by a steady value as stable production is achieved when the wellis lifted from valve No. 4.
Figure 14
The producing gas lifted
well
24
.eru
sser
Pgni
buTdnagnisaCfognidroceR
neP
niw T
DATE ON
DATE ON
TIME
TIME
AM
AM
11PMMIDNIGHT1AM
2AM
3AM
4AM
5AM
6AM
7AM
8AM
9A
M
10AM
11AMMIDDAY 1PM
2PM
3PM
4PM
5PM
6PM
7PM
8PM
9PM
10PM
1000
800
600
400
300
200
100
500
700
9001000
800
600
400300
200100
500
700
900
1000
800
600
400300
200100
500
700
900
1000800
600
400300
200 100
500
700
900
1000
800
600 40
0 300
200
100
500
700
900
1000
800
600
400
300
200
100
500
700
900
1000
800
600
400
300
200
100
500
700
900
1000
800
600
400
300
200
100
500
700
900
Open chokebleed off tubingpressure.
Liquid only beingexpelled fromtubing.
First gas to surface.
Switch ongas lift.
Stable productionvia 4th valve
Top valveuncovered
2nd valveuncovered
3rd valveuncoveredCasing head or
lift gas pressure
Tubing headpressure
4th valveuncovered
5.1 Safety FactorsSeveral safety factors are normally introduced when preparing gas lift designs for realwells. This is to account for:
(i) errors in the valve’s pressure settings,
(ii) errors and fluctuations in the well data, lift gas injection pressure and the estimateof the valve temperatures under flowing conditions,
(iii) the pressure drop across the valve’s choke and that required to obtain sufficientmovement of the stem.
One method by which safety factors can be built into the design is illustrated in section3.13.4
5.2 Gaslift Valve Spacing Criteria summarisedThe well unloading process was described in the previous section. The choseninstallation depths of valves Nos. 1 - 4 will have been based on the following criteria:
(i) Specify a minimum number of gas lift valves. This will not only reduce the costbut also the number of potential leak paths.
(ii) Gas lift valves to be installed sufficiently far apart that they do not interfere witheach other’s operation (150m suggested minimum spacing).
Figure 15
Typical casing / tubing
pressure and production
rate measurement during
unloading of a gas lifted
well.
Department of Petroleum Engineering, Heriot-Watt University 25
33Gas Lift
(iii) (continuous) Gas injection through the operating valve occurs as deep aspossible based on current producing conditions.
6. SIDE POCKET MANDRELS
Completion equipment is available so that gas lift valves can be permanently installedas part of the tubing. Such wells require a workover if any repairs or changes to thesettings of the gas lift system need to be made. However, most completions employside pocket mandrels (figure 16) installed at appropriate depths in the tubing string aspart of the permanent completion. Side pocket mandrels allow gas lift valves to beinstalled (and recovered) in a live well using wireline techniques. They are ovalshaped accessories with an outside diameter greater than that of the tubing {figure 16(c)}. This shape allows the gas lift valve to be installed in the pocket placed to oneside of the tubing conduit, thus maintaining fullbore access throughout the completetubing length.
Lift Gas Injection
Lift Gas InjectedVia Port in Casing Li
ft G
as In
ject
edV
ia P
ort i
n Tu
bing
Lift Gas InjectedVia Port in Casing
Pocket for GasLift Valve
Poc
ket f
or G
asLi
ft V
alve
Produced Fluidsand
Lift Gas
Gas Lift Valve
Pressure SensitiveElement (Bellows)
Seal Element
Polished Bore
Produced Fluids
Lift Gas Injection
Produced Fluidsand
Lift Gas
Produced Fluids
(a) Side pocket mandrel for injectionpressure operated valve
(b) Side pocket mandrel for tubingpressure operated valve
(c) Cross section of side pocket mandrelwith gas lift valve installed in pocket
Seal ElementFull BoreAccessThroughTubing
Gas Lift Valve Body
Polished bore
Figure 16
Schematic view of side
pocket mandrel showing
comparison of injection and
tubing pressure operated
valves
26
Figure 16(a) is a schematic illustration of a side pocket mandrel intended for use withan Injection (or Casing) Pressure Operated (IPO) valve; while figure 16(b) shows theequivalent mandrel construction details for a Tubing Pressure Operated (TPO) valve.An IPO valve uses the injection or casing pressure acting on a pressure element(bellows) mounted in the upper portion of the valve to open and close the valve. A TPOvalve uses the tubing pressure for the same purpose. It can be seen that a different,internal arrangement of the mandrel is required for the two types of gas lift valves.
N.B. The mechanical design and operation of the gas lift valve are discussed in section 3.7.
Orienting Sleeve
Side Pocket Mandrel Body
Tool Discriminator to ensure gas liftvalve is inserted in side pocket mandrel
Latch Lug to hold gas lift valve in place
Polished Bore
Pocket for gas lift valve
Port
Figure 17 is a more detailed view of a side pocket mandrel showing some of the designfeatures which are necessary for valve installation and retrieval by wireline (see alsofigure 18). These include:
Figure 17
A side pocket mandrel
Department of Petroleum Engineering, Heriot-Watt University 27
33Gas Lift
(a) Gas lift valve beingrun into well on wirelinetool sling.
(b) Wireline "kickover" toolplacing gas lift valve intoside pocket mandrel.
(c) Recovering wireline toolstring after latching gas liftvalve in place in side pocketmandrel.
(b) (c)(a)
(i) An orientating sleeve. This contains a device (e.g. a vertical slot) which fitsinto its counterpart (finger) on the kickover tool so that gas lift valve and knuckle ofthe kickover tool are correctly aligned with the pocket orientation. This allows thetool to insert a gas lift valve into the mandrel or to recover the gas lift valve, asappropiate.
(ii) A tool discriminator which guides the gas lift valve into the pocket whiledeflecting larger diameter tools back into the main tubing.
Figure 18
Wireline installation of a
gas lift valve in a side
pocket mandrel
28
(iii) A latch ring on the valve locks underneath the mandrel’s latch lug, securing thevalve in place. This prevents the valve becoming detached from the mandrel oncethe well is placed on production.
(iv) Fluid tight seals are created between the gas lift valve seal elements and thepocket’s polished bore. These are situated both above and below the casing port forIPO valves or the tubing port (TPO valves).
6.1 Other Uses of Side Pocket MandrelsThe gas lift valve Side Pocket Mandrels can also be replaced by tools with thefollowing functions:
(i) Chemical injection valves;
(ii) Differential dump/kill valves;
(iii) Circulating valves;
(iv) Circulating sleeves;
(v) Dummy valves;
(vi) Water injection control valves.
7. GAS LIFT VALVE MECHANICS
The mode of action of the upper gas lift valves, in which the top valves are designedto open and close to allow the fluid in the casing/tubing annulus to be unloaded so thatdeep gas injection can be achieved, was described in the previous section. Theoperating valve is different, being designed to allow for a continuous flow of gas. Theupper gas lift valves have ports sized to pass only the required volume of gas, limitingthe rate at which the unloading takes place. A larger port is often installed in theoperating valve so that gas injection can be increased, if dictated by future well orreservoir conditions. Dummy valves are installed in the “Bracketing Envelope”where “live” valves are currently not required.
The following sections discusses the valve’s mechanical construction.
7.1 Casing or Inflow Pressure Operated (IPO) ValvesA schematic diagram of a casing pressure operated valve is shown in figure 19. The:
Department of Petroleum Engineering, Heriot-Watt University 29
33Gas Lift
Pdome
Abellows
Aport
Plug (Removed to charge dome with nitrogen)
Nitrogen Charged Dome
Bellows
Ptubing
Pcasing
Spring(To preventexcessivebellows collapse)
Stem
Ball or Stem Tip
Port Square Edged SeatChoke
Check Valve(Prevents back flow of produced
fluids from tubing into valve)
Chevron seals (Forms seal against polished
bore in body of side pocket mandrel)
(i) dome or top section of the valve is charged, via the plug, with nitrogen to therequired pressure (P
dome).
(ii) nitrogen charge acts on the bellows; exerting a force pushing the ball against thechoke or port and halting the flow of lift gas into the tubing.
(iii) spring prevents damage to the bellows due to excessive collapse when exposedto forces much greater than those generated by the nitrogen charge. Such excessivecollapse would result in the bellows loosing their elastic response to pressurechanges. If this occurs the gas lift valve needs to be changed.
(iv) choke or port prevents excessive (gas) flow rates.
(v) check valve prevents formation fluids flowing from the tubing into the annulus.
Simple, mechanical considerations allow us to derive the following equations:
The Closing Force, Fc, tending to seat the ball is given by:
FC = P
dome * A
bellows
Where Ax refers to the Area of component x and P
y refers to the Pressure at point y.
See figure 19 for further explanation of nomenclature.
The Opening Force (FO) is made up of two components: FO1, and FO2, where
FO1
= Ptubing
* Aport
FO2
= Pcasing
* (Abellows
- Aport
)
Figure 19
Schematic diagram of a
casing pressure operated
valve
30
and FO = F
O1 + F
O2
The Opening and Closing forces are equal just before the valve opens.
Pdome
* Abellows
= Pcasing
* (Abellows
- Aport
) + Ptubing
* Aport
or PP P (A A
(A Acasing dome tubing port bellows
port bellows
=−−
/ )
/ )1
Thus the key factors controlling the gas lift pressure required to open the valve are thedome and tubing pressures and the ratio of the bellows and port areas. IPO valvesnormally have the ratio (A
port/A
bellows) set as small as practical.
7.2 Dome Pressure CalibrationThe dome is charged with nitrogen and the gas lift valve installed in a Test Rack placedin a temperature controlled enclosure set at 60ºF (see Figure 20). The simulated P
tubing
is at atmospheric pressure (or O psig). A gradually increasing gas pressure is appliedto simulate P
casing. The valve opening pressure (P
opening) is recorded as the pressure
when gas begins to flow.
Gas Pressure(Simulates Pcasing)
Abellows
Aport
Simulated Ptubing
Pdome
Application of the above force balance equation gives:
Popening
= Pdome
/(1 - Aport
/Abellows
)
This can be used to confirm that the manufacturing of the gas lift valve is onspecification and that the dome nitrogen pressure, P
dome, has been set correctly.
The valve closing pressure (Pclosing
) may be measured by pressurising both the injectionand tubing sides so that the valve is fully open. The pressure on the tubing side isreduced and valve closure recognised as the pressure at which P
casing no longer
Figure 20
Test rack for gas lift valve
Department of Petroleum Engineering, Heriot-Watt University 31
33Gas Lift
decreases in line with Ptubing
. The design of the valve dictates that Pclosing
will be lowerthan P
opening, since starting from the open valve situation means that the injection
pressure acts on the complete bellows area.
The valve spread is defined as the difference between the valve test rack opening andclosing pressures i.e. (P
opening - P
closing). It is a measure of the difference between the
effective area of the bellows and the port. Figure 21 illustrates a typical valve spreadand its variation with changes in the casing pressure.
Pressure (psi)
Throttling flow region
Gas
inje
ctio
n ra
te (
MM
scf/d
)
Critical flow through a square edged orifice (same diameter as choke installed in gas lift valve)
Sub - criticalflow regionGas lift valve performance with
casing pressures (b), (c), (e) and (d)
Pcasing
P casing (b)
P casing (d)P casing (c)
Valve spread at various values of Pcasing
(e)
(a)(d)(c)(b)
7.2.1 Temperature CorrectionA compressible fluid (gas) is required for charging the dome (to avoid valve ruptureas the fluid heats up when it is run into the well or the well heats up when it is placedon production). Nitrogen is used for this purpose since it is non-corrosive, inflammableand the temperature effect on the pressure is well known.
P2 = P
1 * T
c
and TcTT
= + −+ −
{ . * ( )}{ . * ( )}1 0 00215 601 0 00215 60
2
1
where P1 = gas Pressure at Temperature T
1 (oR),
P2 = gas Pressure at Temperature T
2 (oR),
Tc = Temperature Correction factor
In practice, part of the dome volume is filled with a silicone liquid to dampenvibrations created by the gas flow through the valve. This silicone liquid occupies partof the volume of the dome, reducing the effective volume of the nitrogen charge.However, it will also expand as the valve heats up, giving an additional pressureincrease to that calculated above. This secondary correction becomes less importantfor the larger (1.5") valves where the volume of liquid compared to the gas volume isrelatively less important.
Figure 21
Valve performance
32
7.3 Valve Flow PerformanceThe operating valve at the bottom of the gas lift string, through which gas will becontinually passed, is normally equipped with square edged, orifice choke. Theresulting flow performance is shown in figure 21, curve (a). The gas flow rate passingthrough the choke increases with increasing pressure difference between the casingand tubing until a maximum value is reached when the critical flow rate is achieved.At this point the gas flow velocity has become supersonic and the volume of gas passeddoes not increase further with increasing pressure differential. The valve is oftenreferred to as being “choked” at this point. Orifice flow is described by the “Thornhill-Craver equation” which relates flow capacity to the difference between injection gasand tubing pressure at the valve depth and the port (orifice) size (see chapter 3.13). Theorifice size determines the maximum or “choked” volume of gas that can be injectedto aid lifting the well fluids to the surface.
This equation is often (incorrectly) used to describe the gas flow through an IPO orTPO gas lift valve. The Thornhill-Craver equation assumes that the flow through thevalve’s port is unimpeded by the presence of the ball, i.e. it assumes the valve is fullyopen and the ball does not impede the passage of gas. (Dynamic) Valve Response tests(see below) have shown that the Thornhill-Craver equation can overpredict the actualflow rate by more than 200%.
Further, simple gas lift design procedures assume that the valve will shut immediatelythe casing (IPO valve) or tubing (TPO valve) pressure drops below the preset value.This “instantaneous-closure” assumption is not correct, creating an even greater errorwhen Proportional Response Valves (see section 3.7.4) are used.
7.3.1 Dynamic Valve PerformanceDynamic testing of the valve, in which the injection gas flow rate is measured for afixed casing pressure and a range of tubing pressures, is required to properlyunderstand the valve’s performance. These dynamic valve performance characteristicsdescribe the:
(i) valve’s ability to start passing gas when the ball first lifts from its seat,
(ii) rate of increase in the gas flow rate as this clearance increases (flow capacity),
(iii) corresponding decrease in gas flow rate as the valve closes and
(iv) resistance to vibration under flowing conditions.
The valve's load rate (measured in units of psi/inch) is the pressure differencerequired to move the stem a given distance; while the valve spread is the differencebetween the opening and closing pressures. The shape of the performance curverepresents the valve’s sensitivity to pressure.
The American Petroleum Institute has published a Recommended Practice (NumberIIV2 or API RP IIV2) which describes a standardised testing procedure to measure aparticular gas lift valves dynamic performance. The standard is summarised as follows:
Department of Petroleum Engineering, Heriot-Watt University 33
33Gas Lift
The valve is installed in the test equipment after being set up with a suitable domepressure using the manufacturer’s recommended procedure. Measurements are madeof the:
(i) valve opening pressure,
(ii) pressure increase required to move the stem to its value of maximum travel and
(iii) position of the stem for various intermediate pressures.
The maximum effective travel distance influences the volume of gas that the valve canpass. The maximum effective stem travel (test one) needs to be measured only oncefor each valve design.
The flow capacity of the valve (as a function of stem position) is tested in a separatetest designed to measure the valve’s flow coefficient (or C
v). In this test the stem is
adjusted to various positions between 5% and 100% of its maximum value and theflow rate measured at 5 differential pressures - one of which should display chokingflow (see above). This flow coefficient test and the dynamic test (see below) need tobe carried out for each size of choke (or port) that are planned for installation in thevalve.
Pressure (psi)
Throttling flow region
Gas
inje
ctio
n ra
te (
MM
scf/d
)
Critical flow through a square edged orifice (same diameter as choke installed in gas lift valve)
Sub - criticalflow regionGas lift valve performance with
casing pressures (b), (c), (e) and (d)
Pcasing
P casing (b)
P casing (d)P casing (c)
Valve spread at various values of Pcasing
(e)
(a)(d)(c)(b)
A final test is performed for several set pressures in which the injection (or tubing)pressure is slowly decreased (or increased). The flow rate is measured at a numberof pressures until the valve closes. Either procedure should give the same result. Theresults of such a testing scheme is illustrated in figure 21. Here, the valve’s flow ratehas been measured at four casing pressures {P
casing (b), (c), (d) and (e)} and a variety of tubing
pressures. N.B. A fixed port size and dome pressure was used for all these tests.
Curve (b) shows the performance with the casing pressure adjusted to Pcasing(b)
. The gasflow rate is initially zero, even though the valve is open, since the tubing pressure isset to the same value. Gas starts to flow into the tubing when the tubing pressure isreduced, though the flow rate will not increases so rapidly as recorded for curve (a)- the equivalent test with a square edged orifice.
Figure 21
Valve performance
34
N.B. Pcasing(a)
and Pcasing(b)
have the same numerical value in figurer 21. It can be seenthat the gas flow rate passed by the valve {curve(b)} initially increases at a slightlylower rate than that for the orifice {curve (a)} of the same diameter as the gas liftvalve’s port. This difference is due to the ball, positioned slightly downstream of thevalve’s port (see figure 19), interfering with the gas flow pattern through the valve.
Further reductions in tubing pressure will increase the gas flow rate until a maximumis reached. After this point, the valve’s closing force (F
C) becomes relatively greater
than the opening force (FO); allowing the ball to move closer to its seat and reduce the
gas flow. This gas flow reduction occurs in a (relatively) linear manner (the “throttlingregion”) until it has decreased to zero at P
tubing(b). The slope of the gas flow rate / tubing
pressure plot in the throttling region is a function of the gas lift valve’s mechanicaldesign.
The dynamic valve performance at a slightly lower casing pressure {Pcasing(c)
} has asimilar shape (determined by the gas lift valve’s mechanical design). The maximumgas flow rate and the corresponding valve “spread” are now smaller since the valveopening force (F
O) has been reduced by the lower value of the casing pressure.
This process continues further when the casing pressure is reduced to Pcasing(d)
. Hereonly a small gas flow rate is recorded over a small range of tubing pressures. Oncethe casing pressure is set equal to, or lower than, P
casing(e), the valve will no longer open
and gas can not flow into the tubing.
N.B. It will be discussed later (13.4) that each gas lift valves in a completion stringshould be set up with an opening pressure slightly lower (typically 50 psi) than thesetting of the valve above it. This progressive lowering of the setting pressure for thedeeper valves aids the closure of the upper valves so that the gas injection isprogressively transferred to lower valves during the unloading process. It should benoted here that IPO valves do not control the annulus pressure. A drop in the annularpressure can only occur when the gas injection rate, possibly through both valves, isgreater than gas flow rate into the casing annulus at the surface.
7.3.2 Valve Performance Flow ModelSophisticated, computerised gas lift design programs require quantitative modelswhich describe the above flow tests. The Thornhill-Craver model {figure 21, curvea}, the previous industry standard, describes the flow through orifices reasonable wellbut does not capture the throttling action of “real” gas lift valves. (The originalpublication proposed the equation to describe flow through surface chokes with fixedbeans!) An extension of this model, by Winkler and Eads, uses the same formula butsubstitutes the actual choke or port area open to flow rather than the fully open valveused by Thornhill-Craver. Despite using a constant discharge coefficient for all stempositions, the flow curves generated by the Winkler-Eads model have the correctshape {figure 21(b) et seq.}. It works well (accuracy 20-30% over the full pressurerange) for valves with a port size of less than 0.25", though the errors rise to >100%when ports >0.25" are installed.
Both the Thornhill-Craver and the Winkler-Eads equations have an idealised, mecha-nistic basis - the equations are not “tuned” using the actual flow test data measured in
Department of Petroleum Engineering, Heriot-Watt University 35
33Gas Lift
the API test procedure described above. These gas flow rate measurements can berepresented by:
(i) a Flow Coefficient (Cv). This is a measure of the gas flow rate as a function of
pressure differential across the valve
(ii) the distance of the stem from its seat.
The Flow Coefficient (Cv) also determines the pressure rates at which choke flow
occurs across the valve as a function of the stem position. This is an important factorsince most valves show choke flow when the gas injection is being transferred to thenext lower valve. A higher value of C
v implies a greater valve flow capacity. It’s value
is determined by the:
(i) valve inlet port location with respect to the seat,
(ii) actual profile of the seat or valve,
(iii) design of the cross over ports and
(iv) any flow restrictions downstream of the port e.g. the check valve.
The measured Cv values are useable over a wide range of temperature and pressure
conditions, as well as for both liquid and gas flow. They can be used for all valves ofthe same type and port size. However, any design change which alters the flow paththrough the valve will effect the value of C
v and require that the flow test be repeated.
API RP IIV2 recommends a simplified approach to analysing this flow data using Cvvalues. Alternative approaches have been proposed by:
(i) Tulsa University Artificial Lift Project (TUALP) using a statistical approach todata analysis. This can only be applied over the pressure range used for the flow tests.
(ii) Valve Performance Cleaning-house (VPC). The VPC method uses a mechanisticmodel based on the force balance equation and load rate to calculate a value of C
v;
which is then adjusted so as to be able to reproduce the dynamic test data.
36
7.4 Proportional Response ValvesThe design of a Proportional Response Valve (see figure 22) involves:
Pdome
Abellows
Aport
Plug (Removed to charge dome with nitrogen)
Nitrogen Charged Dome
Bellows
Ptubing
PcasingStem
"Stop" to prevent excessivebellow collapse
Large Ball
PortTaperedSeatChoke
Check Valve(prevents back flow of produced
fluids from tubing into valve)
Chevron seals(form seal against polished borein body of side pocket mandrel)
Spring(To provideproportionalresponse)
(i) The addition of a spring to counterbalance the force trying to expand the bellows,
(ii) A “stop” to prevent excessive bellows collapse,
(iii) A larger ball at the tip of the stem and
(iv) A tapered seat instead of a square edged seat.
The proportional response valve has more parts than the “traditional” valve. It usesmore “O” rings, which are not only expensive but also prone to failure i.e. ProportionalResponse Valves should only be specified when definite operational advantages havebeen identified.
7.5 Dynamic Valve Response and Gas Lift Completion ModelingSimple (manual) gas lift completion design programs assume that the valves will openfully at (casing or tubing) pressures slightly greater than the set value and willcompletely close when the pressure drops below this value. They assume that the well“jumps” immediately from steady state operation when lifting through one (fullyopen) valve to a second, steady state condition when lifting through the next deepervalve. In practice:
(i) the volume of gas injected through a given valve will vary with time.
(ii) the well will only reach steady state inflow a considerable period of time afterthe unloading process has finished.
Figure 22
Proportional response gas
lift valve
Department of Petroleum Engineering, Heriot-Watt University 37
33Gas Lift
(iii) the combination of changes in (i) and (ii) will result in continuous changes inthe well’s Gas Liquid Ratio (GLR) during the unloading process. This is especiallytrue when any fluid lost to the formation during the workover or the early stages ofwell unloading has to be produced first before live formation fluids can enter the welland improve the well outfllow.
(iv) there is a significant volume of gas in the casing/tubing annulus. Pressurechanges will not occur instantaneously due to the limited gas supply rate at thesurface or the finite gas flow rate through a particular valve.
(v) the valves can open and close a number of times during the unloading of a multi-valve completion string e.g. as well temperature changes.
(vi) the temperature changes during the unloading process (being neither thegeothermal nor the steady state flowing temperature gradient).
It can be appreciated that a transient rather than a steady state computer program isrequired to simulate the unloading process. Such programs must accurately predictthe development of the above parameters as a function of time. The output of theprogram will include, as a function of time, the:
(i) tubing and annulus pressure at valve depths as well as surface casing (annulus)and tubing head pressure,
(ii) flow rate of liquid and gas through open valves,
(iii) depth of liquid levels in tubing and annulus,
(iv) volume of fluids (gas and liquid) produced to the surface and
(v) inflow from the reservoir (after making allowance for any liquid lost to theformation, as discussed above).
One commercially available program is DynaliftTM marketed by Edinburgh PetroleumSystems. It has been shown to be useful for troubleshooting problematic wells e.g. theprediction of when the following difficulties can be expected during well unloading- erosion of (gas lift) valves, severe slugging, excessive well stabilisation times etc.
7.6 Well StabilityDynamic Well Simulation programs can also be used to evaluate potential causes oflift problems once the well has been placed on production. These valve problemsmanifest themselves by the:
(i) well not producing or producing at lower than expected rate,
(ii) unstable well behaviour (casing and/or tubing head pressures oscillate regularly)and
(iii) excessive unloading and well stabilisation times.
38
The quickest method to correct the problem is to identify which valve is the cause ofthe production difficulty and then replace it with a correctly calibrated replacement.The alternative, to replace the gas-lift valves in a random, ad hoc manner is often muchmore time consuming and may never solve the problem.
Comparison of the actual and the predicted well behavior allows identification of thetype of problem being experienced as well as which gas lift valves is the cause of theimproper operation. Typical problems which have to be diagnosed include:
(i) valves that have become either:
(a) (partially) plugged or
(b) enlarged (cut out)
(ii) valves set to:
(a) incorrect dome pressures or
(b) operating incorrectly due to mechanical problems and
(iii) multi-pointing (injection through more than one valve at the same time). Thisis often due to an upper valve opening and closing.
An example of a specific operational problem that could be diagnosed is the case whenthe surface gas injection rate is greater than the single valve flow rate calculated fromthe annular and tubing pressures. Possible causes include:
(a) gas injection through two or more valves,
(b) the valve choke (or port) of the operating valve was enlarged (cut out) duringunloading or
(c) an upper valve is opening and closing.
A quantitative evaluation of the flow rates and the tubing and casing pressure valuesand their stability are required to differentiate between the above causes, as well as toidentify the specific problem valve.
8. GAS LIFT DESIGN PROCEDURES
This chapter will discuss the procedure used to design the gas lifted well. It willcombine the flow performance of operating valves discussed above with the conceptsdiscussed in Section 1 on multiphase flow and Nodal Analysis. Table 4 lists the welldata required to initiate this design process which will enable us to specify theoptimum:
Department of Petroleum Engineering, Heriot-Watt University 39
33Gas Lift
Tubing Outside Diameter 4.5 in
Tubing weight 12.6 lb/ft
Deviation Survey Vertical Well
Target Production Rate 5000 bfpd
Watercut 60% vol.
Produced Water Density 1.06 g/cm3
Gas Relative Density 0.65
Packer Setting Depth 9,500ft.
Mid Perforation Depth 10,000ft.
Wellhead Flowing Pressure 100 psig
Shut In Bottom Hole Pressure 3,700 psig
Productivity Index 0.50 STB/d/psi
Oil Gravity 37 API
Bubble Point 1200 psig
Gas Oil Ratio 400
Gas Injection (Casing) Pressure 1500 psig
Gas Available For Injection 3 MM scf/d
Ambient Temperature at Surface 70 F
Flowing Temperature at Surface 140 F
Temperature at Mid Perforation 240 F
Kill Fluid Gradient 0.45 psi/ft. *
* In Annulus / Tubing for Unloading Calculations
(i) Tubing size,
(ii) Injection gas supply parameters (pressure and volume),
(iii) Installation depth of operating valve and
(iv) Separator pressure.
The first item is to check that gas lift valves can be placed at suitable distances abovethe operating valve to ensure that the well can be unloaded using the gas lift parameterschosen (see section 3.5). The gas lift design will have to be changed if it is found thatthe well cannot be unloaded. Further, the design will be influenced by the gas liftdesign philosophy. For example:
(i) A Maximum Production / Expected Case Design would:
(a) maximise production by injecting gas as deep as possible,
(b) install extra gas lift valves around the operating valve to allow the depthof the operating valve to be adjusted and
(c) budget for (some) well entries to be required since the well may stopproducing if well inflow conditions change.
(ii) Worst Case / Robust Design would be designed to:
Table 4
Data set for gas lift design
40
(a) always work by placing the operating valve at a relatively shallow depthcompared to that normally achieved with to the available gas lift pressure.
(b) results in a lower production rate. Deferred production probably alsoimplies a reduced reserve recovery.
The philosophy chosen for a particular well / field will depend on the operationalconditions and the cost scenario.
8.1 An Example Design - optimising the performance of a gas lifted wellThe first requirement is to study the well’s in- and out-flow and confirm that artificiallift is required. This is illustrated in figure 23. It can be seen that the well productionrate will decrease rapidly once well Edinburgh-1 begins to produce water. It will stopflowing when the water cut reaches 30%. Any reduction in the flowing sand facepressure due to depletion of the reservoir pressure and / or the development of a nearwellbore “skin” due to Formation Damage will reduce the well production rate. Thiswill also reduce the critical water cut at which natural flow ceases. It should beremembered that the form of the Inflow Performance Relationship used to describefluid inflow from the formation (e.g. Vogel or “straight line PI”) will have an impacton these calculations.
Well production rate
San
dfac
e flo
win
g pr
essu
re
Reservoir inflow at various water cuts
60%40%20%
80%
0%
Tubing outflow at various water cuts 80%
0%
We will assume that Gas Lift has already been identified as the most suitable form ofartificial lift for this well and an operating valve has been installed at a suitable depth.The impact of the injection of varying amounts of gas on the well production rate isevaluated in Figure 24. Figure 25, a plot of production performance against lift gasinjection rate can now be derived. It can be seen that the injection of the first 1.25MMscf/d of lift gas leads to a rapid increase in production rate. Further increases inthe rate of gas injection yields ever reducing benefits until a maximum production rateis reached at 2.5 MMscf/d. The same data are plotted in Figure 26. Here the lift gasinjection rate is represented as a Gas-Liquid Ratio - the sum of the Produced fluid gasliquid ratio and the gas injection rate divided by the (gross) liquid production rate. Asexpected, the graph shows a maximum production at the 2.5 MMscf/d gas injection rate.
Figure 23
Performance well
Edinburgh - 1 at varying
water cuts
Department of Petroleum Engineering, Heriot-Watt University 41
33Gas Lift
Well production rate
Sandfa
ce flo
win
g p
ress
ure
Natural flow
ReservoirInflow
All at40%Watercut0.5
1.25
3.75
2.5
Lift gas injection rate (MM scf/d)
Lift gas injection rate (MMscf/d)
Well
pro
duct
ion r
ate
0 1.25 2.5 3.75
3.75 MM scf/d
2.5 MM scf/d
1.25 MM scf/d
0.5 MMM scf/d Maximum production rate
Well production rate
Gas
- L
iqui
d R
atio
g
(Technical) Optimum gasinjection rate
Formation Damage - or the presence of a near wellbore skin - will reduce the wellproductivity. Figure 27 is similar to Figure 24, except that the reservoir inflow froma damaged well (skin = +6) has also been plotted. Figure 28 not only illustrates thisloss in well productivity due to the presence of the skin, but also indicate that theconclusion drawn as to the optimum gas injection rate is unchanged.
Figure 24
Sensitivity study for tubing
outflow to lift gas injection
rate at 40% water cut
Figure 25
Production performance
and gas lift injection rate
Figure 26
Production performance
against gas liquid ratio
42
Well production rate
San
dfac
e flo
win
g pr
essu
re
Natural flow
ReservoirInflowskin = 0
40%Watercut
j
0.5
1.25
3.75
2.5
Reservoirinflowskin = +6
x
x
x
x
x
x
Lift gas injection rate (MM scf/d)
Lift gas injection rate (MMscf/d)
Wel
l pro
duct
ion
rate
0 1.25 2.5 3.75
Skin = 0
Skin = +6
Figure 29 examines the effect of changing the depth of lift gas injection. A deepinjection point creates a greater drawdown leading to an increased production rate anda more efficient GUF (Gas Utilisation Factor or Gross production rate / Gas InjectionRate).
Lift gas injection rate (MMscf/d)
Wel
l pro
duct
ion
rate
0 1.25 2.5 3.75
Incr
easi
ng d
epth
of i
njec
tion
Figures 30 and 31 examine the sensitivity of the well production rate to an increase inthe diameter of the production tubing. The production increases rapidly as the tubingOD increases to 4.5 in, but then decreases for the largest tubing (5.5 in). Excessiveslip between the gas and liquid phases occurs at the lower flow velocity associatedwith the larger tubing.
Figure 28
Production performance
and lift gas injection rate
for an impaired well
Figure 29
Production performance
and lift gas injection depth
Figure 27
Sensitivity study for tubing
outflow to lift gas injection
rate at 40% water cut for
an impaired well
Department of Petroleum Engineering, Heriot-Watt University 43
33Gas Lift
San
dfac
e flo
win
gpre
ssur
e
Well production rate
2.875" Tubing OD
Well inflow at40% Water cut
3.5"4.5"
5.5"
Wel
l pro
duct
ion
rate
Tubing internal diameter (in)2 2.5 3 3.5 4 4.5 5
3.5"OD
4.5"OD 5.5"OD
2.875"OD
2 2.5 3 3.5 4 4.5 52 2.5 3 3.5 4 4.5 5
We showed in Chapter 1 (Well Performance) that separator pressure affected theoutflow performance of naturally flowing wells. The same is true for gas lifted wells- Figure 32 show how the well outflow performance improves as the separatorpressure is decreased. This is shown explicitly in Figure 33, where the production rateis plotted as a function of separator pressure. In general, the lower the separatorpressure, the higher the production rate.
San
dfac
e flo
win
g pr
essu
re
Well production rate
Separator pressure (psi)
Well inflow at40% Water cut
250200150100
5025
75
Figure 30
Sensitivity study of tubing
outflow at various to tubing
diameters
Figure 31
Production performance
and tubing diameter
Figure 32
Sensitivity study of tubing
outflow to separator
pressure
44
Wel
l pro
duct
ion
rate
Separator pressure (psi)0 50 100 150 200 250
It should be emphasised that, in addition to the separator pressure, a back pressure onthe well performance is created by pressure losses across:
(i) Valves (should be fully opening type to minimise any pressure losses)
(ii) Choke body and/or choke nipples, if installed
(iii) The flowline (particularly pipe elbows) and long narrow diameter, tortuousflow lines with the separator mounted at a higher elevation than the wellheadcause an increased back pressure.
(iv) The wellhead itself (a Y block design will have a reduced pressure losscompared to the standard T block for a high rate well)
A further parameter that needs to be optimised is the lift gas injection pressure - Figure34 illustrates the effect this has on the well outflow performance while Figure 35records the production rate with changes in injection gas pressure.
Sandfa
ce flo
win
g p
ress
ure
Well production rate
Well inflow at
40% Water cut
Gas injection lift pressure
400
600
800
1000
12001500
Figure 34
Sensitivity study of tubing
outflow curves to gas
injection pressure
Figure 33
Production performance
and separator pressure
Department of Petroleum Engineering, Heriot-Watt University 45
33Gas Lift
W
ell
pro
duct
ion r
ate
Gas injection pressure (psi)500 750 1000 1250 1500
N.B. This well is capable of production with low injection gas pressures because thedepth of the operating valve was not too deep and the well was almost capable ofproducing under natural flow.
The conclusions reached from the above sensitivity studies are valid for the Edinburgh-1 well producing under the chosen combination of Reservoir Pressure and waterproduction studied. The conclusions need to be tested for robustness by repeating thesensitivity analysis for combinations of water cut and reservoir pressures that areexpected to be encountered during the well’s lifetime. Planned recompletions e.g.extension of the perforated interval, should also be included in this. Figure 36 is atypical example of the deterioration of the Reservoir Inflow Performance at varioustimes (T
1, T
2 and T
3) under the twin influences of the decrease of reservoir pressure
and increasing water cut. The tubing (outflow) performance is plotted for theoptimum gas injection rate that maximises the well production for each reservoirinflow curve i.e. the optimum gas injection rate changes with time. The GUF willdecrease with time due to the:
Well production rate
San
dfac
e flo
win
g pr
essu
re
T1
T3
T2 Reservoirdepletion Gas injection rate adjusted
for each reservoir inflow curvein order to maximise production
Reservoir inflow
x
x
x
(i) Increasing average liquid density (increasing water cut).
(ii) Decreasing (own) gas production rate (decreasing net oil production and somereservoir drive mechanisms e.g. solution gas) and
(iii) Reducing reservoir pressure.
Figure 35
Production performance
and gas injection pressure
Figure 36
Performance of well
Edinburgh - 1 at various
times
46
Figure 37 records the decreasing well production resulting from the changing wellinflow and tubing outflow performance curves, being derived from the data presentedin Figure 36.
Wel
l pro
duct
ion
rate
TimeT1 T2 T3
x
x
x
One further variable to be discussed is the choice of the multiphase flow correlationused to calculate the well outflow performance. This can have a great influence on theresults, as discussed in Chapter 1 (Well Performance). The flow correlation can bechosen on the:
(i) basis of production experience in other, similar wells in the same field,
(ii) theoretical considerations based on a comparison of the well type with theparameters originally used to generate the correlation or
(iii) comparison of predicted and actual well performance.
The latter is the best choice when (field) data is available. The actual and predictedgas lifted well performance need to be compared at a range of gas lift injection rates.{Chapter 1 discussed how the flow correlation for a naturally flowing well could bechosen by comparing the predicted and the measured pressure traverse (the pressure-depth profile).} Figure 38 illustrates how the use of an inappropriate correlation canlead to false expectations as far as the well production rate is concerned. Flowcorrelation number 4 refers to “Hagedorn and Brown”, probably the one mostfrequently used to describe the performance of gas lifted wells. The preferred flowcorrelation for the gas lifted well may be different from that previously used todescribe the same well’s performance under natural flow.
Figure 37
Performance of well
Edinburgh - 1 as the
reservoir pressure depletes
Department of Petroleum Engineering, Heriot-Watt University 47
33Gas Lift
Lift gas injection rate (MM scf/d)W
ell p
rodu
ctio
n ra
te
Actual well data
1
23
4,5
6
0 1.5 3 4.5
Flow correlation
Tubing Size 4.5"(O.D)
Separator Pressure As low As Possibleφ
Gas Injection Pressure 1000 or 1200 psi*
Gas Injection Rate 1MM scf/d
Flow Correlation Hagedorn and Brown
* Depends on detailed cost calculationsφ Some operators have profitably reduced back pressure on the
well by "twinning" the flow line to the separator
Table 5 summarises the conclusions from the above gas lift design exercise. Chapter10 of this module discusses the need to develop a well / field model to model currentand predict future production performance as well as to systematically monitor anydifferences between this prediction and the actual measured values. Such a monitoringprocess will gather the basic data needed to test whether the chosen inflow and outflowcorrelations are still appropriate. For example, table 6 illustrates the correlations usedby one operator to achieve the most accurate predictions based on such a monitoringprogramme.
Inflow < 50% Water Cut Vogel
>50% Water Cut "Straight Line" PI
Outflow < 10,000 bfpd Hagedorn and Brown
> 10,000 bfpd Duns and Ross
8.2 An Example Design - Gas Lift Unloading CalculationsThe section above describes how the operating valve location and the designobjectives for the gas lift system can be chosen. An illustrative description of theUnloading Process was given in section 3.5 and Figures 7 to 15. Concerns, to beavoided by careful design of the unloading process to be followed in a particular wellare that the:
(i) Well does not unload resulting in the gas being injected at a shallow depth. Thiscan be overcome by employing a conservative design philosophy to valve spacing.
(ii) Well does not lift at an optimal rate since the above conservative designphilosophy has resulted in the operating valve being situated at too shallow a depth.
Figure 38
Sensitivity of production
performance under gas
injection to choice of
multiphase flow correlation
Table 5
Summary of chosen gas lift
parameters
Table 6
Preferred correlations
48
This can be overcome by carrying out the gas lift design using the expected wellproperties. However, if an incorrect choice of well conditions are made, this mayresult in the well being unable to unload).
(iii) The well is stymied i.e. there is a combination of conditions under which all thegas lift valves close and no more gas can be injected.
These two approaches are compared in the worked example summarised in section 3.13.
8.3 Further Gas Lift System ConsiderationsTwo further points need to be evaluated.
(i) Flow Velocity. Corrosion / erosion of the tubing can become excessive if acritical flow velocity. Exceeding this limit will lead to excessive maintenance costse.g. early replacement of damaged tubing or accessories. This is particularly true ifsignificant quantities of formation solids e.g. sand are being co-produced with theformation fluids. One operator sets a maximum velocity of 12 ft/s in 3.5 in tubingfor these conditions. The American Petroleum Institute published recommendationsin this area tend to be somewhat conservative, one needs to compare them with localexperience. The critical value for the onset of corrosion / erosion can be verysensitive to the composition of the produced fluid; in particular the concentrationsof carbon dioxide and hydrogen sulphide in the gas phase.
(ii) Shear Rates. Exposing water and oil mixtures to high shear rates can lead tothe formation of emulsions. The produced fluid experiences increased velocitieswhen gas lift is introduced. This is due to the improved production rate and theincreased gas oil ratio. The increased shear rates can result in the formation ofviscous emulsions, especially when a choke or bean is included in the surface flowlines, though it is difficult to predict the precise conditions which will lead toemulsion formation. The well model will need to be adapted to the field data by“tuning” the fluid properties if emulsion formation is observed.
8.4 Further Gas Lift System Calculations
(i) Annulus Gas Pressure with Depth
The maximum depth at which gas can be injected into the tubing is controlled by thesurface (casing) gas pressure, the pressure drop across the gas lift valve and the tubingpressure gradient. Most lift gas systems operate at between 1000 and 1200 psi and arefilled with a natural gas with a high methane content. A useful “rule of thumb” is thatthe gas pressure gradient can be approximated to 30 psi/1000 ft. A more exact valuecan be calculated from the equation:
P PRD * L
TL Sgas
av
=
exp. * *53 34 Z
where: PL
= gas pressure at depth L (psia)P
S= surface gas pressure (psia)
L = true vertical depth (ft)
Department of Petroleum Engineering, Heriot-Watt University 49
33Gas Lift
RDgas
= gas relative density (to air)T
av= average temperature of gas column (°R)
Z = gas compressibility factor at Tav and P
av
Pav
= {PS + P
L}/2
(ii) Stored Gas Volume
A considerable volume of gas is stored in the casing/tubing annulus. It can beestimated from the gas equation:
V VP T
P Tannulusav
av
=
**
* *Z
where V = the gas volume at standard conditions (ft3)V
annulus= total annular volume (ft3)
Pav
= average gas pressure in annulus (psia)P = pressure at standard condition (14.3 psia)T = temperature at standard condition (460 °R)T
av= average temperature in annulus (°R)
and Vannulus
= 0.022 * L * {dic
2 - det
2}
where L = (along hole) length between well head and tubing packer (ft)d
ic= internal casing diameter (in)
det
= external tubing diameter (in)
(iii) Gas Compressor Power Requirements
The power required by the gas compressor which will supply gas to the lift system canbe estimated from
HHP 2.23*10 * QPP
4 out
in
=
−
−
0 2
1.
Where HHP = power required (hydraulic horse power)Q = gas injection rate (scf/d)P
out= compressor outlet pressure (psia)
Pin
= compressor inlet pressure (psia)
9. OPERATIONAL PROBLEMS
Typical problems associated with the operation of gas lifted wells include:
9.1 Gas QualityThis can lead to:
50
(i) Blockage of gas injection lines due to (solid) hydrate formation during periods oflow ambient temperatures. Such blockages are most easily removed by depressuris-ing the line since hydrates are unstable at low pressures, decomposing into theirconstituent components of gas and water. The problem can be avoided by properdrying of the gas so that water dew point is below the lowest expected ambienttemperature.
(ii) Loss of casing integrity due to use of corrosive lift gas. This is associated withthe presence of the acid gases (carbon dioxide and hydrogen sulphide) in the naturalgas. Guidelines are available to determine at which concentrations corrosion can beexpected. Often part of the produced gas will be sold, only sufficient gas beingretained in the lift gas system to “top-up” losses. In this case, all the gas is normallytreated to sales gas specification. Such specifications will typically limit the carbondioxide concentration to 2 or 3%, while hydrogen sulphide levels above 4 ppm areunacceptable.
In the case that none of the produced gas is sold, it may be practical to tolerate higherlevels of these acid gases. The resulting problems, from these more relaxed, gasquality specifications which will need to be resolved include:
(i) Natural gas containing carbon dioxide is only corrosive in the presence of liquidwater rather than gaseous water. Corrosion should be looked for in those pointswhere liquid water can collect (the bottom of pipelines, low points in undulatingpipelines etc.) when the partial carbon dioxide pressure exceeds 30 psi.N.B. The partial pressure of a gas corresponds to its concentration (expressed as amolar fraction) multiplied by the total system pressure.
Higher carbon dioxide partial pressures require either the specification of specialmetallurgy e.g. 13 chrome stainless steel, tubing or the continuous injection of acorrosion inhibitor. This corrosion inhibitor will passivate metal surfaces byforming a protective layer. High gas flow velocities (typically above 30-50 ft/s) willstrip the inhibitor from the metal surfaces and remove this protection.
(ii) Hydrogen sulphide partial pressures of greater than 0.05 psia lead to sulphidestress cracking of many of the steels used in completion systems and pipelines (inparticular, those with a Rockwell C hardness of greater than 22). A partial pressureof 0.05 psia hydrogen sulphide corresponds to approximately 50 ppm hydrogensulphide in a system operating at 1000 psi. Higher hydrogen sulphide levels can becoped with by specifying the appropriate metallurgy. Injection of corrosioninhibitors to passivate metal surfaces is possible on a temporary basis to protectagainst excess hydrogen sulphide levels. It is not normally recommended on apermanent basis since the sulphide stress cracking proceeds rapidly if inhibitoraddition is accidentally halted.
9.2 Solids These may either be produced from the formation {e.g. sand - see Chapter 7 (SandControl) or formed from the produced fluids {inorganic and organic scales - seeChapter 4 (Formation Damage)}. Wax is the most frequently encountered solid in gaslift operations}. Whatever the cause, the presence of solids within the production
Department of Petroleum Engineering, Heriot-Watt University 51
33Gas Lift
tubing can be expected to result in operational difficulties when carrying out wirelineoperations such as changing gas lift valves.
Excessive sand production can only be prevented by a suitable completion design e.g.installing a gravel pack. It should be repeated here that gas lift is relatively toleranttowards sand production compared to other forms of artificial lift, such as electricsubmersible pumps.
Recent developments in Production Chemistry has resulted in the development ofscale inhibitors which can be injected into the lift gas at the surface and transporteddown hole to be injected, via the operating valve, with the lift gas and subsequentlymixed with the stream of produced fluids.N.B. Scale inhibitors are chemicals which, at low concentrations, inhibit (delay) theformation of massive deposits of the material in question from the minute “seed”crystals. These seed crystals are formed when the concentration of the chemicalexceeds the chemical’s solubility limit. This occurs because of temperature andpressure reductions undergone by the formation fluids as it is produced to the surface.
9.3 Changes in Reservoir PerformanceReservoir Performance will change as the well/field matures. Typically, the water cut(and, consequently, the hydrostatic head of the fluid in the tubing) will increase andthe reservoir pressure will decrease. These factors, which reduce the well outflowperformance, may also be accompanied by a deterioration of the well inflowperformance due to the relative permeability effect brought about by the increasingproduced water fraction. Since the production target will be to maintain the same netoil (plateau) production for as long as possible, the well will be produced at a higherdrawdown with an increased gross fluid volume produced.
All these changes will require adjustment of the gas injection system parameters e.g.increase in the gas injection rate, moving the operating valve to a different depth(bracketing envelope concept - see section 13) etc.
9.4 Gas Supply ProblemsThese usually manifest themselves in one of two ways:
(i) The volume of lift gas available is less than that required to produce every gaslifted well at its (individual) optimum rate. It then becomes necessary to allocate thegas so that the total field production is maximised. This is discussed in section 10and an illustrative example studied in section 13.
(ii) Fluctuating compressor suction and discharge pressures. This is due to unstableoperation of the process plant, which may in turn be caused by unstable producingwell operation. For example, tubing “heading” leads to the production of a large“slug” of liquid followed by a period of gas only or no production at all (see Chapter9.6 for further details). The surface gas distribution network should also be designedto minimise the propagation of such pressure transients (see Chapter 4.3).
Profitability will be maximised by stabilising the entire production system (wells andsurface facilities) since this will maximise the reservoir fluid production and minimisethe maintenance costs associated with these fluctuating pressures.
52
9.5 Well Start-Up (Unloading)As mentioned previously, good practices are required to prevent fluid damage to thelower gas lift valves as the liquid in the casing / tubing annulus is unloaded whenstarting up a gas lifted. This is achieved by avoiding excessive fluid flow rates. Asuitable Start-Up procedure, as recommended by the American Petroleum Institute,is summarised as follows:
(i) Provision should be made to monitor the tubing and casing pressures andproduction rate (both liquid and gas) during the unloading process.
(ii) The tubing head pressure should be blown down to the separator flowline priorto starting the unloading process. Wells that produce under natural flow will be fittedwith a choke. This should be fully opened while undergoing the “kick-off” process.Continuous gas lifted wells which do not flow naturally normally do not have a chokeinstalled.
(iii) The Unloading Process is commenced under casing head pressure control (eithermanual or automatic). The casing gas injection rate is adjusted so that the pressureincreases at a rate of 5 psi/min until a pressure of 400 psi is reached. An increasedpressure rise rate of 10 psi/min can then be used until the system pressure is reached.In addition, the maximum, lift gas injection rate should be less than 50% of theplanned design rate. The well’s production rate can only be used to measure theunloading rate once the liquid level in the tubing reaches the surface. The progressof the unloading process can not be followed prior to this point in time, unlessarrangements are made to measure the fluid level in the annulus e.g. with anechometer.
(iv) There is often a drop in casing head pressure when gas is first injected into theupper gas lift valve and subsequent, lower lift valves. There will be a change in thegas injection rate, even if this drop in pressure is not observed.
(v) There is often a period of instability during well start up. Apart from the need toestablish a steady state flow regime within the well, there are considerable longerterm changes occurring during well unloading e.g. due to the well heating up. Thusthe pressure exerted by the gas lift valve’s nitrogen charge will increase as thetemperature increases. This can result in a gas lift valve opening and closing morethan once during well unloading.
The gas lift valve settings are normally designed using the operating conditions andwell temperature profile for the final, planned steady production conditions. There areseveral published correlations and many commercial programs available to predictthis equilibrium temperature profile. By contrast, there are few programs capable ofmaking meaningful dynamic predictions during the unloading process, particularlysince the well conditions during unloading tend to be relatively ill-defined.
It is good practice to check that the well unloading process can still proceed with valvepressure settings based on a cooler temperature profile than that predicted for theequilibrium producing conditions. Figure 39 illustrates this situation. It indicates howthe unloading calculations should be carried out using a progressively smaller
Department of Petroleum Engineering, Heriot-Watt University 53
33Gas Lift
difference between the geothermal and the equilibrium producing temperature profilefor the deeper gas lift valves. The suggested profile for a gas lift string with 8 valveswould show a temperature profile of {geothermal plus 12%, 25%, 37%, 50%, 62%,75%, 87% and 100% of the "equilibrium temperature value"} respectively.
Temperature
Dep
th
25%
50%
75%
100%
1st valve depth
4th Operating valve depth
PerforationsGeothermaltemperature profile
equilibrium flowingtemperature profile
Surface equilibrium flowing temperature for valvesetting
estimate of valve temperature forunloading calculations
2nd valve depth
3rd valve depth
It is clear that use of the reduced temperature profile is “safer” i.e. that the additionalsafety introduced by these measures increases the chance that the well will unloadsuccessfully. It is less likely to become “hung up” or “stymied” {see section 3.9.7 inthis chapter on trouble shooting}. This design can thus be viewed as a more robustdesign case. However, its use will normally result in somewhat lower well productionrate as the final lift configuration will be less optimum.
9.6 Well StabilityWell instability can take several forms.
(i) The Casing Head pressure remains constant and the Tubing Head pressure showssignificant, but regular, fluctuations. The size and frequency of these pressurefluctuations will depend on the particular combination of factors that are leading tothe well instability. An example of tubing heading is illustrated in figure 40. Thecasing head pressure is constant at 1000 psi while the tubing head pressure fluctuatesregularly between 60 and 180 psi with a cycle frequency time of 90-120 minutes.
Figure 39
Temperature profiles
during steady state flow and
well unloading
54
.eru
sser
Pgni
buTdnagnisaCfognidroce
Rne
Pni
w T
DATE ON
DATE ON
TIME
TIME
AM
AM
11PM
MID
NIG
HT
1AM
2AM
3AM
4AM5AM6AM
7AM
8AM
9AM
10A
M11
AM
MID
DAY
1PM
2PM
3PM
4PM5PM 6PM
7PM
8PM
9PM
10PM
Liquid flow rate 1300 stb/dFlow rate gas 630 Mscf/dWater cut 98%Cycle frequency 90 - 120 minGas injection rate 350 Mscf/dWell depth 3500ft Vertical well
Well detailsTubing Head Pressure : 60 - 180 psi
Casing Head Pressure : 1020psi
(ii) Both the Tubing Head and the Casing Head Pressure show regularly fluctuatingvalues.
Professor Asheim of Trondheim University defined two parameters (F1 and F2)which can be used to determine whether a stable flow condition exists in either thetubing or casing / tubing annulus respectively. He showed that unstable flowconditions have to exist in both flow conduits before the overall well behavior isunstable.
The tubing (F1) parameter considers whether the changes in tubing inflow and gas
injection pressure resulting from chance fluctuations in the tubing head pressure willtend to die away and return the well to the current producing conditions (the definitionof stable flow). Thus a stable well condition (F
1 > 1) requires that a momentary
decrease in the downhole pressure will result in a greater influx of (high density)formation fluids relative to that of (low density) injection gas. The combination ofthese extra inflows result in a relative increase in the average tubing density,counteracting the original, momentary decrease in downhole pressure. The well isthus flowing with a stable condition since it tends to return to the original operatingcondition. By contrast, unstable flow (F
1 < 1) would have resulted in the well moving
to a different operating point. This stability criterion is equivalent to that discussedin in the 'Well Performance' module for naturally flowing wells.
The casing / tubing annulus (F2) parameter reflects the response of the flow of casing
gas to momentary fluctuations in the Tubing Pressure opposite the operating valvewhen the tubing flow is potentially unstable (F
1 < 1). Thus the well is unstable (F
2 < 1)
if a momentary decrease in downhole tubing pressure causes the casing pressure todecrease more slowly than the tubing pressure. The resulting increased pressure dropacross the gas lift valve will lead to a greater inflow of lift gas, providing the valve isnot “choked”. This would tend to decrease the bottom hole pressure further - movingthe well even further from its previous operating point.
Figure 40
Example of tubing heading
Department of Petroleum Engineering, Heriot-Watt University 55
33Gas Lift
Well stability can be improved by instituting choke control of the casing gas injectionpressure. It is more difficult for the well to show unstable behavior when the pressuredrop across the surface choke controlling the gas injection rate is greater than thepressure drop across the downhole operating valve. This arrangement prevents thegas flowing out of the casing annulus and into the tubing at a faster rate than it is beingreplaced by the lift gas supply to the well. {The inclusion of any extra pressure dropin the gas lift supply system will often decrease the maximum well production rate.}
Practical considerations will often result that, if a well is suffering from unstablebehavior, an operator will find it easier to adjust the surface choke rather than thechange the size of the operating valve’s downhole orifice. The pressure drop acrossthe surface choke may thus become the greater pressure drop; resulting in greatertendency for annulus heading (the reverse of what was being attempted). Amomentary drop in tubing press can then result in an excessive gas flow rate throughthe downhole orifice. There will be a delay in the pressure reduction being felt at thesurface, due to the large annular volume and high gas compressibility. The lift gassupply will not increase quickly enough to compensate for this pressure reductionsufficiently, particularly when its supply is being impeded by a small surface choke.The pressure at the bottom of the annulus will continue to drop - eventually the fallingcasing pressure will reduce the gas supply to the tubing. Once again, there will be adelay in the response of the surface choke and it will continue to supply excess gas,leading to a build up in the annulus pressure. These pressure excursions can berepeated regularly - the size and frequency of the fluctuations depending on theparticular well circumstances.
One further cause of casing and tubing head pressure fluctuations is valve multi-pointing. This occurs when more than one valve is passing gas i.e. gas is being injectedinto the tubing at several different levels. This can represent an apparently stablesituation from the point of view of the Tubing and Casing Head pressures. Howeverfluctuations may also be induced by one valve repetitively opening and closing dueto the temperature changes induced by the changing flow conditions.
9.7 Dual Gas LiftThe most common dual gas lift configurations are two strings of 2.375in or 3.5 in ODtubing run inside 7.0 or 9.635 in OD casing respectively. Combinations of other sizesor even concentric strings are feasible, but rarely used.
The first problem with Dual Gas Lift completions is the well completion operationitself - there is insufficient width for the gas lift mandrels in the two strings to pass oneanother. This is overcome by designing the spacing of the mandrels in the two tubingstrings and the dual tubing running procedure so that the gas lift mandrels do not needto pass one another during the well completion operation.
A well with a Dual gas Lift Completion can be treated as a normal single completionif only one tubing string requires gas lift. The difficulty arises in ensuring a properdistribution of the lift gas between the strings if both strings are to be gas lifted i.e. onestring may take most of the gas while the second tubing is starved of gas.
One solution to prevent the valves in the different strings interfering with the operationof the other one is to install valves with significantly different operating characteristics
56
in the different strings. e.g. One string uses IPO valves while TPO valves are installedin the second string. The gas allocation between the strings is then controlled by theselection of the chokes installed in the operating valves. An alternative solution is toinstall TPO valves in both strings, since they do not respond strongly to the annuluspressure. However, practical experience shows that multipointing is a commonoccurrence for this design.
9.8 Trouble ShootingInefficient gas lift operation can be caused by:
(i) Inflow Problems due to:
(a) Incorrect size of choke (too large or too small),
(b) Incorrect lift gas (casing) pressure (too high or too low),
(c) Fluctuating compressor lift gas pressures and
(d) Incorrect lift gas flow rate (too high or too low).
(ii) Outlet Problems can include:
(a) Inadequate or restricted (due to partial plugging) gas flow through theoperating valve,
(b) Increased hydrostatic gradient across tubing due to increasing water cut andconsequent reduction in the produced fluid’s “natural” gross Gas Oil Ratio and
(c) The wellhead choke not having been fully opened or the separator pressurebeing set too high.
(iii) Downhole Problems. These include:
(a) A hole in the tubing and the
(b) Incorrect operation of valves due to either:
(i) Mechanical problems with the valve operation itself,
(ii) Wrong pressure setting of the gas lift valve’s nitrogen charge or
(iii) Incorrect gas lift valve spacing in the tubing string.
9.9 Trouble Shooting TechniquesOperational problems will be resolved most quickly by carrying out a systematicinvestigation to identify the fundamental cause of sub-optimum production from a gaslifted well. The analysis techniques used and the data gathered should include the:
(i) (Computer) modeling of annulus and tubing pressure profile.
Department of Petroleum Engineering, Heriot-Watt University 57
33Gas Lift
(iii) Comparison with a flowing temperature and pressure survey.
(iv) Carrying out a production test with multi gas injection rates.
(v) Performing of an analysis of the historical well test and production data toidentify trends.
(vi) Carrying out an echometer survey to identify the fluid level in the casing /tubing annulus. The echometer is an acoustic device which measures the transit timefor a signal generated at the surface to travel down the annulus and be reflected backto the surface from the liquid level in the casing / tubing annulus. This timemeasurement can be translated into a depth if the acoustic velocity is known.
9.10 Some Field Examples of operational problems:
(i) Injection pressure rises to its maximum value while the injection rate drops tozero. This indicates that an upper value closed before the lower valve opened i.e. thegas injection point has not being passed to successively deeper gas lift valves. Thewell is “stymied” or “hung up” and the setting on one or more gas injection valvesneeds to be changed.
(ii) Casing head pressure fluctuates by 5 to 10 psi but the gas injection rate isconstant. The injection point is being transferred between two valves (one form ofvalve multi- pointing).
(iii) Casing head pressure fluctuates by 10-20 psi while the injection rate is alsochanging. This suggests that a lower valve is opening and closing while a second,higher valve is continually open.
It should be remembered that it is always possible that the problem is more apparentthan real i.e. the accuracy of the flow meters (in particular) and pressure gauges needsto be verified. Once it has been confirmed that the flow and pressure measurementsare correct, then one can adjust the lift gas injection rate to the casing or increase theflowing tubing head pressure by closing the surface choke. These changes can rectifythe problem, though frequently the cause is a valve problem (mechanical damage orincorrect setting) which can only be solved by its replacement i.e. a wireline wellintervention is required.
10. FIELD PRODUCTION OPTIMISATION
We have already discussed how to maximise the production of gas lifted for a singlewell. This involved the development of (theoretical) (computer) model of the singlewell’s performance and its calibration against the actual well data via carrying outmulti-rate well tests. A calibrated model will allow those parameters controlled fromthe surface {injection gas choke setting, injection gas pressure and separator pressure}to be set to their optimum values. This process requires the systematic collection,verification, storage and analysis of the field data since the (computer) model of thewell can never be more accurate than the data against which the model was calibrated.
58
The individual well performance models for the wells in the field should now becombined to give a total field model so that the field performance can be maximisedagainst the relevant constraints. These constraints can either be:
(i) Physical constraints imposed by the overall system e.g. the total, available lift gasinjection rate, the volume of water that can be handled by the separation system, the(net) oil production from wells producing from different reservoirs may have to bemixed in a certain ratio to ensure that the specified crude oil quality is maintained.
(ii) Individual wells or reservoir specific constraints e.g. certain wells may have tobe produced at a specified minimum rate for reservoir management purposes.
The task is then to maximise the NPV operating cash income over the life of the wells.This normally equates to maximising the oil production by ranking each individualwell’s optimum production response to incremental gas injection. This concept isexplained in figure 41, where the net oil production (or revenue) is plotted against thelift gas injection rate (or production cost). A minimum (or “kick-off”) lift gas injectionrate, (V
1) is required to bring the well on production at the initial production rate Q
1.
The well performance curve is then plotted so that the incremental productionincrease, (∆Q
n) can be evaluated for equal increments in the lift gas injection rate,
(∆xn). It can be seen that the incremental GUF {Gas Utilisation Factor or (∆Q
n / ∆x
n)}
decreases as the lift gas injection rate increases. In fact, the GUF represents the slopeof the well performance curve (see Figure 41). This slope is also related to the factor:
{incremental (net oil) production revenue} {incremental (gas lift & other production) costs}
Maximumproduction
Net
oil
prod
uctio
n or
rev
enue
Lift - gas injection rate
Q1
V1
V1 "Kick - off" lift gas injection rate
Q1 Initial oil at "kick - off"
∆xn Equal lift gas rate increments
∆Qn Corresponding incremental oilproduction or revenue.
GUFn Gas Utilisation Factor
Economic limit
∆x1
∆x2∆Q1
∆x3∆Q2
∆x4∆Q3
∆x5∆Q4
∆x6
∆x7
∆x8∆x9 ∆x10
Economic limit where marginal(net production) revenue equalsmarginal lifting costs.
*
*
GUF1
GUF7
GUF4
This factor is plotted against the gas injection rate in figure 42. The economic limitand the technical maximum oil production are the points at which this factor equals1.0 and 0 respectively.
Figure 41
The incremental gas
utilisation factor decreases
with increasing gas
injection rate
Department of Petroleum Engineering, Heriot-Watt University 59
33Gas Lift
2.5
2.0
1.5
1.0
0.5
0
Lift gas injection rate
Incr
emen
tal (
Rev
enue
/ C
osts
)Economic limit where incrementalcosts = incremental revenue
x 1 x 2 x 3 x 4 x 5 x 6 x 7 x 8 x 9 x 1 0 x 1 1 x 1 2
Technical maximum production
"Kick - off" gasinjection rate
Lift gasincrements∆ ∆ ∆ ∆ ∆ ∆ ∆∆ ∆ ∆ ∆ ∆
Allocation of the available lift gas between wells is performed by comparing the GUFvalues. This calculation is complicated by the:
(i) Varying well “kick-off” gas requirements (figure 43 shows three wells exhibitingdifferent “kick-off” behavior).
(ii) Reservoir and well specific constraints, as well as any other production systemconstraints which also have to be honored.
(iii) The numbers of wells involved (large gathering systems can connect more than1000 production wells).
Net
oil
prod
uctio
n ra
te
Lift - gas injection rate
Nat
ural
flow
(wel
l a )
Well a
Well b
Well c
KOGcKOGb
Initial Gas Utilisation Factors
KOGb, KOGc = "Kick - off" lift - gasrequired by wells b and c.
Specialist software packages e.g. FieldfloTM, NetOptTM etc. are available from anumber of vendors which are capable of optimising such large systems. However, itmust always be borne in mind that the optimised production recommendation resultswill only be as good as the input field data. Once again, this points to the need to ensurethat a systematic data collection, validation and management system is installed at thesame time as such optimisation packages are purchased.
Figure 43
Well performance curves
for three wells
Figure 42
Gas lift incremental cost
curve.
60
Experience in many fields has shown that a sustained 2-4% increase in net oilproduction can be achieved by a dedicated engineering effort to provide continuousoptimisation of the gas lift systems performance. Large fields often require employingan engineer solely to manage gas lift and to obtain the commitment required from theoperations staff with respect to data gathering etc. This provides the informationrequired to overcome the typical gas lift operational problems summarised in figure44 i.e. so that deviations from optimum operation can be identified and rectified asquickly as possible.
OIL
WATER
GAS
Q o
il
Q gas
Stable flow
Unknown wellperformancecurve
Leakingunloading valves
Hole in tubing
Unstable injectiongas systempressure
(10) Unreliable test separator data(11) Low well test frequency
(9) Excessive surface pressure lossesOscillatingcontrol valve
Hydrateblockage
Valve multipointing
Incorrect orifice choke
OIL
(1)
(6)
(2)
(4)(3)
(5)
(7)
(8)
Q waterQ oil
Q gas
11. NEW TECHNOLOGY FOR CONTINUOUS FLOW GAS LIFT
It will have become apparent from the above that optimising the gas injection rate atthe operating valve is a key parameter in the control of a gas lifted well. Two newdevelopments in this area:
(a) A surface controlled gas lift valve where the required choke settings aretransmitted from surface via a cable or wireless transmission system. Two-wayinformation exchange can be implemented. Thus, annulus and tubing pressure andtemperature measurements at the gas lift valve depth as well as flow measurementscan be incorporated in the same instrumentation package which transmits data to thesurface. This allows the optimum choke setting to be specified.
(b) A redesigned choke that develops critical flow when the tubing/casing pressureratio is as high as 90% compared to the normal ratio of 55% (figure 45). This criticalflow condition decouples the pressure behavior in the annulus from that in the tubingas well as giving a constant gas injection rate for most producing conditions (thiscertainly simplifies modeling of the process of gas lift and operation!)
Figure 44
Problems for gas lift system
management
Department of Petroleum Engineering, Heriot-Watt University 61
33Gas Lift
Critical flow (Nova “) valves
Critical flow region Sub - criticalflow region
P tubing≈55% P casing
P tubing≈90% P casing
P casing
Tubing pressure (psi)
Conventional port valve
Gas
inje
ctio
n ra
te (
MM
scf/d
)
(c) Coiled Tubing based Gas Lift Completions in which gas lift valves and mandrelscan be incorporated as an integral part of a coiled tubing string. These can beemployed in two manners:
(a) To replace the conventional production tubing (Figure 46). This particularcompletion was designed to test an exploration or appraisal well.
1" gas lift valve in 2.375" mandrel
1.75" coiled tubing
Orifice valve in 2.375" mandrel
Locator / seal assembly
Polished bore receptacle
9.625" casing shoe
18.625" casing shoe
30" casing shoeSurface controlled subsurfacesafety valve (SCSSSV)
Producing formation
Perforations
7" liner shoe
Production to separator
Lift gas
Figure 45
Orifice valve performances
compared
Figure 46
Gas lift completion where
coiled tubing has replaced
the tubing string
62
(b) As a second production string within an existing tubing (Figure 47). This allowsgas lift to be introduced to aid well production without having to mobilise a drillingrig to recomplete the well. This does however require that a polished bore receptacle,into which the coiled tubing seal assembly can be located, has included in the originaltubing string. The advantage of this approach is that it is probably the cheapestmethods of installing gas lift in a well where the casing integrity has been lost.
Producing Formation
Perforations
Gas lift valve mountedin mandrel
Surface Controlled SubsurfaceSafety Valve ( SCSSSV)
Control line hanger
Coiled tubing hanger
Control line for SCSSSV
2" coiled tubing
Orifice
Locator seal assembly
Polished bore receptacle
Gas injection to tubing/ coiled tubing annuls
Production to separator
(c) Alternatively, the orifice valve can be omitted with the coiled tubing suspendedfrom its hanger - the gas injection taking place via the coiled tubing’s open end.
12. INTERMITTENT GAS LIFT
All the above has discussed continuous flow gas lift. The intrinsic flexibility of thegas lift concept allows it to be adapted to a wide range of situations; however it doesbecome inefficient at low formation fluid inflow rates (<150 bf/d with a 2.375 in ODtubing rising to < 300 bf/d for a 3.5 in OD tubing). A partial answer to the problemis intermittent lift where the gas is switched on for a short period of time at regularintervals :
Figure 47
A coiled tubing inner
production string
Department of Petroleum Engineering, Heriot-Watt University 63
33Gas Lift
(i) The formation fluid level in the tubing increases during the periods that the gaslift is switched off.
(ii) The formation fluid that has collected above the valve is lifted out during a periodof lift gas injection.
(iii) The flow of lift gas to the well is halted and the cycle is repeated.
A time cycle controller opens the surface lift-gas valve for a predetermined time atregular intervals and then shuts it. The process typically produces some 2-5 bbls liquidper cycle with a frequency of 1-3 cycles per hour (Figure 48). The valve spread (ordifference between pressure needed to open and close the valve - see section 3.7.2)controls the minimum amount of gas used during each intermittent gas lift cycle.
.eru
sser
Pgni
buTdnagnisaCfognidroceR
neP
niw T
DATE ON
DATE ON
TIME
TIME
AM
AM
5AM
6AM
7AM
8AM
9A
M
10AM
11AMMIDDAY 1PM
2PM
3PM
4PM
5PM
6PM
1000
800
600
400
300
200
100
500
700
900
400300
200100
5
1000
800
600
400300
200100
500
700
900
00 00 00 400
300
200
100
500
00010
00
800
600
400
300
200
100
500
700
900
1000
800
600
400
300
200
100
500
700
900
Wellhead Pressure
Casing Pressure
Cycle time 30 minInjection time 2 minProduction rate 144 bfpdInjection gas rate 350 M scf/d
Well details
This approach is more efficient in terms of the volume of gas required to lift a givenvolume of liquid than continuous gas lift for these low rate wells. However, it stillrequires relatively high gas volumes due to the fall back of the liquid (of the order of10% of the tubing volume) present in the tubing at the time when the gas is switchedoff. This inefficiency can be removed by incorporating a plunger to displace all theliquid to the surface, the plunger itself having been displaced up the tubing by injectinglift gas underneath {see figure 6(e)}. The process is as follows:
(i) The plunger displaces the fluids that have entered the well during the period thatthe lift gas was shut off.
(ii) The flow of lift gas is halted once the plunger reaches the wellhead and the cyclerepeated.
This process is also known as Plunger Assisted Intermittent Lift (PAIL) or chamberlift. PAIL has particular advantages when the well is suffering from severe waxdeposition since the regular passage of the plunger ensures the tubing remains clearof wax.
Figure 48
Intermittent gas lift
64
13. GRAPHICAL GAS LIFT DESIGN EXERCISE FOR WELLEDINBURGH-2
13.1 Introduction
The earlier section in this chapter have showed how the well production wasinfluenced by parameters such as the gas injection rate, tubing size, depth of gasinjection, and flowing tubing head pressure etc. This sensitivity study can be used todetermine the optimum gas lift completion design parameters. It is now necessary tospace out the unloading valves so that the gas can reach the operating valve at the depthshown above. The procedure used will be illustrated by using the Edinburgh-2 wellas an example (see Table 7 for well design conditions).
Gas injection pressure 1250 psig
Gas injection rate 500 M scf/d
"Kill" brine density 0.465 psi / ft
Depth mid perforations 10,000 ft
Reservoir pressure 3600 psi
(Gross fluid) Production index 0.5 bf/d/psi
Water cut 65% vol.
Oil density 35 API or 0.37 psi/ft
Production water density 1.05 g/cm3 or 0.455 psi/ft
( Average) produced fluid density 0.426 psi/ft
(relative) gas gravity 0.6
Flowing wellhead pressure 100 psi
Average flowing gradients: For producing well
250 bf/d at GLR of 2105 scf/b 0.07 psi/ft
500 bf/d at GLR of 1105 scf/b 0.1psi/ft
750 bf/d at GLR of 770 scf/b 0.157 psi/ft
800 bf/d at GLR of 730 scf/b 0.18 psi/ft
Table 7
Well Edinburgh - 2 gas lift
design conditions
Department of Petroleum Engineering, Heriot-Watt University 65
33Gas Lift
Pressure (psi)
Dep
th (
ft)
1000
2000
3000
4000
5000
6000
7000
8000
9000
10,000
500 1000 1500 2000 2500 3000 3500Pressure (psi)
500 1000 1500 2000 2500 3000 3500
Static (kill) brine gradient
0.465 psi / ft
Reservoirpressure3600 psi
2260 ft
1600 psi surface gas injection pressure
1200 surface psi gas injection pressure
800 psi surface gas injection pressure
500 psi surface gas injection pressure
Gas SG = 0.6Gas SG = 0.8
Equilibrium
level
kill brine
600
0.08
0.9 Gas S.G.
0.8 Gas S.G.
0.7 Gas S.G.
0.6 Gas S.G.
0.07
0.06
0.05
0.04
0.03
0.02
0.01700 800 900 1000 1100 1200 1300 1400 1500
Gas
Gra
dien
t, ps
i/ft.
Surface Gas Injection Pressure (psig)
T = 70ºF at Surface190ºF at 10000 ft TVD
Figure 49b
Initial conditions for
"killed" well
Figure 49a
Injection gas gradients for
well Edinburgh 2
66
13.2 Initial Conditions - the “Dead” WellThe well has been completed and circulated to a “kill” brine of density 0.465 psi/ft.Figure 49 shows that the reservoir pressure is only sufficient to support the fluid levelin tubing and tubing/casing annulus to a depth of 2258 ft. Gas pressure gradients forvarious gas injection pressures have been drawn. The effect of using gases of differentrelative densities (0.6 and 0.8) for the 1200 psi surface pressure case has also beenincluded.
13.3 Construction of the “Equilibrium Curve”The range of production rates achievable using gas lift and the required gas injectiondepths must be calculated first. A nodal analysis calculation is carried out for aproduction rate of 500 b/d using a gas lift injection point at depth D ft as the node andthe data from Table 7.
Calculation from Surface
Pressure at gas injection point = separator pressure + 0.1 D
Calculation from Reservoir:
Pressure at gas injection point = Reservoir Pressure - drawdown - 0.426* (10000 - D)
* We have assumed that the frictional pressure losses below the gas lift valve arenegligible at this low flow rate.
These two pressures are equal and thus :
D = 5400 ft and the gas injection pressure at this depth is 640 psi.
The latter corresponds to a surface gas pressure of 560 psi. The same calculation canbe carried out for gross production rates ranging from 250 to 800 b/d - see Table 8 andfigure 50. (Remember that the flowing pressure gradient has to be changed as well asthe drawdown for each calculation.) This table shows that gas injection pressuresranging from 350 psi to 1825 psi are required. It would not be realistic to use a constantgas density for calculating the surface pressure for this wide range of conditions. Theinjection gas pressure gradient is also dependent on the average pressure. Table 8includes an approximate value which allows us to estimate the surface gas injectionpressure required. A more accurate value could have been calculated using theequation given in section 8.4.
Production rate (bf/d) 250 500 750 800
Gas injection depth (ft) 3540 5400 8400 9594
Gas injection pressure at gas lift valve (psi) 348 640 1419 1827
Average gas gradient at this pressure (psi/ft) 0.01 0.015 0.03 0.045
Required surface gas injection pressure (psi) 313 560 1167 1395
Table 8
Calculation of "The
Equilibrium Curve"
Department of Petroleum Engineering, Heriot-Watt University 67
33Gas Lift
Pressure (psi)
Dep
th (
ft)1000
2000
3000
4000
5000
6000
7000
8000
9000
10,000
500 1000 1500 2000 2500 3000 3500Pressure (psi)
500 1000 1500 2000 2500 3000 3500
The "Equilibrium
Inflow / O
utflow C
urve"G
as gradients SG
= 0.6
313560
11671462
Surface gas injection pressures (psi)
Reservoirpressure800 bf/d
750 bf/d
500 bf/d
250 bf/d
Figure 50 summarises the producing conditions which simultaneously satisfy thereservoir inflow and tubing outflow conditions. This line is called the “EquilibriumCurve”. It allows a quick estimate of the depth, and associated pressures, at which theoperating valve must be installed to achieve a given production rate. As expected, thehighest production rates are achieved with the deepest gas injection. The available gaslift pressure is 1250 psi (Table 7) - hence we will set a target production rate of 750bf/d for the Edinburgh -2 gas lift completion by installing the operating valve at 8400ft. It now needs to be confirmed that the well can be unloaded when completion brine(of density 0.465 psi/ft) is present in the well.
Figure 50
Construction of "The
equilibrium curve"
68
Pressure (psi)
Dep
th (
ft)
1000
2000
3000
4000
5000
6000
7000
8000
9000
10,000
400 800 1200 1600 2000 2400 2800Pressure (psi)
Injection gas gradient
Surface gas injection pressure (1250 psi)
400 800 1200 1600 2000 2400 2800
Flowing wellhead pressure
100 psi
Operating valve at 8400 ft Produced fluid
(750 bpd) gradient
= 0.426 psi / ft1500 psi drawdownrequired to produce750 bf/d
Valvenumber1
Kill brine gradient
0.465 psi / ft
2530 ft
750 bpd produced fluid + 500 MM
scfd
injection gas = 0.15 psi / ft
"The Objective G
radient"
50 psi safety margin
13.4 The Unloading ProcessThe unloading calculation is begun by calculating the flowing bottom hole pressure(2100 psi) required to produce 750 bf/d. The pressure profile for the flowing well isthen constructed from this point to the surface wellhead pressure (100 psi) using theflowing fluid gradients before (0.426 psi/ft) and after gas injection (0.15 psi/ft). Thisline is known as the “Objective Gradient” (Figure 51). The flowing gradients havebeen represented here as straight lines to simplify the calculations. The intersectionof these two lines is the depth at which the operating valve will be installed (8400 ft).Also shown is the intersection between a (static) brine gradient drawn from the(flowing) wellhead pressure of 100 psi to the gas gradient line for a surface pressureof 1200 psi. This pressure is 50 psi less than the nominal gas lift system operatingpressure of 1250 psi. The lower value is as an additional safety margin to ensure thatthe gas lift completion will still unload despite fluctuations in the compressor output
Figure 51
Well unloading stage 1,
maximum depth of first gas
lift valve
Department of Petroleum Engineering, Heriot-Watt University 69
33Gas Lift
pressure. It also compensates for any frictional pressure loss due to gas flow acrossthe valve.
The intersection between the gas and brine gradient lines is at 2530 ft - this is the depthat which the first unloading valve should be placed (Figure 51). The surface gaspressure is sufficient to displace the brine from the annulus into the tubing and out ofthe well.
Pressure (psi)
Dep
th (
ft)
1000
2000
3000
4000
5000
6000
7000
8000
9000
10,000
400 800 1200 1600 2000 2400 2800Pressure (psi)
400 800 1200 1600 2000 2400 2800
100
Operating valve at 8400 ftProduced fluid
(750 bpd)
= 0.426 psi / ft
1500 psi drawdownrequired to produce750 bf/d
Valvenumber1
Injection gas gradient
Surface gas injection pressure (1250 psi)Flowing wellhead pressure
Valvenumber2 4485 ft
2530 ft
Brine gradient
reducing under
the influence
lift gas injection
Brine gradient
"The objective gradient"
Gas injection reduces densityof brine above first gas lift valve.
50 psi safety margin
The gas will now enter the tubing via the gas lift valve and reduce the pressure exertedby the fluid in the tubing above the valve. This reduced pressure also ensures that the
Figure 52
Well unloading stage 2,
depth selection for 2nd gas
lift valve
70
pressure in the annulus remains greater than that in the tubing at all depths. Hence fluidcontinues to be displaced from the annulus into the tubing through the second andlower valves. This process is illustrated in figure 52 where it is shown how the brinedensity is reduced under the influence of lift gas injection until it reaches the“Objective Gradient”. The depth (4485 ft) at which the completion brine gradientintersects the gas gradient line is the maximum depth at which the second unloadingvalve can be placed.
Pressure (psi)
Dep
th (
ft)
1000
2000
3000
4000
5000
6000
7000
8000
9000
10,000
400 800 1200 1600 2000 2400 2800Pressure (psi)
400 800 1200 1600 2000 2400 2800
Operating valveat 6650 ft
Produced fluid
gradient
1070 psi drawdownrequired to produce535 bf/d
Valvenumber1
Injection gas gradient
Surface gas injection pressure (1250 psi)Flowing wellhead pressure
50 psi safety margin
2530 ft
The objective gradient
2
3
4
56
100 psi safety margin
150 psi safety margin
200 psi safety margin250 psi safety margin
2530
2530 ft
1750 ft
1200 ft
750 ft
400 ft
200 ft}
100
Valve spacing
6850 ft6650 ft
6250 ft
5500 ft
4300 ft
50 psi safety margin
Classical design procedures assume that the upper valve shuts immediately the gasreaches a lower valve. Section 7.3, which described valve performance, showed that
Figure 53
Well unloading stage 3,
inclusion of safety margins
Department of Petroleum Engineering, Heriot-Watt University 71
33Gas Lift
this is not realistic. Practical experience has shown that closure of the first valve canbe made more certain by designing the second valve to accept gas at lower pressurethan the valve above it. Safety margins of up to 50 psi/valve are often employed.Comparison of figures 52 and 53 show how this 50 psi safety margin can be included,as well as why it reduces the setting depth of valve No. 2 from 4485 ft to 4300 ft.
The design process is continued in the same manner (see figure 53 and Table 9) withvalve setting depths of 5500 ft, 6250 ft, 6650 ft and 6850 ft being identified. The lattervalve spacing of only 200 ft is obviously impractical - in fact a minimum spacingdistance of 450 ft or 150 m is normally recommended. Hence an operating valve depthof 6650 ft is chosen. Placing the operating valve depth at 6650 ft implies, to a firstapproximation, that a flowing bottom hole pressure of 2530 psi and associatedreservoir productivity of 535 bf/d is achievable, rather than the target value of 750 bf/d.
Initial design Final design Initial design Final design
Valve Depth (ft) Valve Depth (ft) Valve Depth (ft) Valve Depth (ft)
No. No. No. No.
1 2550 1 2530 1 2650 1 2650
2 4300 2 4300
2 4690 2 4690
3 5500 3 5500
4 6250 4 6200 3 6190 3 6190
5 6650* 5 6650 *
6 7100 φ 4 7050
4 7250
7 7550 φ 5 7500
8 8000 φ 5 7970 6 7950
9 8450 φ 6 8400* 7 8400 *
10 8900 φ 8 8850 φ
11 9350 φ 9 9300 φ
12 9800 φ 10 9750 φ
* Operating valve
Bracketing envelopeφ Dummy valves
Expected rate conditions
(no safety margins)
Worst case conditions
(with safety margins)
Further it is normal to install at least one gas lift at the minimum spacing both aboveand below the planned depth for the operating valve. This is done because a higheror lower operating point may be required to adapt the well’s operation to conditionssomewhat different from the assumptions made during the design process and/orchanges that have occurred during the lifetime of the well. This zone with minimumvalve spacing is called the Bracketing Envelope. Some of the valve setting depths maybe modified when developing the bracketing envelope - as shown in Figure 54 andTable 9, columns 3 and 4.
Table 9
Unloading valves for well
Edinburgh - 2
72
Pressure (psi)
Dep
th (
ft)
1000
2000
3000
4000
5000
6000
7000
8000
9000
10,000
400 800 1200 1600 2000 2400 2800Pressure (psi)
400 800 1200 1600 2000 2400 2800
Actual operating valve
Produced fluid
1500 psi drawdownrequired to produce750 bf/d
Valvenumber1
Injection gas gradients
Surface gas injection pressure (1250 psi)Flowing wellhead pressure
50 psi safety margin
2530 ft
The objective gradient
2
3
4
5
6
100 psi safety margin
2530
7
8
9
10
11
12
Extendedbracketingenvelope.
150 psi safety margin6200 ft
6650 ft
7100 ft
Target operating valve depth
The bracketing envelope
1070 psi drawdownrequired to produce535 bf/d
100
4300 ft
5500 ft
50 psi safety margin
In addition, it is shown that the:
(i) bracketing envelope has been extended in the final design as far as the bottomof the well to allow deep gas lift if the reservoir pressure / well productivity decreasesufficiently. This worst case design requires 12 gas lift valves to be installed withthe operating valve placed at 6650 ft and dummy valves installed in the bottom 7mandrels.
(ii) 50 psi safety margin used between each succeeding valve has created a robustdeign with high confidence that the well will unload to a depth of 6650 ft.
(iii) depth for the operating valve is shallower than the target depth of 8400 ft. Thisreduces the well production to 535 bf/d, lower than the target value of 750 bf/d.
Figure 54
Final design: worst case
conditions
Department of Petroleum Engineering, Heriot-Watt University 73
33Gas Lift
By contrast, the target fluid production can be achieved if an “expected rateconditions” design is made (Figure 55 and Table 9). The final design shownincorporates the modifications made due to introduction of the “Bracketing Enve-lope” concept. It:
(i) uses two fewer gas lift valves.
(ii) achieves the target production rate of 750 bf/d.
(iii) depends for its success on the well conditions being exactly as prognosed. Thisis required to ensure that the unloading process works efficiently.
Pressure (psi)
Dep
th (
ft)
1000
2000
3000
4000
5000
6000
7000
8000
9000
10,000
400 800 1200 1600 2000 2400 2800Pressure (psi)
400 800 1200 1600 2000 2400 2800
1500 psi drawdownrequired to produce750 bf/d
Valvenumber1
Injection gas gradient
Surface gas injectionpressure (1250 psi)
Flowing wellhead pressure
4690 ft
2650 ft
2
3
4
5
6
6190 ft
7
8
9
Extendedbracketingenvelope.
7250 ft
9300 ft
9250 ft
7500 ft
8400 ftOperatingvalve
Recommended spacing
- 7050 ft
- 7950 ft
- 8400 ft
- 8850 ft
Finaldesign
"The Objective G
radient" for
750 bpd + 500 Msc f/d gas
7970 ft
2650 ft
2040 ft
1500 ft
1060 ft
720 ft
430 ft
Valve Spacing
The bracketingenvelope
Figure 55
Final design: expected rate
conditions
74
13.5 Gas Lift Optimisation Exercise
QuestionFigures 56 to 58 show the (gross) fluid production rate from three gas lifted wells. Thefollowing two exercises illustrate a procedure to manually allocate a limited volumeof gas (2 and 5 MM scf/d) so as to produce the maximum volume of oil.
6000
5600
5200
4800
4400
4000
3600
3200
2800 1 2 3 4 5 6
Gas injection rate (MM scf/d)
(Gro
ss fl
uid)
Pro
duct
ion
rate
(bf
/d)
Water cut = 25%
Figure 56
Gas lift performance - well 1
Department of Petroleum Engineering, Heriot-Watt University 75
33Gas Lift
1 2 3 4 5 6 Gas injection rate (MM scf/d)
(Gro
ss fl
uid)
Pro
duct
ion
rate
(bf
/d)
3200
2800
2400
2000
1600
1200
800
400
0
Water cut = 40%
1 2 3 4 5 6 Gas injection rate (MM scf/d)
(Gro
ss fl
uid)
Pro
duct
ion
rate
(bf
/d)
3200
2800
2400
2000
1600
1200
800
400
0
Gas Lift Performance - Well 3
Water cut = 45%
Figure 57
Gas lift performance - well 2
Figure 58
Gas lift performance - well 3
76
AnswerFigures 56 to 58 allow the incremental net oil production to be calculated for as thelift gas injection rate is increased in increments of 0.5 MM scf/d (see Figure 59 andTable 10). The simplest method is to allocate each (0.5 MM scf/d) increment of liftgas to the well which shows the highest net oil production. The process is thus repeatedfor the second and subsequent increments until all the available gas has been allocated.Table 10 shows the case when 5 MM scf/d of lift gas (ten increments of 0.5 MM scf/d) was available. 2.5 MM scf/d of lift gas has been allocated to well 2 to produce anadditional 1074 b/d of net oil production. A similar gas volume was allocated to Well3 and an additional 1188 bopd produced.
Lift Gas Allocation To Maximise Oil Production.
0.5 MM scf/d Well 1 Well 2 Well 3 Total incremental
Gas increment Incremental oil production (bo/d) oil production (bo/d)
1 228 228
2 228 456
3 228 684
4 222 906
5 194 1104
6 324 1428
7 252 1680
8 242 1922
9 176 2098
10 168 2266
Net oil production 0 1074 1188 2266
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
6000
5600
5200
4800
4400
4000
3600
3200
2800 1 2 3 4 5 6
Gas injection rate (MM scf/d)
(Gro
ss fl
uid)
Pro
duct
ion
rate
(bf
/d)
Well 1 (w
ater cut = 25%)
120 bfpd = 90 bopd
140 bfpd = 105 bopd
130 bfpd = 97.5 bopd
110 bfpd = 82.5 bopd
110 bfpd = 82.5 bopd
120 bfpd = 90 bopd
Figure 59a
Incremental net oil
production with increasing
lift gas volumes
Table 10
Lift gas allocation to
maximise oil production
Department of Petroleum Engineering, Heriot-Watt University 77
33Gas Lift
3200
2800
2400
2000
1600
1200
800
400
0 1 2 3 4 5 6
Gas injection rate (MM scf/d)
(Gro
ss fl
uid)
Pro
duct
ion
rate
(bf
/d)
Well
2 (w
ater
cut =
40%
)
380 bfpd = 228 bopd
380 bfpd = 228 bopd
380 bfpd = 228 bopd
370 bfpd = 222 bopd
280 bfpd = 168 bopd
240 bfpd = 144 bopd
1 2 3 4 5 6 Gas injection rate (MM scf/d)
(Gro
ss fl
uid)
Pro
duct
ion
rate
(bf
/d)
Well 3 (w
ater cut =
45%)
360 bfpd = 198 bopd
590 bfpd = 324.5 bopd
560 bfpd = 252 bopd
440 bfpd = 242 bopd
320 bfpd = 176 bopd
280 bfpd = 154 bopd
3200
2800
2400
2000
1600
1200
800
400
0
Figure 59b
Incremental net oil
production with increasing
lift gas volumes
Figure 59c
Incremental net oil
production with increasing
lift gas volumes
78
Table 10 shows that this calculation is not quite as simple as it appears due to the needfor a small volume of "Kick - off" gas to bring well 3 into production. This is illustratedwhen only 2 MM scf/d of lift gas is available. The simple approach used abovesuggests that it should all be allocated to Well 2 where it will recover an extra 906 bopd.However, allocating the same volume of gas to Well 3 will produce 1016 b/d of extranet oil (Table 11). This calculation illustrates why the slope of the curve (the GasUtilisation Factor) is a better optimisation factor rather than the simplistic incrementalproduction approach suggested above.
0.5 MMscfd Incremental oil production (bopd)
Gas increment Well 1 Well 2 Well 3
1 90 228 194
2 105 228 324
3 97.5 228 252
4 82.5 222 242
Total 375 906 1012
Commercial network simulation computer programs are essential to carry out thistype of optimisation calculation. There are many types of constraints that thesimulator has to honour while at the same time maximising net revenue. These caninclude:
(i) well constraints e.g. sand production,
(ii) gas/water coning or other reservoir constraints,
(iii) gas/water separation, compression or disposal facility constraints or bottle-necks,
(iv) production constraints, such as mechanical equipment failure, and
(v) export quality requirements.
Table 11
Allocation of 2MM scf/d of
lift gas
Department of Petroleum Engineering, Heriot-Watt University 79
33Gas Lift
14. FURTHER READING
(1) Beggs H. D.“Production Optimisation using Nodal Analysis”ISBN 0-930972-14-7published by Oil and Gas Consultants Inc., 1991.
(3) Economides M., Hill A. & Economides C.“Petroleum production Systems”ISBN 0-13-658683-Xpublished by Prentice Hall, 1994.
(4) Economides M. J., Watters L. and Dunn-Norman S.“Petroleum Well Construction”ISBN 0-471-96938-9Published by Wiley, 1998.
(5) Golan M. & Whitson C.“Well Performance” 2nd editionISBN 0-13-946609-6published by the Norwegian University of Science and Technology (NTNU), 1996.
(6) Mian M. A.“Petroleum Engineering Handbook for the Practicing Engineer”, Volume 2ISBN 0-87814-379-3Published by PennWell Books, 1992.
(7) API - EXPLORATION AND PRODUCTION DEPARTMENTAPI Gas Lift Manual. Book 6 of the Vocational Training Series, 3rd editionAPI, 1994
80
����������������� �� ��������� ��� ����� ��� ���� � ���������� !�"������ # ���� ���$���%�&�'%&!�(��$# ���� )��*+(�%�&�'%&,!�$��*+(�%�&#-��.�/� 0 1����10 1�����2� ��� ����//� 0 1����10 1�� 2��� /��� 3�%�& �3� 3"�4 ��%��%&" �$��3�%�&, 2����� /��� 3�%�&�3� 3"�0 �(53&%,� , �+�%&"��%��%&" 2���� /�+%*6�,, 2������ ��3 �5��%,�� 2���� /%����7�3'�/���3�%�&8 9:;:<8 =>?@ABC>8DECFGH 2�� /���3�%�&�3�3"��+�%&"����&�%&" 2��� /��� 3�%�& � 3� 3"� � +�%&" )��I��3�%&" 2�-� /���3�%�&�3�3"��+�%&")��*+(�%�& 2�-��� ��+�(�,�I/���3�%�&�3�3"�7/%&�, 0 �J���&� 2�-��� ��+�(�,�I/���3�%�&�3�3"�7 ���$���%�&3&*� ��'�J��/�+%*, 2�-��� ��+�(�,�I/���3�%�&�3�3"�7�(3�� 2�-�����&��"3&%(�(3��, 2�-�������"3&%(�(3�� 2�-�-� ��+�(�,�I/��� 3�%�&�3� 3"�7 K3(���%3 2�-�2� ��+�(�,�I/��� 3�%�&�3� 3"�7 )��,,+�� �*+(�%�& 2�-�L� ��+�(�,�I/���3�%�&�3�3"�7 ��%�+�3�%�& 2�-�M� ��+�(�,�I/���3�%�&�3�3"�7� 3��� &N�(�%�& L� /� 0 1����10 1���� �� � � ���� �)� 1���� L��� � ��'�J��/�+%*O+3�%� �+%*��%&�,�� 0 %&%�%,�/���3�%�&�3�3"� L��� � ��'�J����(5&%P+�,��0 %&%�%,� /���3�%�&�3�3"� L��� �(�"&%�%�&�I�5�)��,,+���I /���3�%�&�3�3"�M� /� �.� �1���Q� 1))���R1 STUVWXYZ[\ZW]]_] a[b]W[cdeYfUghiZj `W]ik[l`j `mYgno`WZ]p[\_]q]f]kmnSr Yaa[sYZiZj `k_YStudqZ va]_[hZ _WvZ]kmtwkfW xaY[hiZj `W]ikf[ kq[y`kq[zikWZia{V]a[ kq[|`f[sZi_Yff]km{}r `WYZ[u`kqa]km}V]a[ kq[|`f[h]Yaq[VeYZ W]ikfST
������������� ����������������������� ������!�"��#��������!� �������$$�%���%$����&'()*+,-./(012/(-/3(140(5467-8.4/(3-7-92(:4/86.;<82=(84(+446(02,,(+265467-/:2>'(?32/8.5@(812(:6<:.-,(/2-6(02,,;462(-62-(=<=:2+8.;,2(84(5467-8.4/(3-7-92>'(A-,:<,-82(812(:4=8(45(5467-8.4/(3-7-92(B./(8267=(45(,4=8(+643<:8.4/C>'(?32/8.5@(812(7-D46(=4<6:2=(45(5467-8.4/(3-7-92>'(E2,2:8(-++64+6.-82(62723.-,(862-872/8=>'(A-,:<,-82(812(.7+-:8(45(36.,,./9(-/3(:47+,28.4/(5467-8.4/(3-7-92(B32+81(-/3((( 2*82/8(45(.7+-.672/8C>'(F.=:<==(812(:-<=2=(45(+643<:8.4/(62,-823(5467-8.4/(3-7-92>'(?32/8.5@(812(8@+2=(45(=:-,2(2/:4</82623(./(02,,(4+26-8.4/=(-/3(812(G-6.-;,2=(((01.:1(2552:8(812(=2G26.8@(45(812(+64;,27>'()*+,-./(140(=:-,2(./1.;.846=(-62(<=23>'(E8-82(812(=:4+2H(</326,@./9(:-<=2(-/3(62723.-,(-:8.4/(62I<.623(84(32-,(0.81(-(0-*(((32+4=.8.4/(+64;,27>'(E8-82(812(=:4+2H(</326,@./9(:-<=2(-/3(62723.-,(-:8.4/(62I<.623(84(32-,(0.81(((-/(-=+1-,82/2(32+4=.8.4/(+64;,27>'(E8-82(812(=:4+2H(</326,@./9(:-<=2(-/3(62723.-,(-:8.4/(62I<.623(84(32-,(0.81((( +26546-8./9(3-7-92>'((J64G.32(9<.32,./2=(546(7./.7.=./9(5467-8.4/(3-7-92(3<6./9(046K4G26(((4+26-8.4/=>LMNOPQRSTUMVWXMTVUMYZU[[\ZUMW]M]WZ STQWOMPS S_UMRSOMUMQPUOTQaUPMQOMSMYZWP\RTQWO(((46(./D2:8.4/(02,,>
������������������� ����������������������� ���������������� � ������ !��"#"$"%& '"#� ()*�+�,-.�/ %�".0 1%/ %2�,34%��"5 !$"6"+"02378 �&& 9��#.�/ %0$�7:;)6�%<&"$ -�%$�<�"02:=0+�%>&� -.�/ %�".0+ %06 ?%06 @.0��.&A�"& %06 B%+ 9�.$�++"02AC8 %��� ;%06&"02C�"& %06 B%+ -"�&6 �*��%�".0+DEFGHIJKLMNIGKHOPQRSQTQUVWXURYZ[XZQQUXZ[R\W] _QR]XT TTQ]RPWaRbPQRPX[PQTbRcUQTT UQR]UWcTRdUQRQecQUXQZ Q]RXZRbPQRZQdURaQ__fWUQRUQ[XWZTR] QRbWRbPQRTQ\Xg_W[dUXbP\XRZdbUQRWhRbPQRijkllmjknoplqrstkujkvrqpwslxpiluwyuqxkuzmpoupszw{uk|mrqpwsl}uu~xkutjpqptrvuirjr�kqkju]QbQU\XZXZ[RaQ__RcUW] bXVXb�RXTRbP TRbPQRWZ]XbXWZRWhRbPQRZQdURaQ__fWUQRUQ[XWZ�RR�Z�RPdZ[QTRhUW\RbPQRWUX[XZd_RhWU\dbXWZRcQU\QdfX_Xb�RdUQRPdUd bQUXTQ]Rf�RbPQR�T�XZ�R�rvmk�upsqjwomtkou��u�rsu��kjops�ksursou�mjlqu�lkku�klkj�wpju�szw{u� womvk��uruop�kslpwsvkllusm��kju�rqxk�rqptrvv�ursrvw�wmluqwuqxku�v�uqjrslykjutwky�tpksqupsuPQdbRbUdZThQU��WU\dbXWZR�d\d[QR WZ QUZTRbPQRhWU\dbXWZRWhRdRVW_\QRWhRUW �RaXbPRdRUQ] Q]RcQU\QdfX_Xb�RXZRbPQRZQdURaQ__fWUQR�WZQ�RROPXTRcQU\QdfX_Xb�RUQ] bXWZRdZRfQR] QRbWRdR\ _bXb]QRWhRd TQTRgRf bRXZRd__RdTQTRXbRaX__RUQ] QRbPQR�ZdbUd_�RcUW] bXVXb�R] QRbWRbPQRp�iwlpqpwsuwyursuk�qjruijkllmjkuojwiurluqxkuzmpouzw{luqwuqxku{kvv�wjk}uu~xplutxriqkjuaX__RfQ[XZRaXbPRdR]XT TTXWZRWhRbPQRTW UQT�RQebQZbRdZ]RbPQR\W]Q__XZ[RWhRhWU\dbXWZR]d\d[Q�RR�bRaX__R WZbXZ QRaXbPRdR\WUQR]QbdX_Q]RQed\XZdbXWZRWhRbPQR\Q PdZXT\TR_Qd]XZ[RbWR�UX__XZ[RdZ]R�UW] bXWZRXZ] Q]RhWU\dbXWZR]d\d[Q�RhW__WaQ]Rf�RdRUQVXQaRWhRbQ PZX� QTRaPXPRdZR_Qd]RbWRXbTRdVWX]dZ Q�OPQRhW__WaXZ[R\W] _QTRaX__R]XT TTRaQ__RTbX\ _dbXWZRbUQdb\QZbT�RR�bX\ _dbXWZRbUQdb\QZbTRdUQR]QTX[ZQ]RbWRXZ UQdTQRbPQRaQ__RcUW] bXVXb��RQXbPQURf���X R ¡¢£¤¥¦§ RWUR W\c_QbQ_�R©¢ª«¬¦§ RbPQRhWU\dbXWZR]d\d[QRf�RdR PQ\Xd_RRR \dbUXeRbUQdb\QZbR�Q�[�Rd X]XTXZ[ RWU�XX R® °±±¦§ RbPQRhWU\dbXWZR]d\d[QRf�RUQdbXWZRWhRdRPX[PRcQU\QdfX_Xb�RPdZZQ_RRR f�RdRP�]Ud _XRhUd bUXZ[RbUQdb\QZbRWUR]QQcRcQZQbUdbXWZRcQUhWUdbXWZR[ ZT�²EFK³´J³G µOPQRcWbQZbXd_RhWURbPQRcQU\QdfX_Xb�RWhRbPQRZQdURaQ__fWUQRhWU\dbXWZRfQXZ[RUQ] Q]R�]d\d[Q] RQeXTbTRhUW\RbPQR\W\QZbRbPdbRbPQR]UX__fXbRQZbQUTRbPQRhWU\dbXWZR ZbX_RbPQR{kvvuplu�srvv�ur�rsowsko}uu¶jwtkllklu{xptxuvkrouqwuywj�rqpwsuor�r�kuq�iptrvv�urtquqxjwm�xurujklqjptqpwsuwyuzw{u�wsuruiwjkuqxjwrqultrvk�uomkuqwukpqxkj·�X R �P�TXd_Rf_W �d[QRWURUQ] bXWZRXZRTX�QRWhRbPQRcWUQRbPUWdb�XX RSQ] bXWZRXZRbPQRUQ_dbXVQR�WX_ RcQU\QdfX_Xb�RQ�[�R] QRbWR�d]VQUTQ RhWU\dbXWZRuu {kqqr�pvpq�utxrs�kl�uixrlkutxrs�klupsuqxkuijwomtps�uzmpol}�ppp�u~xkur�w�kukyyktqlutrsu�kurttksqmrqkourquxp�xuzw{ujrqku{xksuqmj�mvksquzw{lRRR \d�RRW ` URgR_Qd]XZ[RbWR\^ PR[UQdbQURcUQTT UQR_WTTQTRbPdZRW ` UUQ]RhWURbPQRTd\QRuu zw{ujrqku�kywjkuqxkuywj�rqpwsuor�r�kuqww uivrtk}�WU\dbXWZR]d\d[QR dZRUQT _bRhUW\R\dZ�R]XhhQUQZbRTW UQTRgR]UX__XZ[�R Q\QZbXZ[�RcQUhWUdbXZ[�RW\c_QbXWZ¹[UdVQ_Rcd �XZ[�RcUW] bXWZ�RXZºQ bXWZ�RaWU�WVQU�RTbX\ _dbXWZ�Rkqt}uu~xklkukyyktqlu{pvvurvvu�kuopltmllkoupsu�jkrqkjuokqrpvuvrqkj}uuu�w{k�kj�u�jlqu{kuskkouqwu|mrsqpy�uqxkul psutwstkiqujkykjjkouqwukrjvpkjursouqwuk�rvmrqkupqlup�irtquwsu{kvvupszw{}
����������� ����������� ��� ������������� ���� ������������ ������ ��������� ��� �!����� ������� ������ ������ � �"�#�������!������� ������$� $�����������!������� ������ � �� ������������%���� ����� ��$���� ��$!�$������� ���� ���� �� ����%������$�!%�$����������$!� ����$!� ������� ������� ����"������� ��&����������� � ��������� � ���'� ���� ��(� �� � ��)�������!�)�� ���������)������� �����!�� ������ �������� ������������ ��������� ���� ��$��$��&"���!� ��� ��*+,-..-/+,*012+3+*+4516078*/,2+-9+6:2+;054+;-<+1672*.=5/21+/-7.*=+6-+6:2+<2==>?@A@BC DEEBFDGHDIJKBLMNDGHDIJKOP:2+<2==+Q2-.267R+1S5/+72;2,61+Q2-.2675,*=+,-/15427*65-/1+<:5,:+*=627+6:2+1S5/+T*=02+�������������&��������"������� ���'��� U���������!�V�&������������ ��� ������� ������������������ ��W�� ���&�����������!����� ��������������X���� � ��������� ������� ����������� �� ��$���� �!�� $���� ����� �& !��� ����� ��������������V���� ��&�������� �����!"����� ��&�������� ���������������� �����V���� ���&����&��������������� Y���������������� �������� �������&� ���� � �����&�������� �����!��� ��� ���������� ����&������ ������� � ��"?@ZB[GH\EDI]G BM_] BLM GH \O���� abcdecfghfijkhlmdniodpcdklmqgrskclj��"�"���&������ ����������!%�����++ 1:-76+-7+6--+/*77-<+1:*t2u+5/,-772,6+t:*15/Q+=2*45/Q+6-+;-<+,-/T27Q2/,2+-7�++ 42T5*65-/+97-.+6:2+/-7.*=+;-<+=5/21+26,>v+*/4+5.t242+;-<+97-.+72127T-57��� �����&��"�w����%�������� �������� ������������� � ��������&����� $����!��++ x-/2+*7-0/4+6:2+t279-7*65-/+<:5,:u+59+/-6+72.-T24u+7210=61+5/+*+7240,24+;-<+7*62>�+++ P:2+5/;-<+5/6-+*+t279-7*624+,-.t=265-/+51+/-7.*==R+=211+6:*/+6:2+y6:2-7265,*=v�++ 5/;-<+6-+6:2+y0/5.t*5724v+-t2/+:-=2+-75Q5/*==R+475==24u+5>2>+6:272+51+*+t-1565T2+++ 1S5/>++z-<2T27u+*+:5Q:+42/156R+-9+=-/Q{<542+t279-7*65-/1+,*/+7210=6+5/+*+109|,52/6�++ 5/,72*12+5/+6:2+5/;-<+6:*6+*+/2Q*65T2+1S5/+7210=61>++P:2+5/627*,65-/+826<22/��� ��� ����������� ������������ ���������� ���"�"�������� ����������%��""��� ���������������� ������ ��� ��������������� �!��� ��&��$����� ���� ������ � ��&�������������������� ������ � ��&������ ��&����� ����� �����!"�����&����$���� ���������� ����� ���� ���������� ���"�����}fi~c�d�is�kl��V����&� �������� ����� ���� ��������&��$������� ����� &��� ������� $����!��� ��%��������!�� ��������������&��������"
������������������� ����������������������� ���������������� � ������ !��"#"$"%& '"#� ()*�+�,-.�/ %�".0 1%/ %2�,34%��"5 !$"6"+"02378 �&& 9��#.�/ %0$�7:;)6�%<&"$ -�%$�<�"02:=0+�%>&� -.�/ %�".0+ %06 ?%06 @.0��.&A�"& %06 B%+ 9�.$�++"02AC8 %��� ;%06&"02C�"& %06 B%+ -"�&6 �*��%�".0+DEEEFGHIJKLMINOPQRSTUVSWRXYUZWY[[\R] ZWWTX_R]WR YWUTa TY[[\bR WSYUScRdW]ddScGRR V\cWYZ[TReWY UZWSfRQRgT[[R[SYcRU]RTX WSYfScRTXh]gRYXcRXS_YUTiSRfjTXfRk\Rd[Y TX_GGG lGmEnmGopqrplsEtEuvGolumwlvGxqyrGzppoGE{GumpGxyqrluEy{GuyGumpGwpttsyqp|GGGG }ql~ptGol��E{nGl{zGxql�u�qE{nG�l{GspG�yrsE{pzGE{GumpG��ql�Gl{zG�l���Goqy�p��GuyGGG vEptzG�yrotpuEy{�GwEumGpxxp�uE~pG�l{zG�y{uqytGl{zG{plqG�pqyG��E{�|�����������������������������������wyGp�lrotp�GyxGoqyz��uEy{G��E{�Glqp�DEFG �GqlupGzpop{zp{uG��E{GE�Gyxup{Gys�pq~pzGE{GmEnmGqlupGnl�Gwptt�GDl{zG~pqvGmEnmGGR WYUSR]T[RgS[[fb�RR�VTfRTfRcZSRU]RX]XQ YW\R]WRUZWkZ[SXURh]g�RR¡UfRdWSfSX SRYXGGG spGlG��px�tGE{zE�luEy{GumluGumpGwpttGE�GlGoyup{uEltG�uEr�tluEy{G�l{zEzlup|DEEFG�qyz��E{nGl{GyEtGwpttGwEum�R YbRRR¢Rh]gTX_Rk]UU]£RV][SRdWSffZWSRkS[]gRUVSRkZkk[SRd]TXURdWSfSX SR]eR_YfbR]W¤R kbRR¢RWSUW]_WYcSR]XcSXfYUSRhZTcRf]RUVYURUVSWSRTfRUg]RdVYfSRWS_T]XRYURUVSGGG G opqxyqluEy{�|G�mp�pG�l{GsyumGtplzGuyGoqp���qpGzpop{zp{u¥GqptluE~pGGG G opqrplsEtEuvGpxxp�u�GDl{Gloolqp{uGE{�qpl�pGE{G��E{GyqGlGtywpqGGG G uml{Gp�op�upzGE{�qpl�pGE{Goqyz��uEy{Gl�GumpGzqlwzyw{GE�GE{�qpl�pzF|GGGG G �mpGoqp�p{�pGyxGumpGp�uqlGoml�pGqpz��p�GumpGpxxp�uE~pGopqrplsEtEuvGuyGumpGGG G rl¦yqGoml�p|GG�mE�Gwy�tzG{yqrlttvGspGE{upqoqpupzGl�Goy�EuE~pG��E{|§ ©©ª«¬©®¯°±ª²®³© ±°±µ¬©¶·²³G y�uGxyqr�GyxGxyqrluEy{GzlrlnpGqpz��pGumpGqy��GopqrplsEtEuvGuyGlG�pqulE{GzpoumGYgY\ReW]£RUVSRgS[[�RR¹T_ZWSRºRT[[ZfUWYUSfRUVSRWSfZ[UTX_RdW]cZ TX_RdWSffZWSRdW]a[SRYXcR]£dYWSfRTURgTUVRUVSRS»ZTiY[SXURdWSffZWSRdW]a[SRe]WRYXRZXcY£Y_ScRgS[[�RR�VSRWSfZ[UTX_RS¼UWYRdWSffZWSRcW]dR½¾zFGml�GuyGspG�yrop{�lupzGxyqGpEumpqGsvGlGqpz��pzGoqp���qpGzqyoGl�qy��GumpG�my�pGyqGsvGlG�rlttpqGoqyz��uEy{Gqlup|G�En�qpG¿|G�myw�GumluGumpGoqp���qpGzqyoGyxGumpG{plqGwpttsyqpG�y{pGE�Gy{tvGy{pG�yroy{p{uGyxGumpGqp�pq~yEqÀU]QfU] jRUYXjRh]gRf\fUS£�GGGGGGGGGGGGGGGGGGGGG ÁÂÃÄÅÆÇÈÃÉÉÊÈÃÆÇÈËÌÍÅÃÎÏÐÂÄÑÄÒÃÂÓÔÕÖÊÄÅÆÇÈÃÉÉÊÈÃÆÇÈËÌÍÅÃÎ×ÄÑÄÒÃÂÓÆÎØÂÆÙÆØÓÚÛÜÚÛÜ ÇÝÇÞ ÇÈØ ØÈß È ÈÃàÆáâÖÈÄÆãÈÃÉÉÊÈÃÆÂÈËãÆÂÊÃÆÖËÆäËÈÑÄÖÍËÐÆ×ÄÑÄÒÃåÃÉÃÈæËÍÈ×ÄÑÄÒÃÂÆçËÐà äÍÒÆèéêë ÃÅÅìËÈÃíÃÐÖÈÃÅÍÐÃHîïMINPðñòNPNóóNKLPôóPOõîöPôöP÷NøøPîöùô÷PúINOOMINPúIôûøNOü
�������� � ��������� �������� ���������� ��� �������� �������������������� ����!��������������������"������������#������ $��$%� &������'������("����)��!�������!�� ���� *+�������� ,���� ����-.�+�����������!�� ���� (�*/01�2113�45�6231784923�401�:75;18818�4094�<193�45�=57>94?52�39>9@1A�923�401�54017�BCDECFGFHIJHKLHJDLMGJNEJHKGJHCHLOPJQIJQOONIHRLHGSJQFJTUNRGJVJWKQBKJBCDELRGIJHKGJLXGRLUGJWGOOJIMQFJYRCDJIGXGRLOJZCRHKJ[GLJTGOSI\JJ]KG JWGRGJLOOJSRQOOGSJLHJLRCNFSJHKGJILDGJHQDGJLFSJICDGJWGOOIJQFJGLBKJTGOSJWGRGJSRQOOGSJWQHKJCQOJ_LIGSJDNSJabc dJLFSJCHKGRIJWQHKJWLHGRJ_LIGSJDNSJe bc dJSRQOOQFUJDNSI\JJfHJQIJBOGLRJHKLHJHKGJTGOSJLXGRLUGJWGOOJIMQFJXLONGJQIJXLRQL_OGJ_GHWGGFJTGOSI\JJgNRHKGRPJQHJQIJBCFIQIHGFHOJOCWGRJWKGFJCQOJ_LIGSJDNSJQIJNIGSJLIJLJSRQOOQFUJhNQSJQFJERGYGRGFBGJHCJWLHGRJ_LIGSJDNS\JJiCWGXGRPJQHJ?8�62;<197�j014017�401�9k5l1�l9<618�=57�5?<�k9813�>63�971�5:4?>6>�57�j014017�401m�;56<3�k1�=674017�7136;13n������������������� opqrstuv w u x y z{ru| {ry|u|w|v||x } ~� �~������q�r}���r��p� op���/0?8�2113�45�31417>?21�j014017�401�5:4?>6>�8�?2�l9<61�098�k112�9;0?1l13��?8�933718813�km�401��9j�?28���694?52A�40?8�;9<;6<9418�401�=57>94?52�39>9@1�8�?2���39>9@1��7186<4?2@�=75>�9�;m<?2317�5=�7136;13�:17>19k?<?4m�975623�401�j1<<k571�� �������������������������������������������� �¡¢£¤¥¦ §£� §�£ � � � � � �©� ª«¬® °±²³ ® ¬ °µ ¶¶°· «¹°³ ¶ ·°º»®°·»¼ °½»®¾¿°À¯ °Á ¶Â·ª«¬® °ÃÄ¿ °¹ ¾® ¶°Å»µ°Æ®»ÂÇ«¹¬°ÈÉÈÊËÌ
������������������� ����������������������� ���������������� � ������ !��"#"$"%& '"#� ()*�+�,-.�/ %�".0 1%/ %2�,34%��"5 !$"6"+"02378 �&& 9��#.�/ %0$�7:;)6�%<&"$ -�%$�<�"02:=0+�%>&� -.�/ %�".0+ %06 ?%06 @.0��.&A�"& %06 B%+ 9�.$�++"02AC8 %��� ;%06&"02C�"& %06 B%+ -"�&6 �*��%�".0+DEFGHIJIKFLFJMGIJFGNFOPFNGQPGORSJFMGTGIPNGUGIPNGQPGINNQLQVPWXYZY[\] _a\bYXcadXeGfGgSQNGgVhGJILFiGfGgSQNGjQMkVMQLlmno_a\bYpqrYsX\tsYX_YXcY[\] _a\bYu_ _dcYsvabYabw]c_scsY_sYXcYxc] c_yaza{Yu_ _dcY]_a\Y|v}vu~Yabw]c_scsY\]YXcY]_uaosY\[Yu_ _dcY|]u~Yabw]c_scs�YY�XasYasYazzos]_cuYQPGORSJFG��GGDESMGQ�GLEFG�VJKILQVPGHFJKFI�Q�QLlGQMGJFNSkFNGLVGT��GV�GQLMGVJQRQPI�GjI�SFG|v}vuGfGT��GVSLGLVGIGJINQSMGV�G��GkK�GLEFPGLEFG�VJKILQVPGNIKIRFGM�QPG��u�GQMGFeSI�GLVGT��YYYYYYYYYYYYYYYYYYYYYYYYY ������������ ������� ������ � ����� � �� ��� ¡¢�£�����¤������� ���� ��¤� ���� ��£� ��¥� ��¦� ���� ���� ��¤� ����¢�§� ©��ª«�¡¬�©®©���° �±�¢«²³���³� �©®©��²Y YYYYYYYY µ¶· µ¹ º » ¼ ½ ¾¹ ¾º ¾» ¾¼ ¾½ º¹ ºº º»¾¹¹¹¾¹¹¾¹¾¹¿¾¹¿¹¾ ÀÁ¶ÃÄÅÆÇÁÈÁ·É ÊÅËÉ ÌÉÍÅË ΠÉÏÏÐÅÑÉÒÆÓÔÑÅÂÃÕÓ¶ÅËÀÁÓÉÒÐÐÏÖÂÁÍ ×Â Ø Ù¹È Â×Â Ø ¾¹È Â×Â Ø ¾È ÂÔÉÑÈÉÁж϶ÓÍÅÆÃËÂÁÈÁ·É ÑÉÄÑÉÑÚŶÑØ ¾¹¹ È ÂµÅÑÈÁÓ¶ÅË ÓÛ¶Õ×ËÉÄľ¹ÆÓ¿Î ÉÏÏÐÅÑÉ ÑÁ¶ÃĹ¿ºÙÆÓ¿ÇÑÁ¶ËÁ·É ÑÁ¶ÃÄÙ¹¹ÆӿܶÏÚ¶ÄÕÅĶÓ͹¿Ù ÕÝÇÑÁÞÂÅÞË Ùß¼ ÝĶàÆÆÉÕÓÅƵÅÑÈÁÓ¶ÅË ÇÁÈÁ·É ÅË Î ÉÏÏÔÑÅÂÃÕÓ¶ÅËDEFGJFMS�LMGV�GáINQI�GQPgVhGkI�kS�ILQVPMGSMQPRGLEFGI�VjFGLlHQkI��GHFJKFI�Q�QLlGNIKIRFGJILQVMGIJFG�VSPNGQPGORSJFGâ�GGã F��GHJVNSkLQVPGQMGJIHQN�lGJFNSkFNGIMG�VJKILQVPGNIKIRFGQPkJFIMFM�GhQLEGLEFGNIKIRFGQPGLEFGjFJlGPFIJGhF���VJFGJFRQVPG�OJMLG�FhG�FFL�G�FQPRGXcY \sYa x\]_b�äåæçèéêëìíåîêïðñòóèêåîñèéðôéôêõåòöêðê÷ðèæéèêøðùðæéêèðøåçôêðîøêøðùðæéêèðòåóêíúíøäåæçèéêûüïïéñòêóïêïóèùðòåóîêøðùðæéêóîêõé÷÷êýèóøçñòåóî
��������������� ����������������� ����� ��� ����������������� ��������������������� !��� �!�"!#�$%&!� !�$�'(!�"!�)$!*+�%!,"-��$ �.!/*,01� 23345676894:596;233<;=>?9@?ABCDCEC:596;F<;=>?9@?ABCDCECGHG6;FICJ>KBCLCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCLC � � � MNOP�����Q������R�Q����S�Q������T���T�T�Q����������� ����� ��������UV��� �WXWYZ[\Y] ���������������������_ ���������� ������a������������ ������������V������� T!b)�(!�$+c��� ()��!d$�%$$ !e�%!$"-��$ �.!c �!(f� !�(!������$�!c(!-g��$!h1!!b)$!�'�c��!� !�)$!%$++!��������� !�"!�)$!�c'cg$!(f� !�"!ij&!�c+��+c�$�!� !�)$!$kc'�+$!��(��(($�!$c�+�$�&!�$���$(!��������� !��!� +.!ljm!"��'!%)c�!��!%��+�!)c#$!d$$ !� !�)$!�c($!�)$!����������������P� ��U ��T���������������������� nopqrstu t uv ut wv wt xv tv uvv uvvvyvrvzv{vtv|vxvwvuvv }~o�q�}q����p��n���q���o�o����q������)!�$������ (!� !e�%!$"-��$ �.!%�++!�$���$!�)$!%$++�(!��������� !�!�$+c.� g!����$��!�c.dc�f!c(!%$++!c(!�$���� g!�$�!��$($ �!�c+�$!���-�cd�+��.!/-g��$!�01!!*���)$�&!�)$!$k��c!��$((��$!+�(($(!� !�)$! $c�!%$++d��$!c�$c!/*�g��$!�0!'$c !�)c�!c���-��c+!+�"�!%�++!�������U���� ���������������������������������P�����������V���T���� �P��������� �����P��� ����������� T ��������� ¡¢¡�£ �¤ £¥¥¦¡¢¡�£ �¤ £¥¥�� ¡¢¡�£ �¤ £¥¥ ¦¡¢¡�£ �¤ £¥¥¤£¥¥�§© ª«¬�©�§©£«¬�®£¬� §££�¬�°¡¥ª£ ±�¢£±�¢£§© �¬�²£®£�¡¬�²£¦£³£¨£ � © ª«¬�©��¡� �£ ª«£ � ©£«¬�²¡¥ª£�£ ª¥¬�³©¢�³© ¢ ¡¬�©�� ¡¢¡�£ µ¶· ¹º»¼µ½¾¿»ºÀÁ¶ºÃÂĻź¹ºÆǺǻ¿¶ÈÉ»¶Ã¹ºÆǶ÷»À¾¹ÊÆȶ¾Ã»ÅÆÊÆ·º»Ç˶õ¶· ¹º»Ìµ¾¹ÊÆȶ¾Ã»ÅÆÊÆ·ºÍι¾Å Âȶ¾Ã»ÆÃŻϹ¾ÁÈÆж½¶ÈÄ
������������������� ����������������������� ���������������� � ������ !��"#"$"%& '"#� ()*�+�,-.�/ %�".0 1%/ %2�,34%��"5 !$"6"+"02378 �&& 9��#.�/ %0$�7:;)6�%<&"$ -�%$�<�"02:=0+�%>&� -.�/ %�".0+ %06 ?%06 @.0��.&A�"& %06 B%+ 9�.$�++"02AC8 %��� ;%06&"02C�"& %06 B%+ -"�&6 �*��%�".0+DEFGHIJKLGFHMFMHJNOPQHRFSONOTLUVWXYZ[V\]YXY_ ]aY\]VaabW]ZcWVb_cVbZ]Zc ]d[e]Ve]Zc ]f dd]eWVX]Zc ]XVX \Z]ZcYZ]ghijklmnnjomgjplqgjrisigltgiqjghijuvlwtgmvsxjjynnjzinnjt{gm|mgmiqjsiikjgvjoiji|tn}tgikjeVW]Zc [W]~VZ\Z[Yd]eVW]aYb�[\_]eVWXYZ[V\]YXY_ �]]�c �][\adb ��]�W[dd[\_�]� X \Z[\_�]� WeVWYZ[\_�]�VX~dZ[V\��WY� d]�Ya�[\_�]�WV baZ[V\�]�\�aZ[V\�]� VW�V� W�]�Z[XbdYZ[V\�hijtmwjvujghij�lvk}{gmvsj�i{hsvnv�mqgjmqjgvjt{hmi|ijtsjmsmgmtnj�vzjiup{mis{�jvuj����jY\ ]ZV]XY[\ZY[\][Z]YZ]Zc[�]d� d]bW[\_]Zc ]d[e]Ve]Zc ]f dd]Zc[�][�]Yac[� ]��]\�bW[\_]ZcYZ]~WV~ W]V~ WYZ[V\]~WYaZ[a �]Y\ ]~WVa bW�]YW][X~dX \Z ]bW[\_]\VWXYd]f dd]V~ WYZ[V\�]Y�]f dd]Y�]bW[\_]Zc ]W[dd[\_]Y\ ]fVW�V� W]aYX~Y[_\��]]�VX ]Ve]Zc � ]�VbWa �]Ve]eVWXYZ[V\]YXY_ ]f[dd]� ][�ab�� ][\]Zc ]eVddVf[\_]� aZ[V\���������� ¡¢£�¤�¥¡ ¡¦§�¥�£©©£¤¦�ª«§�¡¢£�¤¬® vqgjklmnnms�j�}mkqjtlijwtkij}rjulvwjrtlgm{}ntgijwtgilmtnqjix�xj{nt�q°jzim�hgms�jt�isgqjZa�±]Y�]f dd]Y�]aV\ZY[\[\_]W[dd]abZZ[\_��]]²\d��]Zc ]f dd][�]W[dd ][\]Y\]b\ W�YdY\a ]{vskmgmvs³jghijrliqq}lijv|iljotnts{i°jli }mlikjoigziisjghijklmnnms�j�}mkjtskjghijW � W�V[W]~W��bW]ZV]� ~]Zc ]f dd]b\ W]aV\ZWVd±]f[dd]W�bdZ][\]Zc � ]Xb ]~YWZ[abdYZ�]oims�juvl{ikjmsgvjghijuvlwtgmvsxj�hijrt�jµvsijzmnnjgh}qjoijrlv�liqqm|in�jpnnikjtskjtjpngilj{t¶ijzmnnjoijuvlwikjvsjghijq}lut{ijvujghijzinnovlixjj�hiqijrtlgm{}ntgijqvnmkqjzmnnjsvgjitqmn�j�vzjot{¶jmsgvjghijzinnovlijzhisjghijrliqq}lij�ltkmisgjmqjli|ilqikjmxixjghijVW]f dd][�]~dYa ]V\]~WV baZ[V\·�]UVWXYZ[V\]YXY_ ]cY�]Zcb�]� \]aWYZ ��lmvljgvjisgilms�jghijuvlwtgmvs°jghijklmnnms�j�}mkjqhv}nkjoij{hvqisjqvjtqjgvjlik}{ijZc ]ZVZYd]W[dd[\_]aV�Z]Zc[�]\VWXYdd�][X~d[�]eY�Z]W[dd[\_·�]]¹\a ]Zc ]~Y�]ºV\ ]cY�]� \]~ \ ZWYZ ±]XY»[X[�[\_]Zc ]f dd]~WV baZ[�[Z�]� aVX �]Zc ]� �]aW[ZW[Y±]� \]YZ]Zc ]»~ \� ]Ve]�dVf W]W[dd[\_�]]�c ]�Vd[]~YWZ[abdYZ�][\]Zc ]W[dd[\_]Xb ]�cVbd]� ]{hvqisjgvjht|ijtjq}mgtonijqmµijqvjtqjgvjuvlwjtjpngilj{t¶ijvsjghijovlihvnijztnnxjj��rm{tnjlintgmvsqhmrqjoigziisjklmnnms�j�}mkjg�ri°j{vqgjtskjghijlmq¶jvuj¼vlwtgmvsj½twt�ijtlijqhvzsjmsjp�}lij¾x
���������� ��������� ����������� ������������������� �������� ���������� ���� ��!�"����� ���#��� #���$��$�� � ��%�� ��������&��'& ��%�������&��'&()*+,*-.*/012134+56+053)+3)17+823*-+9/:*+/;<+3)*+65-./315;+1;=>*;9*+3)*+?@AB�@A�CDEFD�3)*+<-1221;G+.><+823-/3*+1;H/<*7+3)*+65-./315;I++()17+-/3*+56+1;H/715;+./4+0*+9/29>2/3*<+6-5.+3)*+-/<1/2+=5J+*K>/315;7+J13)+/+25J+,*-.*/012134+L5;*M3)*+.><+9/:*N+,2/9*<+;*O3+35+3)*+J*2205-*I++()*+-*7>237+/-*+122>73-/3*<+1;+8G>-*+PI������������� QRSTUVWXYYYYXYYYXYYXYXYVXQRZ[\]T _ \T ab[]aZXcdYe XcdYf YVYYYXYVYYYYXYVYYYX YVYYXYVYYX YVYXYVYX YVXYVX XX XYXY XYYXYYgbh_iRabTj_[\TkllZimn[mo]p QRZ[\]T _ \Tq\]r\_lRZR[sTkrtp Qa]r_[RabTq\]r\_lRZR[sTkrtpQa]r_[RabTq\]r\_lRZR[sT ab[]aZ()17+8G>-*+7)5J7+3J5+-*G15;7uM1N+ v )*;+3)*+823*-+9/:*+)/7+/+.>9)+25J*-+,*-.*/012134+3)/;+3)*+65-./315;I+()*+++ 1;H/715;+-/3*+17+<*79-10*<+/7+0*1;G+>;<*-+823*-+9/:*+95;3-52IwEEx�yDB�Dz?E{z|A@}�~z?AEz|��z?�B@FD��z?�@AEz|�~B?�B@�E}EA��E|�EF@AB��ADB�{z|B��� E|�CDEFD�ADB��z?�@AEz|�~B?�B@�E}EA��D@���BFz�B�ADB��z�E|@|A��@FAz?�E|�Fz|A?z}�++ 21;G+3)*+2*/:+566+-/3*I++()17+599>-7+/7+3)*+,*-.*/012134+56+3)*+823*-+9/:*+1;9-*/7*7��� ?B}@AE�B�Az�AD@A�z��ADB��z?�@AEz|� ��������������������������������������������������������������� � ���������������� ����������������������¡�����������������¢�����������������
������������������� ����������������������� ���������������� �� ������ !��"#"$"%& '"#� ()*�+�,-.�/ %�".0 1%/ %2�,34%��"5 !$"6"+"02378 �&& 9��#.�/ %0$�7:;)6�%<&"$ -�%$�<�"02:=0+�%>&� -.�/ %�".0+ %06 ?%06 @.0��.&A�"& %06 B%+ 9�.$�++"02AC8 %��� ;%06&"02C�"& %06 B%+ -"�&6 �*��%�".0+DEFGHIJHFIKLFMGNHFIOFPQLRLKSGPTUGKELVWTFMMXGJYGKEFGOZUVPWFGPIFGUFHFTUFTKGJTGKEFGKSHFGJYGUILRRLT[GOZUGLTGZMFGPMG\FRRGPMGJTGKEFGUILRRLT[GHPIPOFKFIMGPTUGKEFG]ZPRLKSGJYGKEFGOZUGFT[LTFFILT[GF[_G MGKEFGUILRRGMJRLUMGQZLRUGZHGQFLT[GVJTKIJRRFUGPKGPGRJ\GaPRZFb_GcIFGKEFGOZUGHIJHFIKLFMGQFLT[GIFdPUeZMKFUGKJGKEFLIGVJIIFVKGaPRZFMbDSHLVPRGFYYFVKMGJYGKEFGOZUGHIJHFIKLFMGJTGKEFGLTaPMLJTGUFHKEGPTUGYJIOPKLJTGUPOP[FGfghijklmgklnohpqlrqlistgkluvwllx tylzl{hg|nlml}or~��l�t}l|hgklfkg|km�jk�lij}kgl~m�klpor~olmjjhpnl�tryljkm�lh{{�lmqylmq�l�n|mjj�lnhjrylygrjjrqsl|tylfmg}r~jknl}om}lfmnnl}oghtsol}oklij}kgl~m�k�l}hlrq�mykl}okl{hg|m}rhql}hlml~hqnrykgm�jklykf}owll�oklgkyt~}rhqlLTGHFIOFPQLRLKSGJYGKEFGYJIOPKLJTGdGJIGYJIOPKLJTGUPOP[FGdGGUZFGKJGOL�KZIFGJYGUILRRLT[G|tylij}gm}klrnlorsolmqy�l|hgklr|fhg}mq}j��lhqj�lyk~gkmnknlnjhpj�lpr}olsgkm}kglykf}olh{lrq�mnrhqwllx tyl��l��l~hq}gmn}�l{hg|nlml}orq�lorsoj�lr|fkg|km�jk�lij}kgl~m�kwll�ornlIFMZRKMGLTGFMMFTKLPRRSGKJKPRGYJIOPKLJTGUPOP[FGLOOFULPKFRSGPUePVFTKGKJGKEFG\FRRQJIF�GQZKGKELMGUFVIFPMFMGIPHLURSGKJGPGRJ\GRFaFRGPMGKEFGUFHKEGJYGLTaPMLJTGLTVIFPMFMGGGGGGGGGGGGGGGG ������������������� ����������� ¡¢���£���¤¥¡�����¦¡¥¡�� § ¦� ©§ ¦� ª § ¦� ©§ ¦� ª �� ¡¢�����¤���«�¤��¤¡����������¬���®�����°�¥��¤¥�¡±�®¥¦�²¡³�®���������±¡¢��¥¦µ �±�¤� ¬��²³�¤ ®�±�¥�¤����¤¥�¡±��¥¦ ²¡³�®�����¶¡��¤�±¡¢��¥¦·¸¹º»¼½¾¿ÀÁ ELVEGOZUG[LaFMGQFKKFIGHIJUZVKLJTbDEFGHIJHFIKLFMGJYGKEFGUILRRLT[GOZUGPTUGKEFGUILRRLT[GOZUGÂGYJIOPKLJTGLTKFIPVKLJTGPMGpkjjlmnl}okl~h|fjk}rhql}�fkllprjjl~hq}ghjl}oklpkjjÃnlfghyt~}r�r}�lm{}kglr}lomnl�kkqlVJOHRFKFUÄžºÆ¹Ç¿NLXG ÈJIGPGÉÊËÌÍÎÊÏÍÎÐÌÑÒÓÑÊÔÌÍÎÉÓÕÐÖÌÔ×ÓÏ�GØ ZUGÙGLMGF�HFVKFUGKJG[LaFGKEFGEL[EFIGll pkjjlfghyt~}r�r}�lnrq~kl}okljkqs}olh{l}oklfkg{hgm}rhqlrnlnt{i~rkq}l}hlfkqk}gm}klGG VJOHRFKFRSGKEIJZ[EGKEFGLOHPLIFUGÚJTF�GVJTTFVKLT[G\LKEGYJIOPKLJTGMKLRRGIFKPLTLT[GGG LKMGJIL[LTPRGHFIOFPQLRLKSLLXG ÈJIGPTGÓÐÌÏÎÛÓÖÌÎÉÓÕÐÖÌÔ×ÓÏGKEFGVEJLVFG\LRRGUFHFTUGJTGKEFGF�KFTKGKJG\ELVEGll }okl|tyl~m�klmqylij}gm}klmgklgk|h�kyl{gh|l}oklpmjjlh{l}oklpkjj�hgklhq~kl}okGGG \FRRGLMGHRPVFUGJTGHIJUZVKLJTÜ¼Ý Ç¹ÀÞßàáá¹â»À½áÀã¸äÀ»åæ¹À½¾À»ç¹À¼¾è麼½¾Àä¹æ»çÀé¾äÀ»ç¹Àá½Çã黼½¾ÀäéãéݹÀæǽêë¹
������ �������� ������������������������������� !�" !#$!%��&'())(*+�,-(&�(%�./�!*��� 0��1������� 2��3��4 ����1445���4���3 ���6���4���7����8�1����674���1��� 67�0��5�271���2������4��1� ��� �0�4 �� 2��9���:;:<:�� 2���������7� �0��92��� 2��4��1�����4��1��8��� 2������ �������5 2=��:�:���� 24 � 2��4��1���� 4� ��4��� 2����3��4 �����4�1����8��� 2��3��4 ����1445��3����8��7���3��4 ���:��> #*?�/#"�.'%��#$#'��/'.@�� !�$'.$!'�(!%�./�� !�&'())(*+�,-(&��".*�'.)�� !�&!$� ��.�A (" �3��4 ����1445���4�����0�:B��� CDEFGHIDJKLEDLMEHIMN�O�#*�(*"'!#%!&�&!$� �./�&'())�,-(&�(*P#%(.*�(%�./�!*�� �6���8�1�92���4��4�5������Q��R��Q����2�52������46��� 7�Q�3��4 �������6���5��� &'())!&�&-!��.�'!&-"!&�!/S"(!*"?�./�S)�!'�"#T!�/.'@#�(.*U�V*�/#"��� (+ )?��� (@$#('(*+�).%%!%�./�A .)!�@-&�W(U!U�*.�S)�!'�"#T!�/.'@#�(.*X�"#*�Y!�.Y%!'P!&��� ����� 2��3��4 �������46��� 7����Z�[4��7����5��4 ���B1���1��5���� 2���� 1� 4���1�����5��� �7�=�0����������4��3�0�1��������� ����411� �8����� WY#%("#))?��)#'+!'�$#'�(")!%��.�$'.@.�!�S)�!'�"#T!�/.'@#�(.*X�#'!�#&&!&��.�� !��� &'())(*+�,-(&UB����\LMJK]DMKHIFM�O�,-(&�).%%�(%�#�".*�(*-.-%�$'."!%%�O�� !�).*+!'�� !�.$!*� .)!��� ����_���1� �� 2��1������5�01=� 2��5��4 ��� 2����84�����1� 2:��2��������� 24 ��� � !'!�(%�#�+'!#�!'�'(%T�./�(@$#('@!*��/.'�� #��$#'��./�� !�/.'@#�(.*�&'())!&�S'%���� ��4��1� �� 2��3��4 �����1�����1��4 �����:��2������4� ��0�4��7���� 4� �3����� ).*+� .'(a.*�#)�A!))%�A !'!�".@$)!�(.*�(*�!'P#)%�+'!#�!'�� #*�bccc�@�#'!��� �� �0�����:B�����defghijiklf�O�,-(&�).%%m(*P#%(.*�&!$� �(*"'!#%!%�#%�� !�&(//!'!*"!�(*��� ����0���6� 9���� 2��1������5�01�4�1� 2�������8��������0���B�8��64�4������� �����4���:�n�1���64�4����1������5�4�1����� ���� ��2��o0�����9�6���5�� 0��1� ��1�����9��������4�� ��48��1� 2��� 7���3�3��4 ����1445�:B�8��pqgfrqjfstukivwlx�A())� #P!�#�)#'+!�!//!"��.*�� !�,-(&�).%%� B4�� y4�5��B���5�4�1z���9�1��1�4� ����<� ��{����|���6������� � �:5:�� 46��������4�1� 2���4��7��5��0 ��3�4�7���0�1� �����:5:� ���24�5���� � 2��1�����6� ����41}0� � 2��1���� ���4��1������5�4���6�7�9�������0� ������ � � !�/'!~-!*��%"'#$(*+�./�� !�@-&��"#T!U����'#�,-(&�).%%�A())�'!%-)��#%��� � 2��01��4��������� ���1� ��� ���o0���6��0� 2�������:� B6�� {�52�����0�4 �����4 ���9���������4��� 2��17�4����8��64�4����92���� 2���� � (+ �,-(&�P!)."(�(!%�A())�!'.&!�� !�@-&�"#T!U� !�� (**!'�@-&�"#T!��� � #*&�� !�+'!#�!'�.P!'Y#)#*"!�A())�Y.� �'!%-)��(*�(*"'!#%!&�,-(&�)!#T�.//U� !�'!)#�(P!�,-(&�).%%��EGHMN��4�����4 �1�9� 2� 2��84���0��24�����3� 2��1������5��������#'!�())-%�'#�!&�(*�S+-'!�bbU��� !�".''!%$.*&(*+��'!)#�(P!��D�FMN�3����4�2�24���4���#)%.�% .A*U��� !�'!)#�(.*% ($�Y!�A!!*�&!$� �./�(*P#%(.*�#*&�,-(&�)!#T�.//�P.)-@!�@#?�Y!�~-#*�(S!&�Y?�#�%(@$)!�P.)-@!�'("�"#)"-)#�(.*�#%�% .A*�(*��#Y)!�bU��� !��#Y)!�% .A%�� #��)#'+!�P.)-@!%�./�,-(&� #P!��.�Y!�).%���.�#" (!P!�%-Y%�#*�(#)�&!$� %�./���84����������4���46������0��3��4 ����:�
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EFGHIJKLMK NO KP OM PNQRSTUHVWXXYUHZFUH[Y\]YXH[T]X[WT_YUFW`H]aRWX]bc]YTHZFUUSTZSV]`UH[VWd[VSFbHVWXX EVSFbHVWXXHWRRWXFU]XUYZFVFe]TX EVSFbHVWXXHd^FV]HbTFVVFGWRRWXFU]HbTFVVHRFR] QUYUF\H[FVUTYUFW`Hd^FV]UTFRRFGHbTFVVHRFR]EVSFbHVWXXHfWVS_]PLHX]\ gH_ F KH T KbYh PHbYhX OLHbYhXEVSFbHVWXXHTYU]HiVWGHX\YV]j kF_]ll mnomnpmnqrrmrpm nqpmrstmuunmposmpprptmvvmqsmnosmnnpmmnumnqmrumrqmunmowtvmnwovmnwutmnwsxmnwtsmowrtmowutmowpqmowspmowvumowopmowosmowovowoqowoxm yz{|}y~��y����� {y�� ��u��m mmmmmmmmns��m mmmmmmmmuo��m mmmmmmmmnso��mmmmmmmmmmuoo��mmmmmmmmvoo����������y���� ��{���y��y���|����y���� �������l�� ¡¢£l¤�¥¦§§©ª«¬® §°±²³ ²µ¶±·§¹¶±²ºµ»§¼½¦§§¾ª«¿®À��ÁÂ�ÃÃlÄ��¢Å�lÆlÇÈÉlʤ�¥lËl�ÌÍlÂl�¥ÍÎl l l l l ¤ÏÌÐÑÒÑÒÑÓÔÕÖ×ØÙÚÕÛÓÜØ×ØÝÞÓß ÞàáØÛÚâ×âÓÜãÖÚÛÝÓÜÖÚääÚÛÝÐÑÒÑåÑÓÔäãÚæÓçÕââlè½±§®±½µ¹é®é§êë§ì½½§¶½±§·²»»¹í§®î·§ï»¶²µ¶±§¶½µ¶§»±µð駳³§¶½² îí½§¶½±§ï»¶±²§ñ�Á�l¡òó�lóÉ�lÃ��Å�ó¡�òl¡òñ�¢ �ô¤¡Ìl õòñ���£¡òölóÉ�l¥�ó��l£�ó¢��ó¡�òl¡òlóÉ�lò���l¥�������l����l¤�� ¢ñ�£lóÉ�l����ó¡÷�l§§ ø±²®±µê»¶ë§¶§ »¦§¼ïíî²±§©¾¦«§§è½é§é§øµ²¶î»µ²»ë§®ø ²¶µ¹¶§³²§»ì§ø±²®±µê»¶ëlll ��ñÁ£l¥É���lóÉ�l��Å�÷��l�ÃlóÉ�lù�úó��ûl¥�ó��l£�ó¢��ó¡�òlÅ�üló�Á�l�l��òölll ó¡Å�ýl¡þ�þl¥���lùñ���òl¢ÿûlÅ�üló�Á�lÅ�òülÅ�òóÉ£þllõòlÃ�ñóýl¡ólÅ�ülò�ólÿ��÷�lll ÿ�££¡���ló�l¡ò¡ó¡�ó�lÿ�� ¢ñó¡�òl¡òl�l���£�ò����ló¡Å�l¡ÃlóÉ�l¥���l ��¥ �¥òl¡£l§§ ¹ ¶§éî³ï±¹¶§é¹±§¶½±§¼²±»µ¶º±¦§ø±²®±µê »¶ë§³§¶½±§½ë·² µ²ê ¹§ø½µé±§½µé§§l ��ñ�Å�l£�l��¥þ����������������� �������������������������� !������ �����""�� !�� ������!����� ��!�#�""
������������������ ��������������������������������������������������������������� ���������� !������"##���$%&���!����#��"'�� ��(��)"#"�*��+�����,-�%���./01��� 23456473789:6; 2344<5=9> <5=? @4A<B=B9C6;5?<5=DEFGHEIJKL KME NFOEHIJKL KMEC6P9QR2C3?S<56349@ST<6?S=459U7=9539> <5=?9V;3WX:6;2;<Y9Z<?56W;=823<59[<4B9\?<64 [<4B9\?<648 ]7<?593?9:5_=?92=S=45923483;6B<5=89\?<648 :6;9U?3T;=5 \<8���������� `abcdefghdgidjab kgllamndo amnp qrfdspgtfnmdepgudqrfdvabnwdxywz{|}~{���� ��{ �|�{~���� ��{kfacd�apmr�fnbkgamd�alwdjparl �alwdjparlb �yapm�dgpdqm�npdknunlmdkglbgfrwamnbdjparlb qrfdspgtfnm jab er�d���a�wwrmrgldgidm�rpwdt�abndpnwy�nbd�abdifghkgllamndo amnp jabdefghd�utnwnw����������������������������������������������������������������������� ������������������������ ��������������������������������������������������������� ���������������������������������������������������������������������������� -#��������1�1���+(����"�'����+"#��(%#�"+��(%$��(�*�&�%����!���������"+��+������ '"�&+%����������"����+"#��(%#�"+�� �������"���!���+�(��"+��,-�%���. 01 ¡¢£¤¥¦§©§¡ª¥«¬¢ª®§¢« ¬¢¥«¦®§°¤¦§ª§±¬¦¥§§²³ª µ¡¢£¤¥¦§¶§·°°¢¢ª®§ª §¹¢¥°§¹¬º¦§¥¦°¤ ¦º§£¬º§»ª±
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EEE FGHIJHKGKLMNJO FGHHPIQMR PIQS THUPVQVMWJOISPIQXYZ[\Y] _` _aY bZcY\] _` _aYWJdMefghJLiGKLMNJOMjMTHMR PIQSMklKOLJGHFOPmMnPSIJiOQLFGPIMoPHVMpSPJH oPHVMpSPJHL qKPSIrMGSMNIsQSMFQlQHIMFGHLGOJVPIQLMpSPJHL NJOMtSGuOQI pPLvwxyz{|}z~�����}z���}�z���z�|}z���z������}z��zw��}~�w��z��w�z���z�}zw�~�����w��}�EEE ���������E��E������������E�����E����� ¡E¢��£E������������E¤���E�� ���EEE �£�E�����������¥E��E�£�E����E¤�������E���������¡EE¢�������¥¦E���¥����E �����§� Ezz w�z�|}z���z������}z���z�� ���z��z�|}z©�����w��z��©�~} �z�} ��w~�w�ªz�|}z��}�Ezz ��}�z��z���zv�ª��}z«¬y
�� ���������� ����������� ����� �� � �������� ��������� �������� ��� � � �������������� ��������� � ���� � �!"#$%"&'() (*" +#,"%&'() (*"������� ������� ���� �-��� ��� �-���� .�� /��� 0������� ������ � ���-���� �������� -��123456789:;<34=>4:234?3@536A9B9:C4@3<;D:9=84>@=54A=:24:C?3E4=>4>=@56:9=8FG;9<4HIJKLMNJHOIPQHRRPSKTKISPOIPJUKVWHXP YRMZPNUK[H\JLZPWJUKLKPMLKP[MIZPJZTK\PO]PNRMZXWHHXPYRMZP[OLTUORO ZWHHHXP_[O IJPO]PNRMZPMISa9bc4d2359E:@C4=>4:2345;<4eB:@6:34f4?g468<4E6B989:C46@3495?=@:68:423@3hUKPSH]]KLKIJPJZTK\PO]PNRMZPNMIPiKPLKNO IH\KSPiZPJUKHLPMTTKMLMINKjPPkH LK\PlmPnPolP\NUK[MJHNMRRZPLKTLK\KIJ\PJUKP]OL[\PQUHNUPSH]]KLKIJPNRMZPJZTK\PMLKP]LKp KIJRZP]O ISVqPrstuvwvxy4>=@5E4A==zB3:E4{29D24A@9<734?=@34:2@=6:E4=@4eBB4?=@34{6BBE4ae7;@34|}cqP~uuvxy4>=@5E4>@679B34{9E?C4eB6538:E4{29D24B9834?=@34{6BBE468<4A@9<734:234?=@3EP4 4:235E3Bb3E4a�97;@34|�c���������������������������� �������������������������������������������� ������������������������PPPPPPPPPPPP ���������������������������� �������������������������������������������� ������������������������qP���� ¡ �426E4642=83CD=5A4E:@;D:;@34a�97;@34¢£cqP¤¥ut¦vxyPH\PMIPHLOIPNOIJMHIHI PNRMZPQUHNUPRHIK\PJUKPTOLKPQMRRPWkH LKPolX §¡©ª�«¬§®ª� ¡®°«¡�± ¡ª��° «²©�« ®«�³ «µ¶�³³¡°¨§¡©ª�«¬·«³�¹ º»¡µ�ª� �«¼®®½³� µ«¾ ®³¡°¡ �º§¡©ª�«¬¿«ª¡À ºÁ®ª�«¼ª¡² ¡° «Â³³¡ �º
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,
������������� ��������������������������� ����� �� ������ ����������������� ������������� !�"#$�%&'�()*!+�)&,�,-.�(,/&$!�� !�%-'!+�� !"+!*/!+0��1 !2�"&2�!(� !'�&$ !'!��-���������������������������������������3�4�������� ���� �����5����6������7�� ������+#'8&)!�%'-%!'�(!+�-8�� !�%&'�()*!+�&,$�� !�%-'!�.&**+0��98�� !�%&'�()*!+�$-�,-��&$ !'!��-�� !�%-'!�.&**+�� !2�)&,�)-,�(,#!��-�&,�(,�!',&*�%-'!�� '-&��. !'!�:*-);(,<�"&2�,-.���������������3��=���� ���������7������ �������������� � ����������� ��������������7���>����������6������������ �� ��� �������������?�������������������3@!'2�+"&**�$(&"!�!'�%&'�()*!+�.(**�%&++�� '-#< �� !�%-'!�� '-&�+�.(� -#��:*-);(,<0��A () �-8�� !�&:-/!�%'-)!++!+�-))#'+�$!%!,$+�"&(,*2�-,�� !�B%&'�()*!C8-'"&�(-,D�+(E!�'&�(-�&,$�� !�%&'�()*!�)-,)!,�'&�(-,0��F-'!�� '-&�+�&'!�-8�!,�$!+)'(:!$�&+�:!(,<�"()'-G��������������������������� ��������H���������� ��� �IJKLMNOPNMQORSMNTRUOVOWXYOZO WXYOZOVOJQUNOPNMQORSMNTRUOVO[XYZ JTLMNOPNMQORSMNTRUO\O[XYZ���������������� ] _ abcdefg chig cjefgk _l`mnopnqrstuvwxyzw{mnopnqrstuvwxy|}~|�|a|| |}� a a}� � a| ~|mwon�lowqtvq����p xow�����p�sqt�n]on��n�xu���4���������������������� �������������������������� ��?���������6� ��������� ?�� ����������������������������� ��5���3��4������ ������ �����7������������� ����������������������������������������������������������������� ��5���?�����������������?����������������� �4�����H�������� ����������������� ������������7������������>��6��� ������ �����������������������������������?����6��������7���������� ��� �������� ���� �� ������ 3�����7����������������������������������������������������������������� ��������������� !�'!+�'()�(-,+�%'!+!,�!$�:2�� !�%-'!�� '-&�0��9,�!'"!$(&�!�+(E!$�%&'�()*!+�.(**�'!+#*��� �� �� ��� �������������3��������H���6��������7�������������� �����������������������������������6��������������� �� ������������������������������������������ ��5������ ������ �3��������� ���������� ���� �¡� �� ��¡��������¢��¡� �£��¢��¡�¤¥¦§ ©ª«¦¬ª ®¦« ¨©ª«¦°©±§²©ª¦¬©³°¬±² µ¦«¶°©²¦´¨·°¶°¯¤¦«¶ ± ² © « °¥µ²¥© «³°¸ °©·¯®©ª ² ©ª¯ µ¦«¶°©²¦´¹¯¶¦º°¥»¼½°³²·¾§°±ª¤¦«¶ ±µ²¥© «³°¸ °©µ¦«¶°©²¦´ ±¬«µ°³¿«¦³ ±±À¦© Á²ª ®°«©²³¥ ³¦´³ ´©«°©²¦´±½ÂÃÄ ¾« ·¬³ ±©ª¯ °»¦º º°¥¬¯±À¦©´¯³ ±±°«¼ Å´¦ ·°¶°¯¨¹¯¶¦º°¥ Ʋµµ²³¬¥©©¦ « ¶ ¦º ¨ ÇÈÉÊËÌÍÌÎÏÐÑÒÓÔÕÑÒÖÔ×ÑØÙÔÏØÙÚÛÜÝÞËÌÍßàáÞËÌâãÞáÈâÌäÛåËÌáæÌãÛÜãÌÈçèÌÊáéÌêËÞëËÈÉÛÊÛâìÌÞáíîäÇÈÉÊËÌßÌ ÌÚÛÊâËÞÌïÈîËÌæáÞëÈâÛáçÌêÞáíËääÌèËêËçèäÌáçÌêÈÞâÛíÊËÌäÛåËÌðÌêáÞËÌâãÞáÈâÌäÛåËÌÞÈâÛáÌ
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,E FGHEIEFJKLMHMNEGONMEHPNOKHEEEEFEIEFJKLMHMNEGKNHJQRMEEEEEEEEEEEE STUVWXYZ[\V][ _TZabV Xcdeabf Wgheabf iheabfjTZT[VSZkl^e T kmVj [Za nWohVTm fVpVcdqrhhVe T kmbsYatam_V][ _TZahV ch rh uh ih Whh Wch Wrh Wuh Wih chhSk t [_TkmVv[t[UaVjTwaVxaZ[_TkmbfTybz {jVeabfVjTwab|}~E�PME�LM�P�E�J�MENM�MNNMFEHOENMRKHM�EHOEHPME��E�HK�FKNFE�JM�ME�J�M�EKE�NM��M�HR�E��MFEGKNHJQRME�J�MELMK��NMEJ�EHPMEOJREJ�F��HN�EM~�~E�ONE�NK�MREGKQ�E�K�F~�����������������������PMEQOL J�MFEJ�HMNKQHJO�E MH¡MM�EHPMELK��E�KQHON�E¡PJQPEM��MQHEHPME�ORJF�EK�FE¢£¤¥¦¤§©ª«¦¬©ª®¥¢£§©ª¤¤ §°¥±¦¤©ª¦¥§¬²±±¦¥©¬§³©ª¢ ²¥§µ¶· ©¹ ¬ ²£³ ME�MR�EMºGRK�KHON�~ »¼½½¾¿ÀÁ¼ÂÃÄÅÀ¾ÆÂÇÈɾÂüƽÈÀļ¿ÊË̼ž½ÍÎÂżÏÏÂÄ¿À¼ÂüƽÈÀļ¿ÊÐÂξ¾ÑÂÄ¿ÒÈÏļ¿ÓÈÆÀÄÈÅÂÃÄÅÀ¾ÆÂÇÈɾÂǼ¿ÀƼÅÊÄ¿ÒÈÏļ¿ÂξÑÀÌÂƾÎÍǾÎÂÔÕÌÄÖÌÂѼƼÏÄÀÕ×¼¼ÎÂÃÄÅÀ¾ÆÂÇÈɾÂüƽÈÀļ¿ÊǼ¿ÀƼÅÏÂÃÅÍÄÎÂżÏÏ×¼¼ÎÂÃÄÅÀ¾ÆÂÇÈɾÂüƽÈÀļ¿ØÙÄÖÌÂξÑÀ̼ÃÂÄ¿ÒÈÏļ¿ÂÎ;ÂÀ¼ÂżËÂüƽÈÀļ¿ÂѼƼÏÄÀÕÐÂÚ Ì¼Å¾Â½ÍÎÂÇÈ¿ÂÔ¾ÂżÏÀÂÄ¿ÂüƽÈÀļ¿ÏÂËÄÀÌÂÉÂÛÂÜÝÊξѾ¿ÎÄ¿Ö¼¿Â½ÍÎÂÞÂѼƾÂÀÌƼÈÀÂÎÄȽ¾À¾ÆÂÎÄÏÀÆÄÔÍÀļ¿ßÄÖÂàØÜáâ¾ÆÕÂÙÄÖÌÂã¼ÇÉÂӼƼÏÄÀÕÂÞÂӾƽ¾ÈÔÄÅÀÕÙÄÖÌÂäåæçèéçêëìæåíîëìæåíîïðïäåæçèéçê ñ¾ÎÄͽÂò¼ËÂó¼ÅÄÎÏ ßÄÅÀÆÈÀ¾ ô¿Ä¿ÒÈξÎÂƼÇÉäåæçèéçêäåæçèéçê Eõö÷öøùúûüýþÿú�ø�ýüý��ø��ûÿ��ø��ü��þÿ���PME��QQM��EO�EEKEQK�J��EONERJ�MNEQMLM�HKHJO�EJ�EFMGM�FM�H�EKLO��EOHPMNEHPJ����EO�EHPMENMLO�KREO�EHPMEL�FEQK�M~E�PJ�EJ�EO�HM�EKQPJM�MFE �EG�LGJ��E�KNJO��E¡K�PM�EK�F�ONE�GKQMN�EQO�HKJ�J��EFJ�GMN�K�HEKFFJHJ�M�E���N�KQHK�H��~EE�PMENMLO�KREO�EHPMEL�FEQK�ME¤¥© §¥¬©ª¹¥§¦¬§³²©³£¬¬¸ ©¹ ±¦�§²®¤¦³§®¤ °¬§«§¥¦£¹§ª¤©±§¤¥§¬� �������� ����������������������� !"#$%&'#(& �������)*�+������,��-.��������/
��������������� ��� ������������������������������� ���������������������������������������������������������������������������� �������������������������������������� ��������������� ���� ������ ���!"#!$%&'($#$!")*+'',%-).*-/'!*&*0)'(#%--$*('."1*'"'1*-2'.$3.'0")%-"#'4%$5'#6(('%0#*(('!60)-6##*5'72'(%$)"7#*'"55$)$1*(+''8-69*-'4%$5'#6(('!60)-6#'$('"'0*!*(($)2'($0!*'*:!*(($1*'������������������������������������������ ���;�<������������ ������������� ���������������� ��������� �������� ������ ������$0').*'!"($03'$)(*#=+''>.%(/'5*(9$)*').*'5"&"3$03'0")%-*'6=').*'!*&*0)'4%$5'#6((/'$)'$('06)'06-&"##2'"'9-67#*&'($0!*').*'?#)-")*'@ABCDE���������������������� ������������������������������������>.*'*==*!)'6='5-$##$03'"0'61*-'3"%3*'.6#*'60').*'9*-=6-")$03'*=?!$*0!2'$('6='&%!.'3-*")*-'!60!*-0+''>.$('$('$##%()-")*5'$0'?3%-*'FG'H.*-*').*'96H*-'6=').*'9*-=6-")$03'3%0'H"('(%=?!$*0)'=6-').*'1*-)$!"#'9*-=6-")$60(')6'9*0*)-")*').*'!*&*0)'(.*").+''I)'H"('$0(%=?!$*0)#2'96H*-=%#')6'9*0*)-")*').*'!*&*0)'(.*").'$0').*'.6-$J60)"#'5$-*!)$60+''K6)'������ �������������� ��������L����������MN��� ��� ���������������������������������������������������������������������������������������������������������� �������������������������������� ��������$0').*'$5*"#'!"(*+''O0P3"%3*'5-$##$03'6=').*'9"2'J60*'!"0'"16$5').$('9-67#*&'Q?3%-*'FGR+ STUVVUWXYZU[\U]V]_XYT`ab^Uc[T^X[dUTUX]_XU[X [XeU[V]f_XYTghTi_]jUk[l^iUmno npqrXUUi[sYTU]hVVUWXYZU[\U]V]_XYT`\UTUX]_XUc[XeU[WUfUTXceU_Xe[a_Tk[V]f_XYT[k_f_jUtg u^]f_XYTvTwxy_jU[l^iUb]Yii[zYX[kY_fUXU]{������� ��� ������ �������������� ��������� �����������0")%-"##2'=-"!)%-*5+''>.*'4%$5'#6(('!60)-6#'"55$)$1*'"55*5')6').*'!*&*0)'(#%--2'H$##'7*'$0*==*!)$1*'$0'(%!.'H$5*-'5$"&*)*-'=*")%-*('|'(6'!*&*0)'H$##'46H'$0)6').*'=-"!)%-*(/'������������� ����������������� �}���������������������������������� ���������������~������������������������ ������������������������������������������������������������������������� �;��������H*##'$046H+����������� �������������������������� ���������������������� �}������������������ ����������������������������������}������� ��������������������� ���� ��������������������������������������������������������������������������� ���������������������������������������������������������������������������������
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EFGFHIJKLMNOJPHQMLMRSHQTKOPRHUSKVJKMNOPRWXYZ[Y\] \_abcZ[\ dYeeZ\YefgeZabZ_bZ [YbZ[Y\] \_abZYbdgeYhZiaXabZ_ZgiZjklmknopqprstuvwktvxtxlnyrzlk{t|lnpw}tnw{tlkmwnwr}tvxtr~ktjklxvlnrpw|ty~nl|kt��|zlkt����tt�~p}t�{nmn|k{�tnlkntlk}zqr}tpwtntlk{zyk{tpw�v�tpwrvtr~ktjklxvlnrpvw�tt�v�k�kl�tr~ktmkn}zlk{t�kqqt�lv{zyrp�prst�w{k�t�p�k�t�zp{tjlv{zyrpvwtjkltzwprt�kqqt{ln�{v�w�te �Ya�YeZabd\Y_eYeZiaXZa�YZ_eZXYZiYggZaeZ[\ hfdYh�ZZWXaeZ�dgY_babcZf[�Z[\ dYeeZaeZ]YbZ_ed\a�YhZ ZXYZ\Y� �_gZ]Z[Y\] \_abcZhY�\aeZ�dX_\cYZhY�\ae�Z\ d�Z]\_c�YbeZnw{tr~ktqv�tjklmknopqprstylz}~k{tuvwk��tnqqtvxt�~py~tlk{zyktr~kt�kqqtpw�v��tt�~p}t\Y� �_gZabd\Y_eYeZXYZ\_be�aeea�aga�Z�YiYYbZXYZiYggZ_bhZXYZ] \�_ab�Z�f\XY\�Z}vmktvlp|pwnqqstyvmjqkrkqstoqvy�k{tjklxvlnrpvw}tmnstvjkwtrvt�v�tn}tr~kt�kqqt{ln�{v�wtpwylkn}k}�tt�~p}t�yqknwtzj�tjlvyk}}tynwtoktnyykqklnrk{tost}jkyp�yt�kqqt\Y_�YbeZefdXZ_e��p�t �ny�t�v�pw|t�jlv{zyrpvwtzw{klt~p|~t{ln�{v�w��aa�Z dahaeabcZ� Zhaee g�YZXYZa�[_a\abcZ�_Y\a_g��aaa�Z¡Y\] \_abcZiaXZXYZiYggZfbhY\�_g_bdYZ�XYZd\feXYhZ_bhZa�[_a\YhZ[Y\�Y_�aga�ZZZ \ d�Z_eZiYggZ_eZXYZdX_\cYZhY�\aeZ_\YZ\Y� �YhZ_eZe bZ_eZaZaeZ] \�Yh�ZZZZZZZZZZZZZZZZZZ ¢£¤¥¦§©ª«¬®°©±²³©ª¬ £µ©¶·£ ¬©±¤©¦¹²²º» ¼±¦²µ¬½±¦§®±²µ¾¶·£ ¬©±¤©¦¹²²º»¿©ª®§©ª¬©±¤©¦¹²²º¬¿£§À Áµª¦¤¦¾©ª¬¿£§À©±Â£±¦ ²£µ¬¢Ã¦±¾©¬Ä©¹±²Å¢¦Å²µ¾¬Æ¬¢©¤©µ¨Ç¥©µ¬©±Â£±¦ ²£µ¬È®µµ©¯EFÉFHIJKLMNOJPHQMLMRSHQTKOPRHUKJÊTËNOJPWXYZg_eZeYdabZhaedfeeYhZ_Zd_eYZabZiXadXZXYZiYggZ[\ hfda�a�ZabhYÌZabd\Y_eYhZiaXZa�YZhf\abcZb \�_gZiYggZ[\ hfdab�ZZÍb] \fb_Yg��ZXaeZ]_� f\_�gYZe_YZ]Z_]]_a\eZ]\YÎfYbg�Zh YeZb Zddf\ZÏZ\YhfdabeZabZiYggZ[\ hfda�a�ZabhYÌZ]YbZddf\ZiXYbZ_ZiYggZaeZbZ[\ hfdab�ZZW�[ad_gZ[\ hfdabZ] \�_abZh_�_cYZ[XYb �Yb_ZX_ZgY_hZ`ZefdXZ\YhfdabeZabZiYggZ[\ hfda�a�Z_\Y�Ð�Z�abYeZ� �Y�Yb��tÑ}ktvxtpwyvmjnrpoqkt�vl�v�klt�zp{}Ò�ZÓb \c_badZ_bhZ\c_badZed_gYZ] \�_abÔ�ZÕ_dY\a_Ö�Z¡\Yeef\YZ\Yhfdab×�ZØ a�fg_abÙÚÛÜÝÞßàáßâãäÞåæçÚãßèéßêæåæÛÞêßÝÞÛÚèëßæÝèÜëêßìÞÝéèÝæçÚèëßéèÝåÞêßêÜÝÚëÛßçäÞßìÞÝéèÝæçÚèëßìÝèãÞíí
��������������������� ������ ������������������������������������������� ��!�"��#$%&�'�($#$)��*�"&'� �+ �,�#�'%-./0123/450670489::08;3/79:0<97=;>:/40?>:9@4A0B/:C4<9740/=>D0E.;>.097/0B6F3C0630=./0���G���G��������������� �����H���I�������H�G�����J������������K����������E;=.0=./07/4/7L6;70MF;C40CF7;3N0N/6:6N;>9:0=;8/O00P68/=;8/4A0=./0>63C;=;6340E/7/04F>.0=.9=0=./023/407/89;3/C0940C;4>7/=/0<97=;>:/4079=./70=.930Q/;3N0>/8/3=/C0=6N/=./7093C0=60=./0493C0N79;3076>R0B9Q7;>O00-./0MF;C40;30=./0<67/04<9>/40CF7;3N0=./4/0N/6:6N;>9:0=;8/0</7;6C4097/0/;=./704=9=;>0670M6E09=0L/7@04:6E0L/:6>;=;/4O00S6E/L/7A0=.;40/TF;:;Q7;F804=9=/0;40C;4=F7Q/C063>/090E/::0;40C7;::/C093C0<:9>/C0630<76CF>=;630U0.;N.0MF;C0M6E0L/:6>;=;/4097/0/V</7;/3>/C0;30=./03/970E/::Q67/07/N;63O00W=0E9406Q4/7L/C0;30=./02/:C0���X�I�������Y�G��� ��������������Y����X�����������������������X�������������������� ����I������J��Y���������� �������� ����I�������������Z��������I����[� ��������������K��H�Y����������������������� \]_ ab_\cd_e_fghigjk]]ld_mgij]cn_oacnljclohl]pj_hjlg_bq]pq_lh]rrghn_s]cgn_iatgigclsah_paik]cgu_a]_v_bjlgh_saboahg_fghigjk]ld ` ab_wgap]ld `]r_xyz{\]_v_bjlgh_sab mgu|pgu_gssgpl]tg_}ghigjk]]ldu|g_la_hgjl]tg_}ghigjk]]ld_gssgplny~�I��������� ����������������������������Y������Y������I�����������I��������������������I���G�������������J���� ������������������������������������������E9=/70M6EO00W=0E9406Q4/7L/C0=.9=0=./0</78/9Q;:;=@0=606;:0M6E07/89;3/C0>634=93=A0/L/309=07/94639Q:/0M6E0L/:6>;=;/4A0E.;:/0=./0</78/9Q;:;=@0Q/N930=60Q/07/CF>/C09=090>7;=;>9:0M6E0L/:6>;=@0E./30=./0>67/0E9404FQ�/>=/C0=6090>68Q;3/C06;:�E9=/70M6E07/N;8/0?2NF7/0��DO00�F7=./70/V</7;8/3=40E/7/0</7B678/C0;30E.;>.0=./09>=F9:0M6E0=.76FN.0=./0<67/04<9>/40>6F:C0Q/06Q4/7L/C0L;4F9::@0U02NF7/0��904.6E40=.9=0=./0489::A0:664/0<97=;>:/40<7/4/3=0630=./0493C0N79;304F7B9>/07/89;30;30<:9>/0CF7;3N06;:0M6EO00S6E/L/7A0=./4/023/40Q/>68/086Q;:/063>/0=./0E/==;3N0<.94/0?E9=/7A0;30=.;40>94/D0;40M6E;3N09=09079=/0N7/9=/70=.930=.;40>7;=;>9:0L9:F/0?2NF7/0��QDO00� 6Q;:;49=;6306B0=./023/409::6E40=./80=60��J���������I����������� ����X��G�������������������������G�������I����������I���������������������X����H����������������������G�������I���������Y���� �����������J��Y����������������������J�����I�����������������������������:9Q679=67@0>67/0=/4=4093C0=./02/:C08/94F7/8/3=4O ����������������������������������� ¡¢£¡¤¥¦§¦©
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EEEEEEEEEEEEEEEEEEEEE FGHIJKKLMNOP LMNQRSTTJUGHNVSTTJUGHNOP LMNQOWOP NMOXGKNYXHJZGK[OFGH XG[O\] \L_LK`OaQLGKEEEEEEEEEEEEEEEEEEEEEEEEEEE bcdefghij klmnokgpiqnkfgrstuvwxyzvs bcdefghij klmnbcdefghidfcbfhi{|}{~�d~fcm�ij klmnij mlibfgm�i�e�f��i��~�m��mglc�icmkp�ildi�dnmil�ndkli~cd��fgh� ���d~fcm�ij klmnij mlibfgm������E��������E��E����������E������ ¡¢£¤¥£¦§¦�¢£ £© £ ¦ª§¥ ¢« ¦¬¥ £¥® §¥®¥£¦ ¢£ ¦¬¥ ® £© °§�£ ®ª§«¤¥® £© E ±��²��E�²�E³�������� ´µ¢¶ ·¥µ¢¤�¦ ¢« ¦¬¥ ¹ª�© ¶¥¦¦�£° ¬ ®¥����� º ¥¦¦»�µ�¦ ®¦¦¥ ¢« ¦¬¥ ¼£¥®½¾�����¿�E��¿�À���E±²��²E²¿��E����EÁ�À��E��E��E�ÁÁ������E��E���À��E�²���E�ÁÁ����E����À���Â�ÃE Ä���������ÅE���±E��¿�EÀ³E�ÁE�²�E±���ÆEEDz��E��E���¿À��E�����¿�E�ȳ��������E��E ¦¬¢®¥ ©¥®¤§�»¥© »¢·¥ ¬ ·¥ ®¬¢¶£ ¦¬ ¦ ®É µµ¥§ ¤¬ £°¥ �£ ¹¢¶ ·¥µ¢¤�¦EEE ���À���E�²�E�À����E�ÁE³¿�������E�²¿�E¿��E���������E¿�E¿��E���E����ÅE��¿���ÊEEE ��E¿E��±��E�²¿���E�ÁE����Ë¿Ê�E¿�E�²�E³���E�²��¿��ÆEEÌ�¿��E�����¿���E��E §¢©ª¤¦�¢£ µµ¢¶ ¦¬¥ £¥® ¦¢ »¥ ¼¤µ¥ £¥© ¢ª¦½ ¢« ¦¬¥ «¢§É ¦�¢£ ¦ µ¢¶EEE �¿���E¿��E��������¿�����ÃÍE��E�²¿�E�²�E�������E±���E³���À�����EE�¿��E¿��EEE ��¿±��±�ÃE�¿�E��E¿�²�����E±²���E�¿���¿����ÊE�²�E��¿�α�������E���ËEEE ³����¿������E��E³��������ÊE³���E�²��¿�E����Ê��ÊÆÂ��ÃE�����E�������E���¿������E�¿���E��E��������¿���ÊE�²�E��¿��E³¿�������E��EEE �²�E���������E���ËE±��²E�����ÏÀ���E���À�����E��EÁ���¿����E³����¿������ÃE��EEE �²�EÀ��E�ÁE�À�Á¿��¿���E��E�²¿�Ê�E�²�EÁ���¿����E±���¿������ÐÑÒÓÔÕÖ×ØÙÚÙÛÕÔÖÜÕÛÖÝÞÕßÖÙÔÕÖÑààáâÑãÕÖäÓÔÑÞÒÖáÑãÖåÔáäÓæÛÑáÞÖáÞãçÐÑÒÓÔÕÖ×ØâÚÙÛÕÔÖÜÕÛÖÝÞÕßÖâÕæáàÕÖàáâÑãÕÖÜÑÛèÖÜÙÛÕÔÖåÔáäÓæÛÑáÞéÖãÕÙäÑÞÒÖÛáÖåáÔÕÖÛèÔáÙÛÖâãáæêÙÒÕë
���������������� ���������������������� ����������������������� ������������� !"#$"#�% ��&��'!%$�"�(�%%!)$*$%+�%���$*�(�%��!"�����*%�$"�,-"��.�'�/�'�"%�0���%��% ��1�(��&��$*2345464789:;<=>79?7@9;ABCD9E7FBABG=7H7I9AJK=CD9E7BEL7M 9;N9O=;7@K:DL>P ��1�$0��%��( $� �% ��&��'!%$�"�$���QR���0�0��$"#�% ����$#$"!*���'R*�%$�"�!"0���S��T�� ����U������������ ������S������ ��������V��� �����������V����������S���W�W�� ���������������������� WXY����������S����V�� �������� ���������������Z������ [W\W]W] W_����������V�����S�S�V� ������������������������������������ �� ����������W%�� �(��% ��($0��%��!"#���&���!�%$�"��($% ��$*-�*0�)�$"��a�P $��$��$**��%�!%�0�$"�-#����bc�( ����% ��R��'�!)$*$%+�0�#�!0!%$�"��&�!��'��%$%���*!+���"%!$"$"#��!"0�%�"��������0���%��R!��!#���&�!����$����&�)�$"���$��� �("a�P ��-#����� �(��% !%de!f� gh�(%�!''�"$�'�� *��$0��eij[k*f�)�$"��1�(�#$/���% �� $# ��%�R��'�!)$*$%+�� ��������� �e)f� k !"#$"#�%��!�lh�(%���0$�'�� *��$0��ei!k*f�)�$"��#$/���!���0���0�R��'�!)$*$%+X� Y���������S����V�������� ������ ��������� ������������������� ��"��"%�!%$�"�$����0���0�%��gh�(%�i!k*e0f� m��� �e0$�%$**�0f�(!%���1�(�#!/��% ��*�(��%�R��'�!)$*$%+��&�!**�% ��!)�/�e�f� m*�(��&�!�-"!*�1�� ��&�gh�(%�ij[n�����S�� �����o���������o� �����o���� ������S����V���� �S������������������o� ��X��o���������S����VWY�� �������� ����o��������������� �������S��S�� �������� ����� � � pqrstuvwxyy y vyy xyy zyy {yy |yy }yy~���s������sp�����~�����q�q���s�� x� s��{��|� s����x� s����p����s� ������� ��� ��� ��� ���Y���������S����V�������� ��������SV����S������� ����V����������S����� ������V���������Wn��V���� �������V �o������V����o������������ �����SV� �����S������� X�������������������������� ��������WY�����V������ oX ��S������V�����������S���S�� ���� �������SV�������� �S���S��� �������������� �����S�� �W��o����]������������������������������������ ����� ¡¢£¡¤¥¦�§ ©¡¦ �ª «�©�¬¥¡ �� ¬�® ¡®°¡¡±ª §¬�¬ ¡§®� ¡²³ ¡®²� �¡µ ��®�±¡¶·¹º»¼
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EFGHIJKILMNHOPHQOEJRRLMNHIKJSRTHUFLKGHEFGHRUGKKLVWHXOPHJKKHEFPGGHIKJSRHLRHPGYMIGYHJEHEFGHFLWFGPHZPLVGH[\J]KHIOVIGVEPJELOVR_\`_HaUGKKLVWHOXHEFGHIKJSRHVOEHOVKSHNGJVRHEFJEHEFGSH[QJPELJKKS HZKOIbHEFGHQOPGHRQJIGcHdefgfhijgklmngkllnogfhigplkjgqkrfsplimgfngdriktguriigkfglnoirgvnogwilnpsfsimxglikysz{gfngZKOIbJWGHJEHEFGHQOPGHEFPOJER_HH OEFHNGIFJVLRNRHKGJYHEOHPGYMIELOVRHLVHQGPNGJZLKLES_|LWMPGH}~HLNQKLGRHEFJEHbGGQLVWHEFGHIKJSHLVHEFGHQOEJRRLMNHXOPNHLRHQPGXGPPGYHIONQJPGYHEOHEFGHIJKILMNHOPHROYLMNH[EFGHUOPREH�JPLGELGR_H� JVSHEGRERHJVYHREMYLGRHOXHEFGHMVYGPKSLVWHqhjmspmghkwigmhnozgfhkfgg��gofgqnfkmmse�gphlnrsyig���l�g{swimgfhiglikmfgmoillsz{�gg�fgLRHEFMRHEFGHQPGXGPPGYHZPLVGHIONQORLELOVHXOPHVOV�YJNJWLVWHUOPbO�GPHJVYHIONQKGELOVHvesym�gg�hsmgiu�pkpjgnug��lgsmgyeigfngfhigms�ignugfhigqnfkmmse�gpkfsnzgohsphgkllnomgsfgEOHGJRLKSHLVRGPEHLERGKXHLVEOHEFGHIKJSHKJEELIG_HH�FGHJNNOVLMNHIJELOVHLRHOXHJHRLNLKJPHRL�GH�ghizpigsfmgkdslsfjgfng{swigfhighs{himfgqir�ikdslsfjg�ikmeri�izfgszg�{erig���HHHHHHHHHHHHHHHHHH ������������� ������ �¡¢�£�¤��¥ ¦���§ ©ª«¬«ª©¬©ª¬ª¬ª ®� ¤����¯�£ �¤�� ª °± �²��³ § ¬± �²��³�¡¯�µ ��µ�¶¤�¤�·¡¤� �µ ��µ�¶¤�¤����¶�µ ��µ�¶¤�¤�¸NNOVLMNHIFKOPLYGHLRHMRGYHJRHJHVOV�YJNJWLVWHZPLVGHXOPHRQGILJKHRLEMJELOVRTHRMIFHJRHYMPLVWHJILYHRELNMKJELOVRTHUFGVHLERHRMZRELEMELOVHZSHQOEJRRLMNHIFKOPLYGHUOMKYHKGJYHEOHEFGHXOPNJELOVHOXHYJNJWLVWHQPGILQLEJEGR_¹º»º¼º½¾¿ÀÁÂÃĽ¿Å½Æ¿ÁÇÈÉʿ˽ÌÈÇÈÍýν¾ÂÈÏÃaIJKGHPGXGPRHEOHEFGHQPGILQLEJELOVTHLVHZOEFHEFGHUGKKHLERGKXHOPHEFGHVGJPHUGKKZOPGHXOPNJELOVTHOXHOPWJVLIHOPHLVOPWJVLIHNJEGPLJK_HH�FLRHRIJKGHNJSHZKOIbHEFGHQOPGRHLVHEFGHXOPNJELOVHnrgiwizgfhigqirunrkfsnzmg�riyepsz{goillgszvno�gnrgdlnptgfhigfedsz{g�riyepsz{goillgnefvno�gOPHLNQGYGHEFGHOQGPJELOVHOXHUGKKHJIIGRROPLGRHG_W_HQPG�GVEHEFGHOQGPJELOVHOXHEFGHaMPXJIGH]OVEPOKKGYHaMZHaMPXJIGHaJXGESHÐJK�GH[a]aaaÐ _HHÑVHEFGHKJEEGPHIJRGHEFGHRJXGHOQGPJELOVHOXHEFGHUGKKHUOMKYHZGHIONQPONLRGY_HH|LWMPGH}ÒHRFOURHJHEMZLVWHUFLIFTHnzgripnwirjgurn�gfhigqrnyepsz{goillÓgokmgunezygfnghkwigfhigkrikgkwkslkdligfngvnogriyepiygdjgÔÕ�gyeigfngkzgsznr{kzspgmpkli�Ö×ØÙÚÛÜÝÞÜßàÛáá×âØÜãÙÛÜäåÜàæäÛÚÜæçèåÚéä×åâÜ×èÜêåâäÚåááÛãÜçëÜêáæëÜäëéÛÜæâãÜçÚ×âÛÜêåìéåè×ä×åâ
����������������������������������������� ����� ���� ���� ������ ���� �� ������������� ����� �� ������� !!�" #$#�%&$�'($")')% %)*��)#�%()++$($,�-.� �"&$/)" !�)�#% -)!)%.�%& %�& #�-$$��"($ %$,�0123245367829:29582:;<7<63=28>?<=<0;<?@24:6A<9<:6B2345<8C8A20129582D:;@39<:62E?<A2*F$(�+$*!*+)" !�%)/$#G�H&)#�" ��-$�,I$�%*JK<L2 M2A84;83B82<6298@N8;39?;82A?;<672E:O2?N295829?0<67K<<L2 M2A84;83B82<62N;8BB?;82A?;<6729582;8B8;C:<;2A8N=89<:62:;2E:O2?N2<6295829?0<67K<<<L2P <Q<672O<95232458@<43==12<64:@N39<0=82E?<A28R7R24:663982O398;2O<952<6S849<:6��� T %$(�*(�T %$(�U(*/� �#$' ( %$�($#$(F*)(HT*�%.'$#�*U�*(+ �)"�#" !$#�V�T W� �,� #'& !%$�$#�V� ($�$�"*I�%$($,�T&)!$� �/I!%)%I,$�*U�)�*(+ �)"�#" !$#�& F$�-$$��*-#$(F$,G��XYZY[Y\Y] _ abcd]_e]f_aghij_k]lhghmc]n]ok_amhkjb]bhpcdq�T),$�( �+$�*U�)�*(+ �)"�#" !$#�& F$�-$$��$�"*I�%$($,�,I()�+�T*(!,�T),$�'(*,I"%)*��*'$( %)*�#G��H&$#$�)�"!I,$JK<L2 r;84<N<939<:62:D2s3t=2B3=92A?829:24::=<672:D2O8==2E?<AB236Au:;28C3N:;39<:62:D2958��� T %$(�)�%*�%&$�+ #�'& #$�,I()�+�'(*,I"%)*��*U�# %I( %$,�-()�$#�%*�%&$�#I(U "$G� v w!�x#*!�y�z �v w!{�x#*!),yw !")I/�#I!'& %$�" ��-$�'($")')% %$,�,I$�%*�"**!)�+G�H&)#�)#�,I$�%*�%&$�U "%�%& %�-$!*T� �"()%)" !�%$/'$( %I($�%&$�!$##�#*!I-!$�)#*/*('&|� �&.,()%$|�-$"*/$#�%&$�$}I)!)-()I/�U*(/�%& %�)#�'($")')% %$,G� w ~~�~����VV�z �w ���{�x#*!),yx))y� �($")')% %)*��*U�" !")I/�" (-*� %$�,I$�%*�'($##I($�($,I"%)*���%&*I+&�*�"$�22 9582E?<AB23;82<6295829?0<67295812?6A8;7:2324:@0<68A28DD8492:D2N;8BB?;8236A��� %$/'$( %I($�($,I"%)*�G� w x�w��y��z �w w��{�~�w����~���� �������������������������������������������������������������������������
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EFFFGHIJKLKMNHLOMNPLHOMJHQPHPJORKJSPTPUHPVWVHLKNXYKTHXTPOFXFSMSFRJHZTR[H\PT]HLRKTHHH WMLH PNNLHYMLHQPPJHRQLPT\PUHFJHLP\PTMNHXTRUKOFJWHMTPMLVHH_`VHaTPLLKTPHSTFWWPTPUHLOMNPHZRT[MSFRJHFLHLPNZHMWWTM\MSFJWHFVPVHMJHPbSTMHXTPLLKTPHUTRXHFLHOTPMSPUHMSHSYPHRJLPSHRZHLOMNPHZRT[MSFRJVHHcYFLHNPMULHSRHMHZKTSYPTHUTRXHFJHXTPLLKTPHMLHSYPH PNNHRXPTMSRTHSTFPLHSRH[MFJSMFJHSYPHSMTWPSH PNNHXTRUKOSFRJVHHcYFLHZKTSYPTHXTPLLKTPHUTRXHOTPMSPLHPbSTMHLOMNFJWHXTRQNP[LVEF\GHd FbFJWHRZHS RHTPLPT\RFTHQTFJPLHEJRT[MNN]HFJHSYPHXTRUKOSFRJHSKQFJWGHRTHQPS PPJHee feghihgjklgemglnheohpqpeglrselnetfgluvwefnxefnelnyhrzhxe{ulxeohpqpeihfe|fzhgwHHH OMJHNPMUHSRHXTPOFXFSMSFRJVH `M}}eoghihgjklge{ulxwe~e�����EFJ�POSPUHLPMH MSPTGH� H M����HELRNFUG�SYPTHFJLRNKQNPH[MSPTFMNLHPVWVH�T���HMTPHZRT[PUHQ]HMHLF[FNMTHXTROPLLVHH�KOYHLOMNPLHMTPHZTP�KPJSN]HZRKJUHSRHQPHORJSM[FJMSPUHQ]HNR HNP\PNHTMUFRMOSF\FS]VHcYFLHURPLHJRSHJRT[MNN]HTPXTPLPJSHMHYPMNSYHYM�MTUHKJNPLLHMHUT]HLOMNPHLM[XNPHFLHUTFNNPUHEKJXTRSPOSPUHQTPMSYFJWHRZHSYPHUKLSHFLHYM�MTURKLHUKPHSRHTMUFRMOSF\PHXMTSFONPLHNRUWFJWHFJHSYPHNKJWLGV� PNNHXTRUKOSF\FS]HOMJHRJN]HQPHTPPLSMQNFLYPUHQ]HTP[R\FJWHSYPHLOMNPH[MSPTFMNVHcYFLHFLHMOYFP\PUHQ]�EFGH ����������HSYPHLOMNPHZTR[HFSLHXTPOFXFSMSFRJHXRFJSHEFJHSYPHZRT[MSFRJ�H PNNHRTHHH ZMOFNFSFPLGHKLFJWHMHLKFSMQNPHLRN\PJS�HPVW�H EMGH _M�NHLMNSHOMJHQPHUFLLRN\PUHQ]H MSPTH EQGH �MNOFK[H�MTQRJMSPHE�M���GHOMJHQPHUFLLRN\PUHQ]HY]UTROYNRTFOHMOFUH EOGH `MTFK[H�KNXYMSPHE M���weliejhg�exl��ru�zezkexliik�jhpe�kvvhgrlf���HHH H M\MFNMQNPHLRN\PJSLHSPJUHSRHQPHPbXPJLF\PHMJUHOMJHRJN]HUFLLRN\PHMHNF[FSPUHH H M[RKJSHRZH[MSPTFMNVEFFGH��������ekuzezshe�xl��ru�zezkexliik�jh�eirf�hierku��hxe|lzs el�enhrhiifg� HHH TPXPTZRTMSFJWHSRHQ]XMLLHXPTZRTMSFRJLHQNRO¡PUHQ]HLOMNP¢HYMLHQPPJHZRKJUHSRHQPHPZZPOSF\PVEFFFGHcYPHXTPZPTTPUHSPOYJF�KPHFLHSRH£¤¥£¦£§SYPHZRT[MSFRJHRZHSYPHLOMNPHQ]HOFTOKNMSFRJHHH EFJHSYPH PNNGHRTHFJ�POSFRJHFJSRHSYPHZRT[MSFRJHRZHMHLOMNPHFJYFQFSRTVHH©JHSYPHNMSSPTHHH OMLPHSYPHFJYFQFSRTHMQLRTQLHRJSRHSYPHZRT[MSFRJHLR[PHUFLSMJOPHZTR[HSYPH PNNQRTPHee fnxelie�gkxurhxei�k|��e|lzsezshe�gkxurhxe{ulxiekjhgefe�hglkxek�evfn�evknzsiHHH �HXTRSPOSFJWHSYPHZRT[MSFRJ�H PNNHMJUHZMOFNFSFPLHZTR[HUM[MWPHQ]HLOMNPHXTPOFXFSMSFRJVHHH cYPHFJYFQFSRTH RT¡LHQ]HXTP\PJSFJWHSYPHFJFSFMNN]HZRT[PU�HL[MNNHLOMNPHXMTSFONPLHHH EJKONPFGHZTR[HMWWNR[PTMSFJWHMJUHZRT[FJWH[MLLF\PHLOMNPHOT]LSMNLHFVPVHSYPH[FJKSPHHH XMTSFONPLHRZHSYPHLOMNFJWH[FJPTMNHMTPHTPSMFJPUHFJHLKLXPJLFRJVHH©SHLYRKNUHQPHHH P[XYMLFLPUHSYMSHSYPHFJYFQFSRTHFLHLNR FJWHUR JHSYPH¡FJPSFOLHRZHLOMNPHMWWNR[PTMSFRJ�HHH ¤ª§HSYPHSYPT[RU]JM[FOLHRZHLOMNPHZRT[MSFRJV
������������� ����������������� ��������������� ����� �����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aV CP>JC>H=>C>HACILOGEA>ACI@?BAC<=OGEAC1&%!,#*!1)# #,(/*],)+*'#&!0&/0 9+(#/ ,,342(#- 7#*,#!)%/ ++"#6(01#*!#,)+&1*)!#*!#12(#$%&'(#)*+#9"#(!,&%*!;#12 1#12(#1(/0(% 1&%(#*,#/ *!1 *!('# 9)8(#12(#$+)&'#0)*!1# 1# ++#1*/(,3##42*,#*!8)+8(,#12(#*!,&+ 1*)!#).#bLPCEFNA<C>=Nc<CA>IC=<CPAEEC=<C>HAC?<ACLRCHA=>A@<CPHA@AC=GG@LG@F=>AZCCdA=>ABCI@?BAC)*+#1 !6(%,# %(#&,('#.)%#,( #)%#%) '#1% !,0)%13##42(#% 1(#).# ;;+)/(% 1*)!#).#12(#- 7#$%",1 +,#$ !# +,)#9(#%('&$('#9"#12(# ''*1*)!#).# !#*!2*9*1)%#]#)!$(# ; *!#6*!(1*$,#).#12(#G@AIFGF>=>FLNCG@LIA<<CI=NCMACFNb?ANIABeCNL>C>HAC>HA@OLBKN=OFI<ZCCf@LGA@COLBAEEFNQC).#12(#1(/0(% 1&%(#'*,1%*9&1*)!#*!#12(#-(++,# !'#. $*+*1*(,#!((',#1)#9(#0(%.)%/('#1)#(!,&%(#12 1#12*,#/*!*/&/#1(/0(% 1&%(#*,#/ *!1 *!('# 1# ++#1*/(,3##42(#/ *!1(! !$(#).#12*,#/*!*/&/#1(/0(% 1&%(#*,#)98*)&,+"#( ,*(%#-2*+(#12(#-(++,#=@ACbLPFNQCgC<FNIACHA=>CF<CMAFNQCILN>FN?=EEKC<?GGEFABC>LC>HAC<K<>AOZCDhLNQC>A@OJC<H?>CBLPN<C@AG@A<AN>C=CG@LMEAOCgCA<GAIF=EEKCRL@CELNQCbLPCEFNA<CL@CGFGAEFNA<ZCC[HACG@LMEAOC*,# $$(!1& 1('#.)%#,&9,( #-(++,5#,*!$(#- 1(%#*,# !#(..($1*8(#/('*&/#.)%#2( 1#1% !,.(%3##[HACELNQCbLPCEFNA<C=NBCELPCP=>A@C>AOGA@=>?@A<CDijklJC=<<LIF=>ABCPF>HCOLBA@NC, 1(++*1(#'(8(+)0/(!1,#*!#'((0#- 1(%,5#%(0%(,(!1# #$2 ++(!;(#12 1#*,#$&%%(!1+"#9(*!;#%(,( %$2('#(3;3#'(8(+)0/(!1#).#/)%(#(..($1*8(#*!,&+ 1*)!#/ 1(%* +,3##:!#12(#/( !1*/(5#,&*1 9+(#)0(% 1*!;#0%)$('&%(,#!(('#1)#9(#'(8(+)0('#1)#(!,&%(#12 1#12(#+*!(#')(,#!)1#9($)/(#$)/0+(1(+"#9+)$6('# !'#2 ,#1)#9(# 9 !')!('#W(3;3#0&%;(#12(#+*!(#).#$%&'(#)*+#0%*)%#1)#,2&1')-!X3##42*,#1"0(#).#0*0(+*!(#9+)$6 ;(#2 ,# +%( '"#2 00(!('#)!#/)%(#>H=NCLNACLII=<FLNZCd=@BEKC=NKCLG>FLN<CAmF<>CLNIAC>HACG@A<<?@AC@An?F@ABC>LC@A<>=@>CbLPC %(#;%( 1(%#12 !#12(# 8 *+ 9+(#0&/0,#$ !#;(!(% 1(#)%5#/)%(#0 %1*$&+ %+"5#(7$((',#12(#0*0(+*!(#1(,1#0%(,,&%(3#:/0%)8('##W0*0(]*!]0*0(X#*!,&+ 1*)!#,",1(/,# ,#-(++# ,#12(#HA=>FNQCLRCbLPCEFNA<CMKC=CP=>A@Co=IcA>CL@CAEAI>@FI=EC>@=IFNQC=@ACMAFNQCBA_AELGABZCp =mCPFEEC>KGFI=EEKC<>=@>C>LCRL@OCLNC>HACFNNA@C<?@R=IACLRC>HAC>?MFNQCL@C>HACbLPCEFNAeC<FNIAC12*,#*,#12(#$)+'(,1#0)*!13##q!$(#.)%/('5#12(#- 7#$ !#9(#%(/)8('#9"rW*X# � ($2 !*$ +# $1*)!#W,$% 0*!;XW**X#s*,,)+8*!;#*!# ]# 2)1#$%&'(#)*+#W$2( 05#%( '*+"# 8 *+ 9+(X# ## # ]# ,)+8(!1,#W/)%(#(70(!,*8(5#9&1#$ !#).#'*,,)+8(#2*;2(%### ## # # - 7#$)!$(!1% 1*)!3#� "#9(#2"'%)$ %9)!#)%#- 1(%#9 ,('3X
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EFFGHIJKLMNOPQPJRSTUVWXTUYZ[\SW]TVT[_UTaT[UbUWVUWXb \Sbc[dUefcXfVbc[dU[b]fgUXWTfW]UhSfiSU\ Tif\fWT[UabXU[bXTUicgT[jUkUaTlcTV]mUibVWfV[UWUSfnSUXc]foW]TVUXTW]UibViTV WfbVdUUU[ciSUW[UpWVWgfcXjUUkUf[UXWgTUc\UbaUWUibX\]TqUXfq c TUbaUW[\SW]TVT[dUT[fV[UWVgUXW]TVT[UhSfiSUhTTUb fnfVW]]mU\ T[TVUfVUSTUicgTUbf]UcVgTUb fnfVW]UT[TobfUibVgffbV[UW[UWUXTW[We]TUib]]bfgW]Ugf[\T[fbVjUURSTU\ Tif\fWfbVU\ biT[[Uf[Urstuuvsvwxyzx{sv||}svxsvw}~rt��|x�x�|{���rv�vx{sv~t{tr�rt��xt|x��rv�x�s|rx�y|vs�vwx�v�sxSTUecee]TU\bfVUfjTjUSTUiSWVnTUfVUicgTUbf]UibX\b[ffbVUgcTUbUSTUTXboW]UbaU[bXTUbaUSTU]bhTUXb]Tic]WUhTfnSU[\TifT[UabXUSTUicgTUbf]UgT[Wef]f[T[USTUib]]bfgW]Ugf[\T[fbVUSWUXWfVWfVTgUSTUW[\SW]TVfiUXWTfW]UfVU[c[\TV[fbVjUU�Web Wb mUTq\TfXTV[UfVUhSfiSUSTUTaaTiUbaU\ T[[c TUbVUSTU[Wef]fmUbaUWVUW[\SW]TVfiUicgTUhW[U[cgfTgUSWoTU[SbhVUSWUSTTUWTUhbUTnfbV[UbaU[Wef]fmjUU�VTUbaUST[TUf[UVb XW]]mUWeboTUSTUecee]TU\bfVU\ T[[c TU�W[Ugf[ic[[TgUTW]fT�UhSf]TUSTTUf[UWU[TibVgUTnfbVUbaU[Wef]fmUWU]bhU\ T[[c T[jUUZ[\SW]TVT[UWTUSc[UXWfV]mU\ Tif\fWTgUWUfV TXTgfWTU\ T[[c T[jUURSTUaWiUSWUSTTUf[UWU[TibVgUWTWUbaUY[b]cef]fm_Uf[UbaU]fXfTgU\ WifiW]Uc[TUeTiWc[TUW[\SW]TVT[UgbUVbUTW[f]mUYTgf[[b]oT_jUURSf[Uf[UeTiWc[TUSTmUhTTUb fnfVW]]mU\ T[TVUfVUSTUicgTUbf]UW[UWUib]]bfgW]Ugf[\T[fbVdUWST USWVUfVU[b]cfbVU�W[Uf[USTUiW[TUbaUhWq�j��vx������t�ux|}���st|v|x�{vs�rt����x�v�wxv�{vstv�~vx�s��x�x�}�yvsx��x�v�w|��f�U RSTU[ToT fmUbaUb\T WfbVW]U\ be]TX[Uf[UVb UT]WTgUbUSTUicgTUbf][UW[\SW]TVTUUU ibV TV jU� cgTUhfSUT]WfoT]mU]bhUW[\SW]TVTUibV TV [UiWVUnfoTUSTUXb[UUU fV WiWe]TU\ be]TX[�tt�x�t�~vx�|{���rv�vx{sv~t{tr�rt��x�~~}s|x�rx�x|{v~t�~x{sv||}sv�xtrxt|x���zx�y|vs�vwUUU fVUWTW[UbaUSTU\ bgcifbVU[m[TXUhSTTUSf[Ub UWU]bhTU\ T[[c TUf[UTq\TfTViTgUUU TjnjUW[\SW]TVTU\ Tif\fWfbVUXWmUeTUfVffW]]mUbe[ToTgUfVUSTUaWif]ffT[jUUZ[UUU STUT[TobfUgT\]TT[dUSTU]bhT[U\bfVUbaU\ Tif\fWfbVUhf]]UXboTUbUSTUb\UbaUUU STUcefVnUWVgUSTVUgbhVUSTUhT]]UcVf]UW[\SW]TVTU\ Tif\fWfbVUiWVUW][bUbiic UUU fVUSTUT[Tobf�ttt�x�v|{trvxr�t|�x�v�wxsv{�sr|x��x��s��rt��xw���uvxw}vxr�x�|{���rv�vx{sv~t{tr�rt��UUU fVUSTUT[TobfUWTUWT�fo�U�b]fgUW[\SW]TVTUf[UVbXW]]mUTXboTgUXTiSWVfiW]]mUabXUSTUhT]]UWVgUaWif]ffT[�o�U � TiSWVfiW]UTXboW]Uf[UVTiT[[WmUeTiWc[TUW[\SW]TVT[UWTU\bb ]mU[b]ce]TUfVUUU [b]oTV[�of�U�WnTU[iW]TUW[\SW]TVTU\ Tif\fWfbVUbiic [UhSTVUSTUW[\SW]TVfiUicgTUbf]Uf[UUU ibVWiTgUemUWifgU�UebSUaT[SU��c[U\cX\TgUfVbUSTUhT]]�Ub U[\TVUWifgUUU �WaTUibVWiUhfSUSTUab XWfbV�jUURSf[UTaaTiUf[U\W fic]W]mU[ToTTUhSTVUSTUUU WifgUibVWfV[UaT fiUiWfbV[U��T����UTjnjUabXUTWifbVUbaUSTUWifgUhfSUc[jUUU RSTUfX\WiUbaUSf[UbVUWifgU[T]TifbVUhf]]UeTUgf[ic[[TgU]WTj������H ¡¢£¤PJH¡¥H¦¡£§MOF¡QH M§M©PHªH«M¤OP£FMRSTU\ T[TViTUbaUeWiTfW]Uib]bVfT[UWVgdUXb TUfX\b WV]mUSTUT[fgcT[USTmU\ bgciTdUiWVU]TWgUbUWUVcXeTUbaUfV WiWe]TUb\T WfbVW]U\ be]TX[jUUZ]SbcnSUSTmUWTUXb[U
������������ ������������������������� ����������� �� ��������������� ���������������� !"# $%!#&!!'#()*'+#,'#$--#."/!0#)(#1!--0#$'+#($2,-,.,!0#3#($2,-,.,!0#$4!#)/!'!+#.)#. !#$.5)0/ !4!#()4#,'0/!2.,)'#$'+#&). #1!--0#$'+#($2,-,.,!0#$4!#$.#)22$0,)'$--"#!6/)0!+#������� ������������7�������7����� ��������8�����9�� ��������������������7��������$'+#!0.$&-,0 5!'.#)(#. !#&$2.!4,$-#2)-)'":##� !#()--)1,';#0*55$4,0!0#0)5!#)(#. !#5)0.#1,+!-"#()*'+#/4)&-!50<=,># ?@ABCDE@FGHFEICGJGACFFKF@DGDCIL@LBM#!.2#1,--#/ "0,2$--"#/-*;#. !#()45$.,)'### /)4!0#,(#,'N!2.!+#,'.)#. !#1!--:#� !#0)-*.,)'#.)#/4!%!'.#. ,0#,0#.)#!'0*4!#. $.### . !#&$2.!4,$-#2)-)',!0#+)#').#&!2)5!#!0.$&-,0 !+#,'#. !#1!--0#$'+#($2,-,.,!0:#�� O��P����������� ����������������� ���7��������������� � �����P���������#�� ������� ��Q�������������������������P�������������������R���������ST�U��R����## V)1!%!4W#,.#)'-"#4!X*,4!0#)'!#&$2.!4,$-#0/)4!#.)#&!#5,00!+#1 ,2 #2$'#. !'#�� ���������������Y=,,>#ZD[LG?@ABCDE@#$4!#$!4)&,2#&$2.!4,$#1 ,2 #2$'#!0.$&-,0 #. !50!-%!0#*'+!4### $'#,4)'#+!/)0,.#. $.#. !"#()45#. !50!-%!0:##� !,4#;4)1. #/4)2!00!0#. !'#4!0*-.### ,'#. !#!0.$&-,0 5!'.#)(#$'#$'$!4)&,2#&$2.!4,$-#2)-)'"#!:;:#)(#0*-/ $.!### 4!+*2,';#&$2.!4,$W#-!$+,';#.)#$#/,..,';#."/!#2)44)0,)':##� !"#$4!#."/,2$--"#()*'+#�� ������������������� ����������������� �������Q���������9����������\��#�� �������Q�������T�������P����������� �����������������������P����� ������### ,4)'#+!/)0,.:##]!;*-$4#5!2 $',2$-#4!5)%$-W#!:;:#&"#/,;;,';#. !#-,'!W#,0#. !#5)0.### !((!2.,%!#.4!$.5!'.:=,,,>#HKF _@BCG CaKAELbG?@ABCDE@#$4!#$'$!4)&,2#&$2.!4,$#1 ,2 #$4!#2$/$&-!#)(### /4)+*2,';#-$4;!#X*$'.,.,!0#)(# "+4);!'#0*-/ ,+!#=Vcd>#1 !'#&). #. !### 0*-/ $.!#$',)'#=*0*$--"#(4)5#0!$#1$.!4>#$'+#$#2$4&)'#0)*42!#=24*+!#),->#$4!### /4!0!'.#,'#. !#$&0!'2!#)(#)6";!':##Vcd#/4!0!'.0#$#5$N)4#0$(!."#$'+#2)44)0,)'0### ,00*!#0,'2!<=$># e6/)0*4!#.)#Vcd#$.#$#2)'2!'.4$.,)'#)(#$#(!1#.!'0#)(#//5#,0#($.$-#.)# *5$'#-,(!=&># f !--#$'+#($2,-,."#2)5/)'!'.0#5$+!#(4)5#2)'%!'.,)'$-#0.!!-#$--)"0#1,--#($,-#�� ������9�� ���������� ��������������������� ��������cd:g'N!2.,)'#1$.!40#2)'.$,',';#!%!'#-)1#-!%!-0#)(#. !#(!44,2#2$.,)'#=h!iii>#1,--#/4!2,/,.$.!#. !#!6.4!5!-"#,'0)-*&-!#0)-,+#(!44,2#0*-/ $.!#=h!cdj>#1 !'#!6/)0!+#.)#Vcd:##h!44,2#2$.,)'0#$4!#')45$--"#()*'+#,'#,'N!2.,)'#1$.!4#=0,'2!#,.#,0#/4)2!00!+#,'#0.!!-#%!00!-0#$'+#)6";!'# $0#&!!'#/4!0!'.#,'#. !#0"0.!5#!.2>:##� !#0)-,+#h!cdj#1,--#/-*;#. !#()45$.,)':klmlklGH[KDACMG[nGo[DI@BE[LGp@I@bCGqGrDCMMKDCG CaKABE[L� !#4!0!4%),4#/4!00*4!#)(#. !#%$0.#5$N)4,."#)(# "+4)2$4&)'#4!0!4%),40#1,--#&!#4!+*2!+#+*4,';#,.0#/4)+*2,';#-,(!.,5!:##� ,0#2$'#.4,;;!4#$#'*5&!4#)(#)/!4$.,)'$-#/4)&-!50#1 ,2 #4!0*-.#,'#()45$.,)'#+$5$;!:=,># � !#/4!2,/,.$.,)'#)(#,')4;$',2#$'+#$0/ $-.!',2#02$-!0#1$0#+,02*00!+#!$4-,!4:=,,>#� !#4!+*2.,)'#,'#4!0!4%),4#/4!00*4!#,'24!$0!0#. !#()45$.,)'#!((!2.,%!#0.4!00:#### � !#,5/$2.#)(#. ,0#,'24!$0!+#0.4!00#2$'#-!$+#.)#()45$.,)'#($,-*4!#4!0*-.,';#,'<
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EFGH IFJKHLMNKOPQRNJESGH TNULFPQRNJHNVHQWXHMXYXMZNRMHLMNKOPRJ[HRJQXMZF\]HHWRYHMXYXMZNRMHKMRZXHUXPWFJRYUH__ a_bcdcefgh_adfc_i_j hh_hcgk_il_gd_dfmcgack_mcflncmo_lp_iqc_lh_lmrdghhoHHH RJHL\FPXHRJHQWXHMXYXMZNRMsHSOQHRQHUFtHF\YNHSMRJ[HXuQMFHNLXMFQRNJF\HPNYQYHKOXHQNHHH YOMVFPXHYOSYRKXJPXHFJKHKFUF[XHQNHQWXHvX\\YHQWXUYX\ZXYHwEPFYRJ[xQOSRJ[HSXRJ[HHH YyOXXzXKHRJQNHFJHNZF\HYWFLXHMFQWXMHQWFJHSXRJ[HMNOJK{HNMHXZXJELFMQRJ[HKOXHQNHHH YWXFMRJ[G|]EPGH }FO\QHPMXFQRNJHNMHMXFPQRZFQRNJHwR]X]HEURJRGHXFMQWyOF~XYG]HHTMXFQRNJHNVHXuQMFH__ pg�hia_�go_hcgk_il_��mlnck_d�lj_jqhc_pg�hi_mcgfingild_�go_mca�hi_dHHH LMXZRNOY\tHYXF\RJ[HVFO\QYHSXPNURJ[H�\XF~t]�EKGH �XKOPXKHVNMUFQRNJHLXMUXFSR\RQt]HH WRYHRYHJNMUF\\tHURJNM{HSOQHPFQFYQMNLWRPHHH LNMXHPN\\FLYXHQNHzXMNHLXMUXFSR\RQtHWFYHSXXJHNSYXMZXKHVNMHYNUXHPWF\~YHvWXJHFHHH PXMQFRJHMNP~HXVVXPQRZXHYQMXYYHRYHXuPXXKXK]�`�_�cacmnlm_��hi�qgac_�lj_cppcfia_iqgi_�go_bc_pghacho_giimb�ick_il_plm�gild___ kg�grc�__�_a�pefcdi_mck�fild_d_�mcaa�mc_al_iqgi_iqc_mcacmnlm_��k_�gaacaHHH QWXHSOSS\XHNMHKXvHLNRJQHMXYO\QYHRJHQWXHLMXYXJPXHNVHFJHXuQMFHLWFYXHRJHQWXH__ mcacmnlm_jqfq_j hh_mck�fc_iqc_�mchginc�_�cm�cgb hio_il_lh_lm_rga_�lj����������������������������������� ������¡�����¢ chh_ai��hgild_a_�mlbgbho_iqc_�lai_pmc£�cdi_l�cmgild_jqcd_��ka_gmc_����ck_dil_g_jchh�__¤qc_emai_iqmcc_pgfilma_kaf�aack_bchlj_gmc_ldho_g��hfgbhc_il_iqc_����dr_lp_qrq_mcgfinio_��ka_a�fq_ga_gfk�__¥ljcncm¦_iqc_mc�gddr_cppcfia_fgd_bc_cdfl�dicmck_j iq_gdo_��k_iqgi_a_����ck_dil_iqc_jchh¦_�dhcaa_�ml�cm_iql�rqi_a_[RZXJHQNHRQYHYX\XPQRNJ]ERGH �XFPQRNJHLMNKOPQYH[XJXMFQXKHStHQWXHMXFPQRNJHSXQvXXJHQWXHRJ§XPQXKHFPRKHFJKHHH QWXHVNMUFQRNJHMNP~HUFtHLMXPRLRQFQX{HPFOYRJ[HFHMXKOPXKHLXMUXFSR\RQtHEVNMUFQRNJHKFUF[XGERRGH^WXHFPRKHUFtHvXF~XJHEKXPNJYN\RKFQXGHQWXHMNP~{HStHFQQFP~RJ[HQWXHRJQXM[MFRJHHH PXUXJQHYNHQWFQHEJNMUF\\tHQXULNMFMtGHYFJKHLMNKOPQRNJHRYHNSYXMZXKHvWXJHQWXHHH vX\\HRYHMXQOMJXKHQNHLMNKOPQRNJHRJPNJYRYQXJQ�`�_¤qc_gblnc_kcfldalhkgild_�mlfcaa_�go_rcdcmgic_edca©_jqfq_fgd_� rmgic___ gdk_bhlfª_�lmc_iqmlgia�__¤qa_�mlfcaa_fgd_lff�m_plm_liqcm_��ka_�acc_acfild___ «�¬�«�_ld_��k®mlfª_fl��gib hio�ERZGH¯PRKHRYHNVQXJHRJPNULFQRS\XHvRQWHPMOKXHNR\H\XFKRJ[HQNHVNMUFQRNJHNVHFHYN\RKHHH �Y\OK[X�HvWRPWHPFJHS\NP~HLNMXYHEP]V]HYXPQRNJHKXYPMRSXKHQWXHLMXPRLRQFQRNJHNVHHH FYLWF\QXJXYHStHFPRKGHNMHFHZRYPNOYHFPRKHxHNR\HXUO\YRNJHVNMUFQRNJ]EZGH ¯HVOMQWXMHVNMUHNVHFPRKxPMOKXHNR\HRJPNULFQRSR\RQtHRYHQWXHVNMUFQRNJHNVHFHWR[W\tHHH ZRYPNOYHvFQXMxNR\HXUO\YRNJ]HH WRYHXUO\YRNJHYQFSR\RQtHRYHNVQXJHRJPMXFYXKHStHQWXHH_ edc_�gmifhca_kcafmbck_d_�`�_gblnc
������������� ��������������� ���������������������������������� ������ ��������� !�����"#$%��&' "�� ((��#(�����'����!)*"���(#� ���+��,�#- ( �.�����'���� ��/0 ��1�'.1��!#�-�2�+'#"��3�� �'��� 2!��#" 24����1�!��#" 24� �"�5����#���������������6�� ����6��7�������������� �����������67��������������� 28�!� �2��9��'�����#�,�2��50 1����������:���������������������������:��66� ��������7���������6������ ����6��7�����;��<���=����7:����6>?��;;���������@7���������7����� ����6��7������������:�A6���7���������7?��������6�������������������6���������� �:������������������6;&'��)�.����#�� 1 24�#((��'�"��+��-(�,"� "�+��+���"�(�!� �2��9��'��#! 1�������'���50 1$��;;����6��7���������7�����6�������������������6�������7�;BCDCECFGHIJKLMFHNFOHJPQRSHTFUQPQVLFWFX QRLJFYTZLKRSHT<�������7�6���� ��������[�������������6�����6�����:�667������������������:���������������:������������������������=6 :��>��������9��,��'�� 28�!���"�����'��+��10!��"%��&'��!#+#! �.��9�#2� 28�!� �2���((� 2�#�4 ��2�\�(1��6���������]��������6�����������:���7����� ��;����6������ ������7�6���:��?����2�9������.�",#((�!�2�#, 2#� �2�(���("��9��'�� 28�!��1�50 1_��'#���#+ 1�+(044 24��9��'��+��9��#� �2"�!#2��!!0�%��&' "� "�� �,+( \�1� 2�\40���aa��'����"����#(� 28�!� �2� ���6�������:7�:������7���������������������:������������67����6����7������������[��������7�;����6��6�6�6�� �����b���� ���7������67����� ����:�7�6������������6�����������;�����6��6������ 67�:��6���6������ ���� �6������7����?�������c��������6�7���� $�d (��#� �2��9��' "�"0�9#!���#����� �'�#�ef�2�, 2#(�\(����1�!��#"�"��' "��#����9��� ��:������ $�g"���9�#2����2�\2���\(���_�hf_�1�!��#"�"��' "����2�90��'����������� ijklmnopqqorrqsrtqq snr tr oonr pq psnrulvlw xyz{l|}lij~yz{z�ly�lol��lvlw xyz{l|}lij~yz{z�ly�lrl��lvl�{����z�lw xyz{l��y{zxyz�|lvl��y{zxyz�l��{�x�zlw xyz{��~��zl���z�yz�l��jy{z��z{���u�z{xkz�l�z{�zx�j~jy�l�zx��{z�l��{zl��� ������ ¡¡ ¢£¤¥¦§� ©�� ª�«ª ��«�§ª« ª© �¤¬ �¬®� ª�©¬ ¤ª�� °�¤§�ª±
������������������� ����������������������� ���������������� �� ������ !��"#"$"%& '"#� ()*�+�,-.�/ %�".0 1%/ %2�,34%��"5 !$"6"+"02378 �&& 9��#.�/ %0$�7:;)6�%<&"$ -�%$�<�"02:=0+�%>&� -.�/ %�".0+ %06 ?%06 @.0��.&A�"& %06 B%+ 9�.$�++"02AC8 %��� ;%06&"02C�"& %06 B%+ -"�&6 �*��%�".0+DEFGHGIJKLMMKNOGKLPEHGKHEMQRGSKETKFLNGIKUVWGXNGYKLIGKSRLMMKXERZLIGYKNEKNOESGKIG[QUIGYKUVKLXNQLMKUVWGXNUEVKZIEWGXNS\K]VKZILXNUXG_UK aOGKUVWGXNUEVKFLNGIKUSKETNGVKRQXOKXEMYGIKNOLVKNOGKIGSGIHEUI\KKaOUSKXEEMUVbKETKNOGKKK UVWGXNUEVKUVNGIHLMKIGYQXGSKNOGKIGSGIHEUIKSNIGSSKUVKNOGKVGLIKFGMMPEIGKLIGLKLVYKKK LMMEFSKTILXNQIGKXIGLNUEVKLNKLKRQXOKMEFGIKZIGSSQIGKNOLVKFEQMYKVEIRLMMcKPGKKK GdZGXNGYK_NOGIRLMKTILXNQIUVb \KKaOUSKTILXNQIGKGSSGVNULMMcKUVXIGLSGSKNOGKTEIRLNUEVKKK SQITLXGKLIGLKEZGVKNEKNOGKFGMMJKUVXIGLSUVbKNOGKFGMMKUVWGXNUHUNcKLVYKIGYQXUVbKUNSKKK SGVSUNUHUNcKNEKTEIRLNUEVKYLRLbGeffghfiijkljimnojkpfjqhkrmksltjrfuvsrfwftkijgfxftgjltj_LK aOGK[QLMUNcKETKNOGKLHLUMLPMGKFLNGIKSEQIXG_P K y OGNOGIKNOGIRLMKTILXNQIUVbKEXXQISKLVY_XK aOGK[QLMUNcKETKNOGKTEIRLNUEVzkjiplvhgj{fjtlkfgjkpmkjkpfjqtfjqhkrmksltjfuvsxwftkjpmij{fftjrfwl|fgj}rlwjkpfjst~f�ksltj�mkfrjkrfmkwftkjkrmstjl}jwlikjelrkpj�fmjqfhgij{f�mvifjl}jkpfj�llgj�mkfrjuvmhsknjmtgjNOGKVLNQILMKNGVYGVXcKTEIKNOGKNLIbGNKTEIRLNUEVSKNEKTEIRKNOGIRLMKTILXNQIGS\�ss�j�pfj}lrwmksltjswxmsrwftkjgvfjkljilhsgijxrfiftkjstjkpfj}ffgj�mkfrj�q�vrfj���jl}KKK IGLMKUVWGXNUEVKFGMMSKUSKSQZZMGRGVNGYKPcKRLVcKGdNILKSEQIXGSKETKYLRLbUVbKKK ZLINUXMGSKG\b\KXEIIESUEVKZIEYQXNSJKEUMJKPLXNGIULMKIGSUYQGSKLVYKENOGIKSEMUYSK_SQXOKKK LSKSXLMGKZIGXUZUNLNGSKZUX�GYKQZKYQIUVbKUNSKZLSSLbGKNOIEQbOKNOGKTLXUMUNUGS\KKaOGKKK MGHGMKETKXEIIESUEVKZIEYQXNSKLVYKPLXNGIULMKXEVNLRUVLNUEVKLIGKIGYQXGYKPcKNOGKKK UVWGXNUEVKETKXEIIESUEVKUVOUPUNEIJKYGEdcbGVLNUEVKETKNOGKUVWGXNUEVKFLNGIKLVYKUNSKKK XEVNUVQEQSKNIGLNRGVNKFUNOKPLXNGIUXUYG\K�QPSNUNQNUVbKZIEYQXGYKFLNGIKTEIKTIGSOKFLNGIKPIUVbSKLKVGFKILVbGKETKZIEPMGRSKYQGKNEKNOGKZIGSGVXGKETK_MEFKMGHGMSKETKEUMKLVYKSRLMMKSEMUYKZLINUXMGS\KKaOGK_IGMLNUHGKFLNGIKuvmhsknj�mtj{fj~vg�fgj}rlwjhstfj�jstjq�vrfj���jj�vrkpfroj{fst�jplkojkpfrwmhj}rm�kvrst�jUSKQVMU�GMcKNEKEXXQI\���������������������������� ������������������aOGKZIGHUEQSKSGXNUEVSKETKNOUSKXOLZNGIKOLHGKYUSXQSSGYKNOGKbGVGILMKXEVXGZNKETKTEIRLNUEVKYLRLbGKLVYKOLSKbUHGVKSGHGILMKGdLRZMGSKFOGIGKUNKEXXQIS\KK VGKETKNOGKLIGLSKFOGIGKTEIRLNUEVKYLRLbGKTIG[QGVNMcKEXXQISKYQIUVbKNOGKMUTGKETKLKZIEYQXUVbKFGMMKUSKYQIUVbKFEI�EHGIKEZGILNUEVS\KKaOUSKUSKYQGKNEKURZLUIRGVNKETKNOGKZIEYQXUVbKTEIRLNUEVKPcKSEMUYKZLINUXMGS\¡tnjilhsgijxrfiftkjstjkpfj�lr¢l|frj£vsgoj�shhj{fjst~f�kfgjstkljkpfjrfifr|lsrjgvrst�j�fhhj¢shhst�jlxfrmkslti�jj�psijsijipl�tjstjq�vrfj�¤j�pfrfjkpfj�fhhjxrlgv�ks|sknj�mijIGYQXGYKPcKVGLIMcKLKNOUIYKFOGVKNOGKFGMMKFLSK�UMMGYKFUNOKYIUMMUVbKRQY\
������������� ������� ��� �� �� �� ��������������� �������� �������������������� ������!����"�!���#$�� ���%&� ���%&�' �((�)�((���' �!*�+�� �����,-./.�/0123�4567281./�95:�51/0�;.�<.=.657.3�;:�362112=<�8.9.=7>��/86542=<�85/2=<>�?@AA@BCDEFGHDDI@CJKEDLMDNOPQNDOPQDQPKRPSEKDTJ@UDAPNNEND@[email protected]:�80=/75=7�Y06�Z�35:/�[-.=�5�458\.6�[5/�;.2=<�9211.3]�� =8.�7-.�[06\0X.6�P_EKWF@PBD_KPCKENNEUDFPD?@AA@BCDPBDWDNODFOWFDQWNDAPUCEUDWGKPNNDFOED_EKVPKWF@PBNaDFOEDAPNNENDUEGKEWNEUDbcDMdeDVPKDFOEDBEfFDFQPDUWcNgDUKP__@BCDS@KFJWAAcDFPDhEKPDWVFEKDUWcDij]��k7�2/�X.6:�12\.1:�7-57�7-.�062<2=511:�04.=�4.6Y065720=/�-53�;.809.�80941.7.1:�;108\.3�;:�7-.�.=3�0Y�7-.�04.65720=]��l�6.3m8.3�[.11�4603m8720=�85=�;.�.n4.87.3�[-.=�7-.�[.11�[5/�6.7m6=.3�70�4603m8720=]oDNEGPBUDEfW?_AEDPVDUW?WCEDCEBEKWF@PBD@ND@AAJNFKWFEUD@BDCJKEDLpDqDFOED@B@F@WADQPKRPSEKDTJ@UDAPNNDKWFEDUEGKEWNENDbcDrMeDQOEBDFOEDGE?EBFD@NDUK@AAEUDWBUDNFP_NDWAFPCEFOEKDQOEBD7-.�85/2=<�2/�/8654.3�=.56�7-.�4.6Y065720=/]s PKRPSEKDTJ@UDtJWA@FcDCJ@UEA@BENDWBUDKEGP??EBUEUDP_EKWF@PBWADFEGOB@tJEND6./4.872X.1:�56.�5X5215;1.�70�92=292/.�7-2/�86.57.3�Y0695720=�3595<.]� u v w x y z { | } u~ uu uv uw ux uy uz u{ u| u}����������������������������������������������������������� ���������w �� �������������� ��������� �����������������������������|�vy ¡¢£¤¥¦§©¤ª«£¬¡ª®¦¡°ª¤±¦° ª²°¦³ª¤ µ¡ª®¦«µ µ¢¥¦¬¤¥µ¥«¦¶±¦²¥··¦¡··¡®¢ ¡¢£¤¥¦§¹ºª¤ ª»¥¤¦¡°ª¤±¦° ª²°¦²¯¥®¦³ª¤ µ¡ª®¦«µ µ¢¥¦ª¬¬£¤¤¥«
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EEEEEEEEEEEEEEEEEEEEE FGHIIJKLMLNOPQGRSLJKRTHNUJJVLRGJWLGXYGROHYNTWLGXYGROHYNTPOHIIJZSLNJ[\] ^ _ ` a b cFRdTFRHIdJeYTTLTJfggITh iHUJjk blmnmop qrsqturovwxyzo{x|wy}~o�xyzuwy�u�o}qo� y�y�y�uovqr�|}yq�o�|�|�u�����E����������E���E��E����������E��E���EE������������E��E���������E������E����E����E����E���������E������������ �¡¢£¤�¥¦£�¦§�©¦�ª¤££«�¬ ¡£¤£©�¤�®£¦���«��ª�¡¢£�¤£¦¥¡�® �°¥�©�§¦¦�¤±¡£��¦�¡§§EEE �����²E���E�E³���E ����������Eµ �������E¶³´µ ·E ����E��E��������¹E��º��E��E�� ª§¤»�±®�£¼¡£¤®±�¬ ¡£¤�½±¾£�§®�¡¢£�ª§¤»±¡�§®�ª±½£�§¤�¿£¤ª§¤±¡�§®¦À��Á�¦§¥£EEE ³´µ E��E���E���������E������E����E ��²E������E��������E��E�������E���������E�� ¿±¤¡�½£¦��¤£»§Â±�§ª�± �§ª�¡¢£¦£�¤£Ã¥�¤£�±�¢Ä©¤§½¢§¤�½�±½�©�°¥¦¢�¦�®½£�¡¢£EEE �����E���������E������E������EE ���E�E���¹���E����E��E��������¹E����E ���EEE ����E��E³´µ ·�EEÅ��E��E���E���������¹E�E �������E�����E��E��Æ�����E ���EEE �����E������E��������E����������E�����Ç �®�»�¦£�½§®¡±»�®±¡�§®�§ª�¡¢£�¬ ¡£¤£©�¤�®£À��È¿£¤±¡�§®±�¡£½¢®�Ã¥£¦EEE ��E�������E����E�������ÉEE ¶�·EE ´����E����E¶����E�����·²E������E���E���E�������E����E���E���Ê�²EEEE E �����²E�����²E��������E����E�����E��E���EE ¶�·E µ �������E���E��E����E����E ���E�������E������EE ¶�·E Ë��E���������E���Ê�EÌE�����������¹E���E������E���ÊE���E���������E�� � §¿£¤±¡�§®¦�Í¢£¤£�¡¢£�°¥�©�§¦¦�¤±¡£¦�¡§�¡¢£�ª§¤»±¡�§®��¦�¢�¢À��Φ£�§ª�¡¢£�¦±»£EEE E �Æ�������E���E���������E����������E�������E��E�������������E����E��E��� � £¤Ä�©�ª¬½¥¡�¡§�¤£»§Â£������ÏЧ�©¦�ª¤££Ñ�¤�®£��¦�®§¤»± Ä�©£¬®£©�±¦�½§®¡±�®�® �£¦¦�¡¢±®�ÒÓÓ�¿¿»�Ô§¡±E�� Ð¥¦¿£®©£©�Ч�©¦À��ÕÓÖ�§ª�¡¢£¦£�¦§�©¦�¦¢§¥©�£�¦»± £¤�¡¢±®�§®£�¡£®¡¢�§ªEEE ���E�������E����E��º��EE×ØÙÚÛÜÝÞßàáÛâáãÜÛÝäØåæáÛçÝØèÜéæØêÜåÝëìÚåÜåÝáíÝíáÛîìæØáéÝèìîìÙÜ
��� ����������� ����� ����������������������� �� ��������������������������������� ������������ � ����� ����� ����������������������������� ��� ������ ���� !"#�$%&'(#"�$!")*+&#,�)-!)�-!.#�/%)�0*,,%&.#0�%"�-'0"!)#0�$"%$#"&'�1%2)#/��� 03#�)%�*/!0#43!)#�(*5*/678��9�+-#(*+!&�:"#!;#"�,-%3&0�!&,%�:#�!00#0�)%�)-#��� .*,+%3,�,%&3)*%/�)%�#/,3"#�)-!)�)-#�.*,+%,*)'�0#6"!0#,�$"%$#"&'<�*8#8�,%�)-!)��� ���������������������������������=��������� �� ��������������������� >#&&:%"#�!)�)-#�#/0�%2�)-#�%$#"!)*%/8���� ? ����� �������������������� ��������������� ����������� ���������������� 1%.#":!&!/+#7�:#)>##/�)-#�"#,#".%*"�$"#,,3"#�!/0�)-#�-'0"%,)!)*+�-#!0�%2�)-#��� ���������������@��������ABB�� ���� �������������������������������������� � ������ ������������������ ���� � �CDEDFG HIJHKLIFMLNOPQRSLTFUHFV QPQWQTLFXHIWYUQHPFZYWY[L\)�*,�+&#!"�2"%(�)-#�!:%.#�)-!)�(%,)�%2�)-#�2%"(!)*%/�0!(!6#�%++3",�>-#/�)-#�>#&&�*,�;*&��$"*%"�)%�)-#�,)!")�%2�)-#�>%";%.#"8��]"%$#"�>%";%.#"�$&!//*/6�+!/�#/,3"#�)-!)�(!/'�)'$#,�%2�>%";%.#",�+!/�:#�+!""*#0�%3)�>*)-%3)�-!.*/6�)%�;*&&�)-#�>#&&8�� %(#������� �������� ��������� ����_ �������a���������� a������������� � ���������������������������� bcdefghigjklmhnefolkodfmnphkkfqdkkflonhrfsjkktdlrheufvkjhrfhemnfmtdfcdivnilmhnewxwdfgkdlefyfenezrl{luheuphkkfchkkw|f}dwhuefvnifgn{ckdmdfid{nolkflvmdifqnipnodifvhehwtdr~ tdefqnipheuflsnodfmtdfclgpdiwmncfvkjhrwfsdheufhe�dgmdrfhemnfmtdfvni{lmhnefs�fcklgheufckjufhefmjsheufmlhkfchcd�inmdgmfcdivnilmhnewfvin{rl{ludfqhmtfknwmghigjklmhnef{lmdihlknifwlerfckjuhefglwheu �huf�|��1*7� � #+-!/*+!&&'�*,%&!)#�)-#�:%))%(�%2�)-#�)3:*/6�>*)-�!�$&36�*2�)-#�>%";%.#"��� +%/+#"/,�#43*$(#/)�$&!+#0�!:%.#�)-#�$!+;#"81**7��,#�>*"#&*/#�>%";%.#"�(#)-%0,�%/�!��&*.#��>#&&�#868�� 1!7� ��3($�:!*&#",��)%�$&!+#�(!)#"*!&�>-*+-�>*&&�*,%&!)#�1$"%)#+)7�)-#���� � $#"2%"!)*%/,8���"!.#&�*,�%2)#/�3,#08���-*,�-!,�)%�:#�"#(%.#0�!2)#"�)-#���� � ����������� � �� ���� �������������������������������������������� 1:7� � !/'�>#&&�,#".*+*/6�%$#"!)*%/,�,3+-�!,�$#"2%"!)*/6<�$&!+#(#/)�%2�:"*06#���� � $&36,�#)+8�+!/�:#�+!""*#0�%3)�:'�>*"#&*/#81***7�� *"#&*/#�%$#"!)*%/,�!"#�/%)�)-#�%/&'�)'$#�%2�%$#"!)*%/�>-*+-�+!/�:#�+!""*#0�%3)��� %/�!�$"%03+)*%/�>#&&���+%*��)3:*/6�!/0�,/3::*/6�3/*),�+!/�:%)-�%$#"!)#�>*)-��� !�-*6-�>#&&-#!0�$"#,,3"#8 ������������������������ ¡��¢�������£ ¤����¢ £ ���¢������¥��¦�¡���
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EFGHIJKLMFNOIPKFQRISTULIVWXIRKTSYZXIVUIVWXIMXTSUTYVFUNRIVWTUKOWIVK[FNOIVWYVIWYRII\ ] _\_abcdde\fgh\c\_ij] h\gf\k cha\lhcba\bmg\nhg]djaoEYHI pWXIVK[FNOIFRIZUUqXQI[rIVWXIZUqQIPKFQIYNQIZUNVTYZVRsIItSIYIVK[FNOIYNZWUTIWYRI[XXN\\\ ia e\ uu\bv \lgjnd bg_\eg a\wxy\lg_bc_\c_k\jgz _{\a cda|\bv a\\\ lg_bhclbg_\gf\bv \bi]_{\jck\i_acb\bv \egm_vgd\ncl} h\gh\nidd\bv \bi]_{\cnchbE[HI pWXIMKLMXQIPKFQI~FqqIZqXYNIVWXIFNRFQXIUSIVWXIUqQIVK[FNOIYNQIVWXITXRKqVFNOIQX[TFR\\\ m dd\] \_�lbe\bg\bv \n hfghcbg_au�a \gf\c\abh_{\gf\lg de\bi]_{\�gh\a_i]]_{\i_b\bi]_{�\mv h\bv \_ae \gf\bv \bi]_{\vca\] _\ldc_ e\] fgh\ia \m dd\czge\]gbv\nhg]djau\\yv \dcbbh\nhg]dj\jck\] \heil e\]k\c\nhgn h\nijn_{\alv eid\f\bv \m dd\c_e\nhaaih\lg_e bg_a\ch\cnnhgnhcbu\\�\edib|\_v ] be\cle\a\nijn e\_bg\bv \bi]_{\c_e\eandcl e\bg\YIRYSXIQFRVYNZXIY[UGXIVWXIMXTSUTYVFUNRIERUIVWYVINUNXIUSIVWXIPKFQIZULXRIFNVUIZUNVYZVIm bv\bv \n hfghcbg_a�u\\�b\a\bv _\lhlidcbe\]cl}\bg\bv \aihfcl \u{u\]k\gn __{\c\ade_{\ae \eggh\_abcdde\_\bv \bi]_{\c]gz \bv \ncl} h\c_e\h z ha \lhlidcb_{\]k\nijn_{\_bg\bv \c__idiau�\aj dch\nhgl eih\�m bvgib\bv \_ l aabk\bg\gn _\ade_{\ae \eggha�\jck\] \ia e\bg\ldc_\bv \_ae \gf\bv \lg de\bi]_{�a_i]]_{\i_b\g_l \b\vca\] _\_a hbe\_bg\bv \m ddu�����������������������������������������������������IMTUQKZVFUNIqUOI~WFZWILXYRKTXRIVWXIPU~ITYVXIYVIYNrIMUFNVIFNIVWXI~XqqIFRIUNXIj bvge\bg\e bhj _ \f\fghjcbg_\ecjc{ \vca\bc} _\ndcl u\\� {ih\��\\alv jcblcddk\TXqYVXRIVWXIPU~ITYVXIYVIGYTFUKRIQXMVWRIVUIVWXI�KYqFVrIUSIVWXIEMUVXNVFYqqrIWrQTUZYT[UNInhgeil_{�\ac_e|\ca\j caihe\]k\bv \{cjjc\hck\dg{u\\�b\a\ldch\bvcb\bv \]gbbgj\ac_e\vca\�ngb_bcddk�\] _\jnche\a_l \b\a\_gb\lg_bh]ib_{\bg\bv \nhgeilbg_u\\\\\\\\\\\\\\\\\ ¡¢£ ¡¢£¤¢¥ ¡¢£¦§ ¥©¢ª«¦§ ¥©¢ª«¦§ ¥©¢ª«¬§§¥®§ ¯ «§°±²«³¥ ¡¢£¥®§ « µ¶¥· ¹º§yv \mghe\»¼½¾¿½ÀÁÂÂÃ\a\ia e\v h\a_l \bv h\ch\gbv h\ngaa]d\hcag_a|\cnchb\fhgj\fghjcbg_\ecjc{ |\bvcb\lgide\cdag\cllgi_b\fgh\bva\dcl}\gf\m dd\nhgeilbg_|\u{u\bv \]gbbgj\ac_e\jck\vcz \cdhcek\] _\e ndbe\c_e\bv \h a hzgh\nhaaih \a\bgg\dgm\VUIYqqU~IVWXIÄUNXIVUIPU~sÅÆÇÈÉÊËÌÍÎÉÏÐÈÑÒÆÏÓËÔÏÇËÕÊÖÇÖË×ØÆÓÓÊÉËÏÉËÙÆÐÊÏÚËÆÐÊÓÒÆÛÊ×ËÜÏÓÊ×ËÊÝÞÆßÆÒÆÓÇËØ××ÆßÔÊËËàÏÉáâÒÆÏÓËÐâáâÇÊ
������������ ������������������������������� !"� #��##����$���%&���'����%�������'�'������������������������%������&��(����)������������ !"*�'�'����������������+$����$���%&���'����%�����������������������������&�#�������,���������������&$���������$����-��(������'�������������������#�������(���������������%��������������)��.�(�/��0������������'�'����������������+$��� ���������������&�������������1%��������������234567896:;<=5>6?@A6B8C?DE8@68F6DGH6F89I?DE8@6A?I?JH6I?K6LH6EAH@DEMHA6N�O���EDP69HI8Q?B6EP6?DDHIRDHAS662GEP6EAH@DEMC?DE8@6R98CHPP68FDH@6E@Q8BQHPT ��� U�������������(�&&������������������'$��#����������������������/�����%&��������� '���������������7EE>6:?99KE@J68VD6PRHCEMC6B?L89?D89K6DHPDE@JW6PVCG6?P6?69HPH9Q8E96C89H6XVPGE@JW6D8�66 AHDH9IE@H6EF6DGH6EAH@DEMHA68RH9?DE8@P6AEA6E@AHHA6BH?A6D86C89H6A?I?JH6F896DGH�66 R?9DECVB?96C8ILE@?DE8@68F6DGH6XVEAP6E@6YVHPDE8@6?@A6DGH69HPH9Q8E96F89I?DE8@ �����U��������������������������(������#��������������&$����&�������1%����%���66 ?P6DGH6PC?@@E@J6HBHCD98@6IEC98PC8RH6D86C8@M9I6DGH6A?I?JH6DKRH6?@A6DGH��� �������&��������������������/�&�#������������(��������/����)�������$�#����+&���$#����'�'��������������������%����������������#�������������(�$�'������������������%�������/�����+�����������%��������$���������������)�Z��(�&&��&���LH6PHH@6DG?D6H?CG6P8V9CH68F6F89I?DE8@6A?I?JH6G?P6EDP68[@W6PRHCEMC6CV9HS6
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EFGHIIJKLMNGHGOPPQRSTUVOVTWVXVWYZZX[\V] V_]ZZ]RV\PQWV] V][ZXT]RVSXZXaQbVSTXaR]WT_V_cYQWVXRSV[QZQSTXT]RVQ_dRTeYQWV[]ZVfORVTR[]SY_T]RV]VdQVgXWTWV] VhQccV_]ZPcQT]RWijklmnojkpqiroqsrtpuvpwxuiryzqs{rxskjkpq|}r~pmujxi�rp�r�xpu�xr�kq�ry�mnio}r����|{������������� ���������������� ��������������������������������������������������� ¡¢£¢¤¥¡¦§¤©£ ª¤¢«¬®°¬«®±²³³¬´±µ¶·¬¹®º±µ¬¶°»¼½ µ½¾¿ÀÀ²±À¿³³¿ ¬»±Á¼±°¬¾°«±¿±À³¿ ±®«±¶À¬ ± ®«¾¬µÂ±Ã°¿»¾¹¬µ±¶±µ ¶³³±¶ ¿¾«±¿À´¶¬° ·³¼¹¿³±¿°±¶³¹¿½¿³±®«Ä¬¹®¿«±Á¬³¿ ±½¼»°¶¬À¿° ¶®¿«±Ã¿®«Å±±Æ«µ¾³¶¬»±°®µ¬°µ±¿°±¾Á®«·Å¿®³± ¬³³²±¾µ¾¶³³¼±¿«³¼±À¿° µ±«¬¶°± ¾»±³®«¬®«±¶°¹®¹±°¬·®¿«µÂ± ¶¼±¶³µ¿±À¿° ±®«±»°®³³®«·¯¾» ®«µ¾³¶¬»±°®µ¬°µ¬¿µ±°¿±¼¶³¹ÇÈÉÊËËÊÇÌÍÎÏÐÊÑÇÏÎÇÈÉÊÒÊËÎÓÔÌÈÒÊÑÇÕÒÊÉÓÏÕÊÖ ÑÈË×ÌØÓÏÌÇ ÙÒÎÒÇ ÚËÓÛÌÜÝÇ ÐÏÊÒÈÇ ÌÓËÊÒÊÉÛÚÞÓÒÕÈÌÇÖÓÛÇ ÉÏÊÕÕÈÏÇÑÊÒÈÌÇ ÖÎßÈÖÈÒÉÝÌØÎÏÓ×ÊÚÇÏÈ×ÍÚÉÊÎÒÌÇÊÒÇÑËÎàÇÏÓÉÈáÇßÓÏÊÓÐËÈØÏÎ×ÍÚÉÊÎÒÇ ÏÓÉÈÇ ÉÈÌÉÌÝÇ ÑÊÒÈÌÇÊÒÇ ØÏÎ×áÈÖÍËÌÊÎÒÌÇÏÓÏÈÇÐÍÉÇØÎÌÌÊÐËÈ ÚËÓÛÇÚÎÒÉÏÎËÇÑÎÏÇØÏÈßÈÒÉÊÎÒáÇÏÈÉÓÏ×È×ÇÓÚÊ×ÑÎÏÇÏÈÖÎßÓËÝÇÖÓÛÇÒÈÈ×ÇÉÎÇËÊÖÊÉÇÏÓÉÈÇÊÒÈâÉÏÈÖÈÇÚÓÌÈ×ÒÓÇÈÉÊËËÊÇÌÈÖÊÉãäåæçèéêëìçãíîíìãäæéêëìçïðîëëãñæ æäãìíîíãçîðíãòóãôôãôçìëêéõçöãòäãêóîëîíéçîææãðæîíî÷ãçíåçìøêðïãçîðçñêíãòçæêëîðîíéçåòóòîðãçíéöã êìîôîùãçñîíøçúûüýúþçîÿçôêäêïãçîæçæøêëëåñúéôòê�ëîìçþòêìí�òãçîÿçôêäêïãçîæçôããöãòíøêðç���èæíêåòøíçãòåöçëëêäæçèëëãñçæêïçéëëê�æ�æ�ìåëóçòãíêñ �ðíòãêíãôçñêíãòèçëåñçöòãææ�òãçÿåòäêíîåðõçîÿ÷ãòéçëåñçöòãææ�òãç��çêóå�íç���çöæîýÿíèçöåòãíøòåêíçæîùãçôåãæðíçäêííãò äêíòî�çíòãêíçñîíøçêëìåøåëçåòçæ�òÿêìãçíãðæîåðëåñãòîðïç æ�òÿêìíêðíõçîð�ãìíçïêæçîðçïêæòãæãò÷åîòçíåçôîæíêðìãçåÿç��çÿíëëê�æ��ìãòçôî�ëÿçôêåëçòååö éç ÿåòäêíîåðç ôãöãðôãðíõç äêéôãìòãêæãçöòåô�ìíîåðçòêíãçîðçæã÷ãòãçìêæãæçåòìê�æãçëåðïçìëãêð�öçíîäã�çç åæíçìåääåðçîðÿåòäêíîåðæç ñîíøç æäêëëç öåòãæç êðôäîìòåöåòå�æçìëêéæ �òã÷ãðíîåðç�çê÷åîôçåòçä îðîäîùãçñêíãòìåðíêìíçêðôçëåñãòçæ�òÿêìãçíãðæîåðçåÿçñêíãò��ãäå÷êëç�çêëìåøåëæçêðôçæåäãçæ�òÿêìíêðíæñåëçéòã÷çíêçëëãñçåíðîçíìã�ðîçðêìçèðî�æçøïîøæÿòãöçðîçëëîÿ òêíãæõç òãöãòÿåòêíîðïç äêéç æøåñç æøêòöîðìòãêæãæ ìëãêðå�íçåÿçòãöãòÿåòêíãíìã�ðîçåíçãëóîææåöä���������������������� �� ���!��������� �����"#��$�%!���!�������"�������"���������!&������%�&�����'#������%����&�#�!��!��#���#��!�#$�������(�"����!�����#��&!�����"��#���# �#)#��#�������!���������# ��!��������!����$��#'�������������������#��%*+,-*./01/0-2345676+/4,8-9:6; 31749-743/3<1=3/47*,3-103/-0/1-,/=4,,3>;11:/7*+/810+-9-10-05/.4?1:4/847409>?:4@*409,A/ 4081*094:4+/ -0/ 1;40/ <1,4817;,49-103/60+/<1:-B1096,/=4,,3 CDE/1:/CDEFCG/68-+/-0/769:-H/68-+/I1.31,J409/=63</?1,,1=/=-9</68-+/-0/=4,,3=-9</ 1-,/ .634/7*+>/ ?167/1:/ I499-058,460*;3/6:4/1?940/*34?*,76A/6,31/.4/4081*094:4+/?1,,1=-05/68-+/1:?:68/39-7*,69-103/1:/=1:21J4:/?,*-+3/=<4:46/+-:9A/=694:/=63/*34+/1:/9<4/=694:/=63<6*,4+/1:/391:4+/-0/6/+-:9A/9602 ?-,94:/9:4697409/?,*-+3/K/*34/8,460/96023860/:4+*84/-0I489-10/-0/?,11+/;:1I4893/K.682?,1=/3<1=3/;6:9-8,43/60+/1-,/86::AK1J4:/:49*:0 .4994:/8,460-05/1?/=694:A,,6:*960/0-/34331,/+*7/4,1<=/45:6,50-55*,;/4:*986:? ?:689*:4+/ ?1:769-103>/ -094:7-99409;:1+*89-10/69/,1=/60+/71+4:694/:6943>-0?:4@*409/:481J4:A/1?/=<1,4/7*+/60+7*+/?-043L/M174/47*,3-103N/43;48-6,,A6?94:/68-+/9:46974093>/76A/6,31/188*:/-?<1,4/-3/;11:,A/8,4604+/+*:-05/+:-,,-05 68-+/*34?*,/-?/+67654/-3/3<6,,1=>/?:68/-?+44;>/91/;:4J409N/-7;:1J4/31,-+3/:481J4:A
�� ����������������� ������������������������������ ����� �������������� ��� ��� ����� ������������� �������� ��� ��� ��������� ���������������������������������������� ������ ����� ����� ���������� ��������������������������������� ����������� ���������������������������������������������� �������������������� ���� ������������ ������������������������������������������������������������� ����� ������������� ��������������������������������������� ������������ �������� �������������������� ������������ ��������������� �� �������������������� ��������������������������������������������������� !"���������� ��������#� �������������� ������������������ ������������������ �������������������� ����������� �� �����������������$���������������� �������������� ���������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ab\cHdCFMGCPBOHDFHVBM@@ABMHBM@FBJEFJPGHJGHFAIAKDB@S@PCFHFPHTDBNHOD@@HCPAGNHDFHVBM@@HNBPV@SEPKPBHBMNNJ@THIBPWGHFPHIKDEUSHMQMGHWTJFMPBHZBDRHVP@@JIKMS @EBDVJGZ[HEAFFJGZHCPBHOMETDGJEDKHBMOPQDKSTPFHPJKHA@MCAKHJCHNMVP@JFHJ@H H_aaHCFHCBPO@ABCDEMSH@PKQMGFH@PDU@HPGHNMMVMBHNMVP@JF@S@POMHJGTJIJFPB@HDQDJKDIKMHCPBHVJVMKJGM@HDGNCMWHVBPIKMOHWMKK@SH@POMHWMKK@HBMeAJBMEPGFJGAPA@HNPWGTPKMHFBMDFOMGFHFTBPAZTODEDBPGJH@FBJGZSH@VMEJDKHIDEFMBJDHDBMHDK@PA@MNHFPHVBMQMGFHVBMEJVJFDFJPG
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�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aUOOQNbHJREOLHNEPKXVJEPFGHOYOGHEWFX\WHVFJTcdefghfihjklmnkjkghomjkhpeflqdrshotrhikkhtgkljktifuvhiqfuhfchoedwfxdkhyefdgexhwkiwitjkhjeu tddmzhw{khzkddhwmhldktuhexhmuhfwihmzusheik|rdkukhxjkcdei{ht{ktghmchtlfghz{kuhw{kxjmydkohfihqumzuhwmhmllejkltcjeihufh}kdtlihwmu~hiwfimxkghgjt{hmwhwcmiikufdhzmdchufhufcctjt� �������������������������������������������� ����� �������� ��� ������� ���������������������������������������� �������������������������������������������������������������������������������������� ���������������������������� ��� ������������������¡��¢�������������������� ���������������������������������������������������������������������������¢���������������������������������������������������������������������������������������������� � ����� ��������� ��������������������� ����£¤ ¥¦§ ©ª« ¬® ®©§©°±®²³«§³ µ® ©³¶§®·ª®µµ®§¹§º§»¼½¾¿ÀÁ §Â°§°§·®¬ µ® ·® §²³«§ µ°· ³ §¹ à ¬³ § µ³¶À©°¬Ä® ©§© «§©§Åì® ®§°µ§°¬ ® ®¦§©°·®© µ± ©§ µ®§¬©°§©®± ¬§±°µ§Â³©ª®µ©³° §°±©ª« ¬³§· ©©¦§©®§ ³Æ©¬Â¶®§ ³³Ä®§°µÅì® ®§²³«§ ³Â§³§µ®©®µÄ°³µ©§²³«§·°µ®« §½Ç¼È §©ª« ¬§°§ªµ®Ä® §©¬Â¶®©¶³µÂ§©°³µ°±µ®ª§°³§Ã¬²°¬©§Ãµ®Ä§©²°¬±µ ªµ°Â ³° §°±§°³¬¦§ © ¬¬Ã§ ©©° ³®Â§²³««®§ªµ®©® ®§°±§ § µ§Â®ª°©³§ ® µ§ª Ãɱµ®Ê ® ¬Ã§«³¶«¬Ã§ ©ª« ¬³¦§· ç ° ³©°·®§² ®µ§« §³©§³®Â§ ª§ ©§Âµ°ª¬®©§°µËª° Ì®©Í§³ §«®§«³¶«§Ä³© °©³Ã§· ©© ©°¬Ä® §©° ̧ ©§ ®®Â®Â¦§®©§©°¬Ä® §²³«© ·ª¬®§°±§µ§Á®±°µ®§Î°Á¦§«® §°±® §«®¬ª©´ ¬³· § µÁ° ®§© ¬® · ç±°µ·§ § 窵®©© µ®§Âµ°ªÉ§®³«®µ§³«®§±°µ· ³° §°µ§³§«®§ Á ¬µ©¦§· ç±°µ·Ä®µÃ§±©§ § ¨ §©« µª¬Ã§¬³·³§ªµ°Â ³°®©ª® ³¬¬Ã§ §¶µÄ®¬§ª ̧³®µ± ®©§°µ§ ® µª®µ±°µ³° ©§³§²®¬¬©§²³«§«³¶«§Âµ²§Â°²¯¨ µ°©©§«®§ª®µ±°µ³° ©¦§· ç° ´ µ§·°µ®±µ®Ê ® ¬Ã§³§® µ¬³®µ§©¶®©§³§©°·®§±³®¬Â©²«® § ªµ®©© µ®§ µ°ª§ ³©§ ·°µ®§©®Ä®µ®ÇÏ© ¬¬Ã§· °µª«° ©É§ ° Ƶ鬬³®§±°µ· ÐÑÒ§°§µ®·°Ä®§ §³«³Á³°µ§°§ªµ®Ä® ¦³«³Á³°µ§· ç Á®§ ©Ê ®®Ó®Â§ ³°§ «®±°µ· ³° § ±°µ§ ¬° ¶®µ§¬³Ä®Â§ªµ°® ³° ¦©°·®§ÐÑҧΰÁ©§ · ç µ³¶¶®µ§´ ¬ ³·´ µÁ° ®§© ¬®§³§µµ®§ ©®©¦§³«³Á³§ ³Â°µ§µ® §²³«§ÔÕÖק³±§«³©§³©§§ªµ°Á¬®·®µ®«²§ª°µÂ§®µ©©®µª§§©·µ°±§Ã¬¬©®¬©§®±¬©§·³¬´ ³  ®Â§ Áç µÁ ¬® ®¦§·°µ®§ ±µ®Ê ® ²«®µ®§«³¶«§© ¬±®§² ®µ©§ ° §«³¶«´ ¬³·§² ®µ©§ §³§ÑØ Ù ±¬°°Â©¦§© ¬®§³©°§ ³Â§©°¬Á¬®¦§· çÁ®§±° §° §° ©³Â®°±§ª ·ª©§ §§³ Ì®©§ §¶ ©§®Åª ¬©³°ª°µ©§ 夀 ¬Ä®©§³§Â°² «°¬®¦§ µÃ©¬©§ µ®« µ ®µ³©³§±°µ§«³©§© ¬® «®·³ ¬§ ° Ä®µ®µ§°µ§Â³©©°¬Ä®µ§±°¬¬°²®ÂÁç¨ ³Â¦§¹Â°§ °§ ° § ° Ä®µ®µ§°µÂ³©©°¬Ä®µ§²³«§ ³ÂÀ¦§ ³Â§³©§ °§ ©®± ¬§Á󩮬±¦§ ³«³Á³°µ©§ ª¬ ®Â§ Áç ©Ê ®®Ó®µ® ·® ©§µ®§©®± ¬§°§ªµ®Ä®
������������� ��� ����� ����� ��� ���������������� �� �������������� � �������������������� ���������� ������������� ������� ���� ���������������������������� ������� ������ ��� ���� � ����������� ����� ����������� ������ ������� ��� �� �� � ���� ������������������������� ��������� ��� �� ������� ���� �� �� ��������� �������� ������� ��� ����������������� ������ ��������� ������� �� �!"#$�� ��������������%�� &'()��*��$�� ����������� ���+���� �� � �� �������������� � �����,���������� ������������� �- ��� ������������� � �� �����������������������������. ������ ������,���� ��� ���, ������� ��������� � ��� �� ���� �� � ��������������������� ���������������������� ���� ��������� ���� � � � �������� ��� ���� ������������ ����������������� ������������� �������� � � �������� ���������������� �������� ���������� ������������ � ���� ������� ���������������������� ������� ���� � �� ���������������� ��������� ���� ���������� �������%� ���� �� ������� � ������������ ���������, �� �����������/�� �����������, �� �������������/�� ����������� ��� ����� ������ �� �������������012����������� ������������ ��� ����.� � ��������������$�-� / ����������� ����%� ������� ����� ��� ��� ������ ���� � ���� �� �� ������ �� ���������� ��� ���-������� $������������������������������� ����� ��� ���������� ������������������� ���������������� ��� ���� ��� ������ ��� �� ����������3��� �������������� ����� ���� ��� �� �� ������������������������� ����� �� ��� ��� � ��������� ���� ������ .������ �� ����� ������� ��������� ���� � �� �� � �������� � ���� ��������������� � -��� ��� �� ������������������ �� �� �� ����������� ������ ���������� � .� �� �������� ��������� ������� ������ ��������� ����-������������ ������� ������������������� ����� ���� ����� �����������������%�� �� ������������������� � � �������������� ��������������������������� � .������������������ �������������������� ����� �� ������� ������ ��� ��������� ������� ������ �� ������������ ���������� �� ����� �� ������� ��� �� ������� ����� �������������� ������ � �������� �� ������������ � �� �������� ��� ������ �� ������ ��� � ���� ���� ������� ���������� � ����%�������� �������������������������� ���� ����������������� ������������ �� ����� ������ �%����� ����
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EFGHFIJKLGMEFHKNOPQQPRFSTGPHSHUJQTPFKSPMJHOSL MJRPOHVFPHFPUFIGMJOGIIGTFPTFPTWJOHPGTFXJKKUFXYJMJFUNMQSOJGMFEMGRNOJRFXSHJMFPUFPTWJOHJRZFPQFHYJFOGKGT[PUF JUHSLKPUYJRF PTF HYJFXSHJMF YSTRKPT\U[UHJIZFPHFOSTFGOONMFXPHYFPTWJOHPGTFGQFST[XSHJMUVFXYJTFHNLPT\FPUFENKKJRZFLMGXTFHGLKSO]FUKPI[FISUUJUFGMFSTF^ _ FGRGMFSMJPTRPOSHPGTUVF LSOHJMPSF IS[F OSNUJFUKGXMJRNOHPGTUFPTFHYJFPTWJOHPaPH[FGQFSFXYGKJQPJKRVFOGIEKJHJFMJIGaSKFGQFHYJFLSOHJMPSFPUMSMJK[FEGUUPLKJFbFHMJSHIJTHFPUFNUNSKK[FGTFSMJIJRPSKF LSUPUVF NTHMJSHJRF XSHJMF PTHMJSHIJTHUVFRMPKKPT\FQKNPRFGMFPTWJOHPGTFXSHJMOSTFUGNMFMJUJMaGPMUFXPHYFcdFLSOHJMPS HMJSHFXPHYFLKJSOYFQGKKGXJRFL[F efFgRGFTGHSKKGXFOGTHSOHFGQFLKJSOYFSTRF efhVFIGMJijklmnlomipokiqolimqkrmsomlotouukprvkwioplkixyomipokiqoliumkpomtjwnpxlomzxn{xzoklzmsktiopxtxzomuw|}ui|smiuoim~|mzwx|smkmlnmo}kqkzmukm~|mu�njuulnxikpn�po~mpnn~ tkllnimsomt|pozm�xijmktxzmnpmqotjklxtkwtjkl}oumxlmijom�owwvm~pnswoqumu|tjmkuutpooln|iumn�mjrzpk|wxtm�pkti|pom�nsu�mjx}jxl�otixnlm ~pouu|pou�m uklzm ~pnz|tixnl�zn�ljnwom utkwom ntt|ppolto�m |luikswooq|wuxnlu�mklzmzn�ljnwom~kpk��xlmklzku~jkwiolomzo~nuxiumkpomtnqqnlvm~pnswoqu�xijmkm�owwmijkimtkllnimsomspn�olmzn�lmnpoyolm ~|q~ozm xlinm ujn|wzmkw�krum sok~~pnktjozmsrmpo~op�npkixl}mijom�oww ������������������������������� ���������������������������������������������������� ��� ����� ����� ����� ���������������������������������������������������������� ����� ������������ ������������������������������������������������������������������������������������������������������������������������������������������������������������������������ ��������������� ���������������������������������������������� ¡¢¢£¤ ¥¦ �§£ ¦ © ¦¡�¦ª� «©¤�«¦ ª¡¬�¦ £ª«��¬ª©¤¬£®¦£¦¢ ¬¦�°© ±�®¬²¦¦³§�¤ ¦ ¬§¡ µ¡©µ�¦ª®µ¦£®¦ ª�¢®�¦£ª¦¬©¶®µ²¦¦·¡�¦¡¡ª�¬¥¦ £ ¦ ¤¡©���¦ ®¤¢©«®µ¦ �� �ª�¤£ªª£�£®¹¦ ±¡¢©®¤¬£®®µ¦ �ª£ª¡¬®µµ©®�¹¦ �¦¢¡ �¦¡®«¦ �¡ª¦£ ¦¬©¶©¢¡ª�¦ª£±«ª¢¢®µ¦£ª¦¢¬¦�¥�¬�±²¦¦º £�¬¦¤£±±£®¤¡©���¦¡ª�¦�¡ª¬§¦�§¬¦¢£¡«�¦¤¡©��«¦¶¥�©¶�«�®¤�¦£ ¦ ª£«©¤®µ¦ £ª±¡¬£®�¦ ¬§¢©«¦¡®«¦�¡®«¦ ¬§«ª¡ ¡¢�¹¦¡¤¬�¦¡©¢¬�¹¡®«¦£ª±¡¬£®¦±£ �±�®¬�¦®�¡ª¦�¡¢¬¦»£®�� ¼¦¤¡©��«¦¶¥¦�¡ª¬§¦�§¬¦£ª¤��¹¦©��¦§�¡ �ª �¦£ª¦±©¢¬ ¢�¦�¬ª®µ�²¦¦½®�ª�¹¦¤�±�®¬¹¡®«¦ ¡¬¤§��¦©��«¦£ª¦ª� ¡ª®¡§¤¦®�««©��¡�¢ µ��¦®¦¾¿À¹¦Á ¿À¹¦ ª���©ª�¹£ª¦¤§�±¤¡¢¦¡®¡¢¥��¦£ ¦ ¡¬�ª ª� ¡ªÂ¦¤£®�«�ª¦¤£ªª£�£®¦¤£®¬ª£¢¦ ª£µª¡±
����������������� ������������������������� !"#$�%&'()*��(�+�,��)$�-�.&�/�*/)*�.��.���$!"#�-)��.&�%&'(�."&#�-�(�0�1��������2 33 442 33 4422 5657 4444 24 4 4 4568 78 8978 89: 9: 99 : : 9; ;<=; ;<=;><= ??; ?;;;; ; <=<=; ? @; @;<=; A ?;;;;B CDEF GHIB CDEF GHIB CJK LM JK LMB B NB B BOOPQQOQQ O ORQSTURSVQSTU WVP WWS57575657B NB WWSWWS WOQVR?? ?;;;; ? ������������������������������������XYZ[\\Y] _ \[a[bYcYdYeffY_gYh i_dajcYZjakYdYZ[\\Yl d]jcmYhYnfYd]i[lYoi��p�qrs�%.tu����� �**,&'��-"�(�.�'��v)�*$�q�qwxyu��z.�"$�{*�##�-�.&�{�'%&'�.��.���{�+�|&#�� ".����.�'&)0��.),"#0�0)#� ��'��.���{�'%&'�."&#�/��'0�$���}����-�{.��&%�{�#�.'�."&#�&%�~r�"#u�z��".�"$�-�/"-�-�.&�/�$��.����&*�u����-'"**"#0�()-�)$�-�.&�-'"**�.��� �**�/&#.'&*$�.�����.�#.�&%�{�'(��,"*".+�"({�"'(�#.��#-�.���-�{.��&%�-�(�0�-�'�0"&#���.�#-$�,�+&#-�.��� �**,&'�u������%&**& "#0�()-�.+{�$��'���}�"*�,*�1� � ���� � � � �� � �� � �� � ����������������������������������� ���� � ��� � ���� � �������������� �¡��������� � � �¢£� � �¢¤� � �¢�� � �¢�¥ �"/��()-�$+$.�(�-&�+&)�'�/&((�#-�%'&(��� �**�{'&-)/."}".+�{&"#.�&%�}"� �%&'��#�&{�#�&*���#-����/�$�-��&*��/&({*�."&#¦§�� �©���ª{{*+"#0��� !"#$�%&'()*���������������������������������������� 2 33 442 33 4422 5657 4444 24 4 4 4568 78 8978 89: 9: 99 : : 9; ;<=; ;<=;><= ??; ?;;;; ; <=<=; ? @; @;<=; A ?;;;;B CDEF GHIB CDEF GHIB CJK LM JK LMB B NB B BOOPQQOQQ O ORQSTURSVQSTU WVP WWS57575657B NB WWSWWS WOQVR?? ?;;;; ?' �p�q�qwxY«Yo¬®¬ Y°Yf±²¬³haYdcbY ljcmYµ bYXY did_[a[il¶Yj±[±Y·-Y°Y®ffY_gYdcbY�'-Y°Y®±Y¹Yf±²¬³Y°Y®±º²¬³Y� » ¼»B CJK LM JK LMB ½¾¾¿¾¾ ¿ ¿ÀÁÂþÁÂÃÄ¿ Å ÆÇ ÈÈÅ È�#-�.���É'&-)/."}".+�Ê�."&�Ëp�Ì"({�"'�-wÌ&'"0"#�*t� "**�,��12 33 442 33 4422 5657 4444 24 4 4 4568 78 8978 89: 9: 99 : : 9; ;<=; ;<=;><= ??; ?;;;; ; <=<=; ? @; @;<=; A ?;;;;B CDEF GHIB CDEF GHIB CJK LM JK LMB B NB B BOOPQQOQQ O ORQSTURSVQSTU WVP WWS575756B NWWSWWS WOQVR?? ?
������������������� ����������������������� ���������������� �� ����� !" �#$#%#&'!(#$�!)*+�,�-./ 0 &�#/1!2&0 &3�-45&� #6!"%#7#,#13489 �''!:� $/ 0 &1%�8;<*7 &='#%!. &%�= #13;>1,�&?'�!./ 0 &�#/1,!&17!@&17!A/1� /'B�#'!&17!C&,!: /%�,,#13BD9 &�� !<&17'#13D�#'!&17!C&,!.#�'7!�+� &�#/1,EFEG HHHH IH H H HEF J KJ KK J J KLLLL L MNMNL O PL PLMNL Q OLLLL R SR R RTUVW XYZ XXUEGEGEFEGR SR XXUXXU X[TY\OO OLLLL O]_ abcc_defcghbae]i]ejkdh_l]mnbmeopq gres_]d]e]tuitub_ vs_rguugw]dxetyu_z{||}~�z�z�~�z{��z��~�|���������~�� {������z���������������������������������������������������� �� ������� ���� ���� � ¡¢��� ��¡�£� ��¤� ��¥�¦� ¤��� ����§~�|�}� �|������©�ª~��{«��¬�������©�����¬��®� ~����©°�� {������� ��z�~�«�||�©������©���}©©�©°�}©����©������«�|������©®�z�©«�����«~�}��z�����|�©�|{|��}|}°��±²³ �µ¶®�c_hba]dxes_fcghbae]i]ejkdh_leg·pq gres_bdht tx_hitub_� {�������¹��z�~�«�||�©�������}�«}z��������«�|������©��z���}©©����� {��¹��~��{«�z�}���°�����|�}�~�����~|}���©��}|}°��º�©����«���}z�������°��z��z»�©�}©���������z���~��{«��¬����~}�������}��������~����©°�|{�z���¼��¬�~®�����z�}�z���������©©�z����½�����}©©���¾_crgcte]dxh_¾es¿Ào]dÁ]moo]dxc_te_cestdes_rgc te]gdht tx_¿Âp]dÁÃs_da_������~��~}���©z�����Ä��}Ä������Ä��}zz�������~|}���©��}|}°����Å���~��~}���©���}©�w]esmbrÆa]_de¾_crgcte]dxh_dm]ejtdhmb]etyu_¾stm]dxmsgbuhy_tyu_egh_u]i_ctd{©�}|}°���«�|������©�±Ç�ÈÇ��³�¶�
�������������� ����������������������������� �����!����" ��!��!� ������#�� ����$���" ��!�%"�"&���"!������ "�!�'��!���'�()*+,--./0+1*+2/3+456.7890+,*0+:);;<=*++>753?@8A5/+BC.728A5/90+D5-?E.+F+:G8H+I3A8A5/=*++>?6-A9H.3+6J+BKLM*+MNOP+QR;<Q;SFRFQR)*F*+I@5/5EA3.90+T *0+UA--0+,*+V*+2/3+IH-AWRI@5/5EA3.9+L*0+:);;<=*++>.875-.?E+>753?@8A5/+NJ98.E9*++>?6-A9H.3+6J+>7./8A@.+U2--*++MNOP+QR)<RXYZXZ<R[*<*+I@5/5EA3.90+T *+2/3+P5-8.0+\*0+:);Z;=*++4.9.7]5A7+N8AE?-28A5/*+>?6-A9H.3+6J+
55
C O N T E N T S
1. INTRODUCTION2. WELL INFLOW AND STIMULATION3. AN INTRODUCTION TO WELL STIMULATION
ECONOMICS4. CANDIDATE SELECTION
4.1. Treatment Timing4.2. Further Treatment Selection Criteria &
"The Stimulation Cycle"5. SELECTION OF (CHEMICAL) TREATMENT
TYPE6. POTENTIAL FORMATION DAMAGE
CAUSED BY MATRIX STIMULATION FLUIDS7. MATRIX STIMULATION FLUID SELECTION8. TYPICAL ACID FORMULATIONS USED FOR
MATRIX ACIDISING8.1. Hydrochloric Acid (HCl)8.2. Organic Acids8.3. Mud Acid8.4. Selection of Acid Composition8.5. Selection of Treatment Volume8.6. Selection of Injection Rate8.7. Selection of Additives8.8. Selection of Treatment Type8.9. Selection of Diversion Technique
9. MATRIX STIMULATION FIELD CAMPAIGNS10. STIMULATION OF CARBONATE FORMATIONS
10.1. Acid Composition Selection10.2. Treatment Types for Carbonate Rock
Acidising10.2.1. Matrix Treatments10.2.2. Acid Wash or Acid Soak Type Treatments
1.1 ACIDISING OF SPECIAL WELL TYPES11.1. Gravel Packed Wells11.2. Horizontal Wells11.3. Naturally Fractured Formations
12. ALTERNATIVE ACID FORMULATION13. APPENDIX A14. FURTHER READING
Acidising and Other Matrix Treatments
Revised 26/07/05
1
2
LEARNING OBJECTIVES:
Having worked through this chapter the student will be able to:
• Explain the importance of matrix stimulation in Production Engineering.
• Identify and contrast the application areas of the various types of matrix stimulationtechniques.
• Generically list the primary chemical reactions in sandstone and carbonate acidising.
• Explain the potential negative impacts of “matrix stimulation” and identify migrationstrategies.
• Select acid formulation on the basis of source of formation damage and rockcomposition.
• Identify and discuss the role of acid additives.
• Discuss placement and diversion techniques.
• Design a matrix acidising treatment (acid volume and injection rate).
55
Department of Petroleum Engineering, Heriot-Watt University 3
Acidising and Other Matrix Treatments
1. INTRODUCTION
Chapter 4 described the many sources of formation damage which can lead to areduction in the permeability of the near wellbore area and creation of an extra,positive skin, as measured by a well test. Matrix (stimulation) treatments are acommon form of well intervention aimed at removing this formation damage andrestoring the well to its natural, undamaged inflow performance. An alternativestimulation technique - propped hydraulic fracturing - will be covered in a laterchapter. This latter well treatment can bypass this damage and/or increase theeffective wellbore radius. Either of these stimulation treatments may be carried outimmediately after drilling the well is completed or at any time in the well’s producinglifetime when they can be economically justified.
The next section will discuss the parameters which control well inflow and describehow this can be improved.
2. WELL INFLOW AND STIMULATION
The well known, steady state, radial-flow equation describes the well inflow:
Qo = Ko h (Pe - Pwf)
141.2µBo{In(re/rw)+S}(1)
Where Qo
= well oil production rate (STB / day)K
o= formation permeability to oil (mD)
h = reservoir thickness (ft)P
e= reservoir pressure (psi)
Pwf
= wellbore flowing bottom hole pressure (psi)µ = oil viscosity (cp)B
o= oil formation volume factor (reservoir bbl / STB)
re
= well drainage radius (ft)r
w= well radius (ft)
S = skin (dimensionless)
An increased well inflow (Q), or well stimulation, can be achieved by:
(i) increasing the factor {k.h} or(ii) decreasing one of the factors S or r
e/r
w or µ.
The frequently used stimulation techniques are catalogued in table1. This table alsoidentifies the parameter which this type of well treatment targets in order to increasethe well’s fluid inflow.
N.B. Mechanical and combined mechanical/chemical methods will be discussed inthe Hydraulic Fracturing chapter.
1
4
Mechanical Methods
• Propped Hydraulic Fracturing Increase rw• Explosive Fracturing Increase rw and k
• Underreaming Increase rw• Re - and Additional Perforating Increase h
Chemical Methods
• Matrix Acidising Decrease S
• Tubing Acid Washes Improve Well Outflow by
• Other Chemical Matrix Treatments (Surfacants, Solvents, Mutual Solvents Etc.) Removing Tubing Deposits
Increase k
Biological Methods
• Microbial Stimulation Mechanism Uncertain
Combined Mechanical / Chemical Methods
• Acid - Fracturing Including Propped Acid - Fracturing Increase re• Closed Fracture Acidising Increase re
Thermal Methods
• Steam Soak Decrease µ• Heat / Gas Generation From Injected Chemicals Decrease µ and Improve Well
Outflow by Increasing GOR
• Electrical Heating Decrease µ
Technique Objective
The choice of which technique is the most appropriate for a particular well can be madewith the help of table 2.
Treatment Type Skin Permeability
Propped Hydraulic Fracture Low Low
Propped Hydraulic Fracture High Low
Matrix* High High
Treatment Probably Not Required Low High
Clastic Reservoirs
* Frac and Pack Bypasses Formation Damage in Medium Permeability
Formations
Treatment Type
Propped Hydraulic Fracture As For Clastic Reservoirs
Acid Matrix Treatments Widen Natural Fractures
Acid Etch (Short) Channels
(Wormholes)
Carbonate Reservoirs
This chapter, entitled “Acidising and other Matrix Treatments”, discusses the chemicalmethods of well stimulation. The common factor among these treatments is that theyare carried out under matrix conditions i.e. the injected fluids flow radially away fromthe wellbore since the treatment fluid is injected into the well at rates and pressuresbelow that required for creation of a hydraulic fracture. This chapter mainly concentrateson the injection of acid (“Acidising”), the most frequently employed of the chemicaltreatments.
Table 1
Available stimulation
techniques
Table 2
Stimulation treatment
selection
55
Department of Petroleum Engineering, Heriot-Watt University 5
Acidising and Other Matrix Treatments
As noted in Table 1, tubing washes are a related technique which share a similartechnology, in terms of fluid and additive selection, with matrix treatments. They areused to improve the well outflow by removing deposits which have formed in thetubing i.e. by increasing the (effective) tubing radius. The treatment fluid is normallyretained with in the tubing during the such a tubing wash treatment and is NOTinjected into the formation. This procedure avoids impairing the formation with anyundissolved particles which have become dispersed in the treatment fluid. Tubingwashes will not be discussed any further - but the requirements e.g. for corrosioninhibition (discussed in section 5.8.7) if an acid wash is selected (e.g. because thedeposit in the tubing is acid soluble) is very similar to matrix acidising.
Matrix treatments, and acidising in particular, aim to remove the excess flowingpressure drop (∆P
d) created by the presence of a volume rock which has suffered
formation damage (i.e. has a lower than original permeability) in the near wellborearea (figure 1).
Zone With Formation Damage /
Reduced Permeability
rw
kd
rd
undamage
Actual shape of
formation damage
Idealised shape of
formation damage
(modeled below)
Formation (k)
Ideal Pressure Profile(Undamaged)
Actual Pressure Profile(Damaged) (kd < k)∆Pd
∆Pd
P2
P3
Prkd k
rw rd re= Extra pressure drop due to Formation Damage
ReservoirDamaged
ZoneWellboreCentreline
The removal of this formation damage will restore the “natural” well productivity.The Hawkins formula:
Figure 1
The effect of a near
wellbore damaged zone on
the well inflow pressure
profile
1
6
S = -1 In KKd
rd
rw(2)
where S = skinK = formation (original or undamaged) permeabilityK
d= (near wellbore) damaged formation permeability
rd
= radius of formation damager
w= wellbore radius
is a convenient tool for analysing the influence of varying levels and depths offormation damage.
Matrix stimulation treatments increase well productivity by pumping a speciallyformulated treatment fluid (frequently, but not always, an acid) which is designed toremove (normally dissolve) the formation damage. However, the keys to successfultreatments are:
(1) the identification of a suitable candidate well which is capable of a greaterhydrocarbon production rate,
(2) the selection of the optimum type of treatment fluid for the removal of theformation damage
(3) the design of the operational aspects of the treatment
such that the required economic criteria are successfully achieved.
The next section of this chapter will discuss some simple, economic concepts whichhelp identify whether a stimulation treatment could be economically viable.
3. AN INTRODUCTION TO WELL STIMULATION ECONOMICS
Well stimulation is only justified when the net (discounted) monetary benefit of theresulting extra oil or gas production is greater than the cost of the stimulationtreatment. Previous field experience from the stimulation of similar wells is often agood guide when predicting the expected gain from the well stimulation treatment. Asecond simple, but approximate, method to estimate the potential benefits from astimulation treatment is to use the fact that the production wells typically show aconstant, long term, annual percentage decline in net hydrocarbon production. Thisdecline can be expressed in the form of a straight line when the logarithm of the nethydrocarbon production is plotted against time (figure 2).
55
Department of Petroleum Engineering, Heriot-Watt University 7
Acidising and Other Matrix Treatments
Actual oil production rate of unstimulated well
Extrapolated long term production decline rate ofunstimulated well
Extra net oil produced by stimulation
Expected productionrate gain due tostimulation treatment
Increased decline rate of stimulated well
The stimulated well production may revert to original decline rate or even lower
Net OilProductionRate('logarithmic' scale)
Time
N. B. A straight line with a steeper slope in figure 2 corresponds to a greater, annual,net hydrocarbon production decline rate. Note the extra oil reserves created byproducing the well above the economic limit for a longer period of time.
The expected, net hydrocarbon production gain from the stimulation treatment needsto be estimated as the first step in carrying out the economic evaluation. This can beestimated if:
1. the well’s skin value is known (equation 1) or
2. by use of the Hawkins formula (equation 2) if the extent and depth of formationdamage are known or can be guessed with reasonable confidence.
Field experience has shown that this gain in production will normally be followed byan increased production decline rate. In time, the well’s production rate will oftenrevert to its predicted, original (unstimulated) value or even drop below this extrapolatedvalue. The latter occurs if the well’s reserves have undergone an accelerated depletionresulting from the increased well production following the well stimulation.
Thus an estimate of the length of time that the well stimulation treatment will increasethe well production is also required. Remember that these prognoses must not onlyconsider:
(1) the well inflow i.e. whether the well has sufficient inflow capacity andremaining reserves;
(2) the well (tubing) outflow capacity and
(3) whether the production facilities have sufficient capacity to process the extrafluid volumes.
The above, or other techniques, allow the net hydrocarbon production gain from thestimulation to be estimated. These volumes should be reduced by the:
Figure 2
Improved oil production
from a stimulated well
1
8
(i) appropriate discount rate (since hydrocarbons produced today are more valuablethan later hydrocarbon production). N.B. Any reductions in facility capacitywhile the spent stimulation fluids are being treated must also be includedin the economic evaluation.
(ii) hydrocarbon production lost while the well was taken off production to carryout the well stimulation treatment,
(iii) expected chance that the stimulation will be successful (this is often significantlylower than 100%).
The resulting increased revenue can be calculated from this discounted increase in nethydrocarbon production multiplied by the net revenue per unit of production (salesprice minus marginal operating expenditure and taxes plus royalties). This has to becompared with the cost of the stimulation which should include the:
(i) cost of mobilisation and rental of equipment (pumps, tanks etc.) and personnelemployed for the well stimulation treatment. Also the cost of returning the wellto production e.g. initiate production by lifting with nitrogen gas. These costsare related to the type of stimulation chosen and the stimulation treatment size.
(ii) cost of consumables e.g. chemicals etc. used for the well stimulation treatment.This cost is related to the size of the stimulation treatment, while the earlier onesare related to the type of stimulation treatment chosen.
From the point of view of stimulation candidate well selection, the well stimulationtreatment yielding the highest prognosed (discounted) rate of return is the treatmentwhich, in priciple, should be carried out first.
A somewhat simpler calculation method is to calculate the payback time i.e. theproduction time required for the increased, net hydrocarbon production to pay backthe costs of the well stimulation treatment. The most profitable well stimulationcandidate is the stimulation treatment yielding the most rapid pay back. Mostproduction companies require a very high rate of return from this type of welltreatment, leading to pay back times of between 6 and 12 months from this type of welltreatment.
N.B. These economic concepts are treated in greater detail in the economics moduleof this Petroleum Engineering course.
4. CANDIDATE SELECTION
The selection of stimulation candidates that potentially meet the economic screeningcriteria discussed in the previous section is the key to a successful stimulationcampaign. This involves two stages:
(i) the identification and (accurate) quantification of those parameters whichcontrol the productivity of the specific well and
55
Department of Petroleum Engineering, Heriot-Watt University 9
Acidising and Other Matrix Treatments
(ii) an analysis to determine whether the well stimulation treatment would actuallyimprove the well production.
As mentioned earlier, it is important here to distinguish between well (inflow)productivity and well production. This is because an improvement in the well“inflow” performance (from the reservoir to the well) can have only a limited effecton the daily well production if the well “outflow” is limited by tubing / artificial lift/ facilities restrictions. The emphasise the need to consider the flow path from thereservoir boundary to the production storage facilities as a complete productionsystem.
Well Boundary
Pwellhead Pseparator
re
Kd
Pr
P1
P2
P3
P1 = -ve SkinFlowing bottom hole pressure
P2 = Zero SkinP3 = +ve Skin
Gas
GASOil to Tank
Skin (Zone of damaged
permeability)
Reservoir
Choke
Reservoir
Permeability(K)
Table 3 summarises typical values for the minimum requirements for a successful(matrix) stimulation treatment. The criteria in this table can be used for the preliminaryscreening of well stimulation candidates. These figures are based on experience froma number of fields. They will be modified when studying specific conditionspertaining to a given field.
Parameter Oil Reservoir Gas Reservoir
Hydrocarbon Saturation >40% >50%
Water Cut <30% <200 bbls/MMscf**
Permeability † >20 mD * >1 mD
Reservoir Pressure <70% depleted twice abandonment pressure
Production System 20% spare capacity ø
* higher value required for viscous oil productionø lower value if several wells are manifolded together† wells with lower permeabilities are potential hydraulic fracturing candidates** evaluate potential for well killing itself due to liquid loading in the tubing
Table 3
Minimum Screening
Matrix Treatment
Candidate Well Selection
Criteria
Figure 3
The Producing System
1
10
The above criteria evaluate whether the well has a sufficient minimum of:
(i) remaining reserves to justify carrying out the remedial stimulation technique,
(ii) well inflow productivity and
(iii) capacity in the facilities to process the extra fluid production.
The well skin (S) is not mentioned in the table 3 criteria - this is because the skin cannot be considered apart from the other parameters that control the well inflow {thereservoir permeability thickness (k.h) and the potential well drawdown (P
e - P
wf)}.
Thus, a small reduction in skin value from a well with a large reservoir permeabilitythickness can yield a much larger production increase than removal of a high skinvalue from a well with a small reservoir permeability thickness. Alternatively,installation of an artificial lift method which allows an increased drawdown orre-perforation of the producing interval may be the more effective methods ofincreasing production.
Each candidate well needs to be evaluated on its own merits.
4.1. Treatment TimingWell stimulations may be carried out immediately after the initial drilling/completionprogramme has been finalised e.g. to correct formation permeability impairmentcaused by the drilling mud. The well would normally meet the criteria set out in table 3.Alternatively, the stimulation candidate may be identified as a result of routine, fieldproduction surveillance e.g. the well is identified as producing less than the surroundingwells with comparable reservoir quality or reservoir permeability thickness (kh).Figure 4 schematically illustrates the typical output from a modern, productionsurveillance, computer package. This type of graphical display helps with the easyrecognition of potential well stimulation candidates. Once a particular well has beenidentified, its attributes must be checked against the criteria in table 2.
Low Medium High
Low Medium High
Reservoir Premeabilty Thickness (kh)
Relative Bubble Size Reflects Cumulative Well Production
Well location on reservoir map
Bubble
PossibleStimulationCandidate
Key:
Figure 4
Visualising field production
data using Bubble Maps
55
Department of Petroleum Engineering, Heriot-Watt University 11
Acidising and Other Matrix Treatments
4.2. Further Treatment Selection Criteria & "The Stimulation Cycle"Further selection criteria which should be considered include:
(i) Is Sdam
>30% of Stotal
(the total well skin ) ?i.e. could other inflow improving measures (e.g. reperforation ) be a moreeconomical approach to increasing well production? Table 4 gives examplesof poor well productivity which can not be “stimulated away”.
Examples of Possible Poor Well Productivity that can not be'Stimulated away'
1. Gas well with > 200 bbls liquid/MMscf2. Three phase production
1. Gas well - high drawdown2. Oil well - > 20 bbl/ft
1. Oil wells > 5 bbl/day/perf2. Partial penetration resevoir
poor tubingoutflow
non-Darcy(turbulence)effects
CompletionGeometricalskin
Observation Possible cause
(ii) Does the well show sand production?Are sand control measures in place? (matrix stimulation treatments of gravelpacked completions have historically shown a lower success rate than whenperforated completions are treated).
(iii) Is the cause of formation damage known (or at least suspected)?Identification of the cause of the formation damage greatly increases the chanceof matrix treatment success since a treatment fluid which efficiently removesthat specific form of formation damage can be selected.
(iv) Is the stimulation feasible?The final stage of stimulation candidate selection is to evaluate the practicalaspects of the stimulation (e.g. what is the mechanical condition of the well?Are there any logistical, scheduling, or other overriding considerationswhich prevent the well being taken out of production?).
Once the above questions have been answered the following choices can be made:
(i) The composition of the pre and post-flushes and any additives are determinedand the volume of all flushes chosen. Remember that the post flush has tobe displaced to the perforations by a compatible brine or hydrocarbon fluid.
(ii) The detailed treatment design (including injection rates and pressures) can nowbe made and the strategy for returning the well to production chosen.
(iii) The treatment is now carried out and, eventually, evaluated.
The complete stimulation cycle described above is captured in Figure 5.
Table 4
Stimulation Treatments
Selection
1
12
Scheduling
Operational
logistics
constraints
The Stimulation Cycle
EvaluateTreatmentSuccess
IdentifyPotential
StimulationCandidate
EvaluateStimulationEconomics
MatrixStimulationSuitable?
IdentifyFormationDamage
(type + location)
Site and JobPreparation
TREATMENTEXECUTION
ReturnWell to
Production
ModifyStimulationGuidlines
SelectTreatment FluidCompositions,
Additives + Volumes
Select TreatmentType, DiversionTechniques andSpecify Injection
Rates and Pressure
OperationalStimulation
Program(including well
clean-up strategy)
Although well stimulation is often a high reward well activity, it must be reiteratedhere that, for both clastic formations and many carbonate reservoirs:
• Prevention of formation damage is nearly always better than the cure (a remedialwell stimulation).
• Acidisation and other stimulation techniques can create formation damage (seesection 5.6).
5 SELECTION OF (CHEMICAL) TREATMENT TYPE
The chosen chemical treatment fluid should be targeted at the particular type andlocation of the formation damage to be removed or treated. As discussed in chapter4 (Formation Damage), the formation damage/impairment may be related to:
(i) drilling, completion or workover operations,
(ii) produced or (continually) injected fluids,
(iii) injected fluids during specific well operations e.g. well killing.
Table 5 can be used to select the optimum type of chemical treatment once the typeand location of the formation damage/impairment has been identified.
Figure 5
The Stimulation Cycle
55
Department of Petroleum Engineering, Heriot-Watt University 13
Acidising and Other Matrix Treatments
Treatment Type Objective Location Comments
Acid (inorganic) andsolvent (organic)washes
Remove flow restrictionseg. inorganic / wax scale,kill pills, etc
Tubing, perforationand completion
Circulated in well-not injected intoformation
Matrix acidising Dissolve rock components(drill solids, precipitates, clays etc)and formation damage toimprove well / formationconnectivity
Near wellbore(< 1m depth from sand face)
Other matrix treatments
Solvents and surfactants removeemulsions, wax etc. from oily wellsand inhibitor residues from gas wells.
Chemical inhibitor squeezesprevent scaling etc. duringwell production
Near wellbore(< 1m)
yp
at depth (< 10m)from wellbore
Large volume treatmentInhibitor returns withproduced fluid
6. POTENTIAL FORMATION DAMAGE CAUSED BY MATRIXSTIMULATION FLUIDS
The reaction of the formation rock / insitu (formation/injected) fluids with anincorrectly chosen stimulation fluid may generate further formation damage/impairment. Such sources of formation damage include:
(i) deconsolidation of the rock matrix due to the acid dissolving the cementingmaterial that holds the sand grains together. (Temporary) sand production isoften observed when a well is returned to production after and acid stimulation.
(ii) generation of migrating, small diameter particles (“fines”) which can blockthe pore throats. These particles result from the acid only partly dissolvingthe formation minerals present between the grains. This allows insoluble,small diameter, particles to be created and injected into the formation porethroats where bridging and blockage can occur.
(iii) the reaction products created by the chemical reaction between the acid andthe formation rock can be insoluble in the spent stimulation fluid. This iscalled secondary precipitation. This precipitation process leads to blockageof the pores and pore throats (impairment). This precipitation often does notoccur immediately - this implies that the options are to either:
(a) immediately produce the (spent) acid (i.e. return the well to production)or
(b) inject the (spent) acid deep into the formation where any precipitationwill have limited effect on the well productivity.
(iv) fluid incompatibilities. A matrix acidising treatment consists of sequentiallyinjecting a series of fluids. It must be checked that these fluids are compatiblewith each other and with the formation fluids i.e. to not form a (solid) precipitateat the prevailing, downhole temperature when mixed in any proportion.
Table 5
Chemical treatment types
1
14
Further, for the acid fluids which react with the formation; both the “fresh”(unreacted) and the “spent” (reacted) acids need to be tested in this way.
(v) acid precipitation of an insoluble sludge when mixed with the crude oil. Thisis particularly true for asphaltenic crudes and for acids containing ferric cations(rust - or ferric oxide - is a corrosion product produced by steel surfaces e.g. thetubing internals which is dissolved in the acid as it is injected into the well).Such sludge precipitation can be avoided by injecting a compatible hydrocarbon based preflush to displace the crude oil away from the wellbore and thefollowing acid.
(vi) surfactants. Surfactants added to the treatment fluids may create a (highly)viscous emulsion with the crude oil leading to blockage of the pores by this lowmobility fluid. Prescreening laboratory tests in which a sample of the treatmentfluid and the crude oil are shaken together and then the mixture examined foremulsion formation. A series of such trials can be used to select a suitable, non-emulsifying, surfactant.
(vii) wettability changes. Surfactants can change the wettability of the pore surfaces.An oil wet formation has a lower permeability to oil than the equivalent waterwet formation. Once again, any surfactants that are planned to be used shouldbe tested as described above.
(viii)“water blocks”. Significant volumes of water are injected into the formationduring the stimulation treatment. This may lead to an increased watersaturation (water block) in the near wellbore area and can take a long time todisappear (months or even years) from low permeability formations. This canbe minimised by the addition of gas to the stimulation fluid or the (partial)replacement of water by more volatile solvents.
All these forms of formation damage were previously discussed. Non-acid, matrixstimulation treatments also show these effects - apart from (i) and (iii). Surfactants arecapable of mobilising loosely bound, small particles (such as clay particles or other“fines”) which can then reduce the permeability by blocking the pore throats, a sourceof formation damage similar to (ii).
It is clear from above that formulation of the acid and the other fluid flushes employedis the key to ensuring that formation damage due to the stimulation treatment does notoccur or, if it is unavoidable, is at least minimised.
7. MATRIX STIMULATION FLUID SELECTION
The mineralogy and chemistry of the formation, together with the chemistry of theformation fluid and that of the formation damage, combine to give the stimulationfluid selection criteria. This can be summarised as a balance between the:
(i) positive effects e.g. solubility of the formation and formation damage in theselected fluid, and the
55
Department of Petroleum Engineering, Heriot-Watt University 15
Acidising and Other Matrix Treatments
(ii) negative effects e.g. deconsolidation, “fines” generation, secondary precipitates etc.
The type of formation damage present in the well has a large influence on thisselection. Acid is not always the most appropriate fluid - this can be seen from tablesA-1 and A-2 at the end of this chapter which give examples of fluids for various typesof formation damage.
Another factor which needs to be taken into account when designing the stimulationtreatment is the formation and well temperatures both during and after the welltreatment. The speed of reaction of an acid with the formation rock and / or formationdamage will be greater at elevated temperatures (typically doubling for every 10˚Ctemperature rise). Acidic fluids are highly corrosive to the steel surfaces that makeup the completion (tubing / casing / packers etc.). The type and concentration ofcorrosion inhibitor required to inhibit (i.e. limit) this corrosion reaction depends on thetreatment temperature and the treatment time during which the treatment fluids arepumped.
Inorganic scalein perforation
Organic or Inorganicscale in tubing
Type and Location of Formation Damage
Emulsion orWater Block/WettabilityChange
Drilling mudfiltration
Mud FilterCake
FracturesPlugged withdrilling mudor cement
Open HoleCompletion
TypesCased, Cemented,
and Perforated
The types and location of typical forms of formation damage are depicted in figure 6.This, together with knowledge of the completion (perforated) length and associatedformation inhomogeneities (variation in permeability and pressures across the completelength), will determine if special “diversion” arrangements need to be made. “Diversion”techniques ensure that the treatment fluids are evenly “placed” across the formationi.e. that at least the required minimum volume of fluid is injected into each perforationopen to flow.
8. TYPICAL ACID FORMULATIONS USED FOR MATRIX ACIDISING
The behaviour and chemistry of the three most frequently used acids for wellstimulation treatments are discussed in the following sections.
Figure 6
Type and location of
formation damage for cased
and openhole completions
1
16
8.1. Hydrochloric Acid (HCl)Hydrochloric acid is widely available commercially at concentrations up to 28% wt.Its main reaction is to dissolve carbonate minerals (or scale) present in the formationand the well itself e.g. calcite (chalk or limestone), dolomite, siderite etc. The amountof the mineral dissolved is a function of the volume and concentration of acid used:
2HCl + CaCO3→ CaCl
2 + CO
2 + H
2O
or 1m3 of 15% wt HCl dissolves 220kg or 0.09m3 of limestone.
Hydrochloric acid is also capable of dissolving chlorite (an iron containing clay).Hydrochloric acid only shows secondary precipitation reactions when it reacts withiron containing minerals. Highly impairing iron hydroxide is precipitated if the acidbecome spent (i.e. the pH reduces towards the neutral value of 7, see section 5.8.7 -Sequestering Agent).
It is essential to inhibit the acid with a corrosion inhibitor since hydrochloric acid ishighly corrosive to the conventional mild steels used in many well completions.Corrosion inhibition is even more difficult when treating wells completed with (highcost) 13 %wt chromium (stainless steel) alloys or one of the special duplex steels. Itmust also be remembered that hydrochloric acid will dissolve any rust (ferric oxide)present on the tubing wall - even in the presence of a corrosion inhibitor. The potentialnegative effects of ferric cations were discussed in section 5.6.
8.2. Organic AcidsAcetic acid (CH
3COOH) is sometimes employed when it is desirable to use a weaker
or slower reacting acid than hydrochloric acid e.g. the reaction with dolomite is:
4CH3COOH +Mg.Ca(CO
3)
2 → Mg(CH
3COO)
2 + Ca(CH
3COO)
2 + 2H
2O + 2CO
2
Calcium acetate has only a limited solubility - this means that 15%wt acetic acid is themaximum concentration that should be used. Acetic acid has two major advantages:
(i) It is non corrosive to aluminium and (chrome) steel alloys at temperaturesbelow 90oC. Therefore a corrosion inhibitor is not required for these lowertemperature treatments.
(ii) It retains ferric iron in solution as the acid is neutralised (“spends”) by reactionwith the formation. The chemical name for this effect is sequestration - itprevents the formation of ferric hydroxide precipitates from the depleted acid.This avoids the highly permeability impairing form of formation damagediscussed above, that can occur with hydrochloric acid based stimulation fluids.
Formic acid (HCOOH) behaves in a similar manner to acetic acid, but is even weaker(more slowly reacting) than acetic acid.
Organic acids are often used to replace hydrochloric acid for treatments of highertemperature wells e.g. for formations with a temperature greater than 120oC. Many ofthe available corrosion inhibitors are less effective at this temperature. Organic acidsstill require a corrosion inhibitor at this temperature, but the corrosivity is less than thatof a hydrochloric acid based of the acid at the same temperature.
55
Department of Petroleum Engineering, Heriot-Watt University 17
Acidising and Other Matrix Treatments
8.3. Mud AcidThe majority of acid stimulations of clastic reservoirs are carried out with “MudAcid”. This acid is a mixture of hydrochloric (HCl) and Hydrofluoric (HF) acids. Thisvery aggressive acid is capable of dissolving minerals such as quartz, clays, micas etc.(hence the name “Mud Acid”). These minerals are inert to hydrochloric acid alone.NB. The aggressive nature of mud acid also means that it is a major safety hazard tothe wellsite personnel involved in the treatment. Exposure to the acid results in almostinstantaneous blistering of the skin and permanent scarring of the cornea (leading tosight loss).
Mud Acid is made by adding the appropriate amount of solid ammonium bifluorideto the hydrochloric acid solution. Typical formulations used in the field are “FullStrength Mud Acid” (12%wt HCl and 3%wt HF) and “Half Strength Mud Acid”(6%wt HCl and 1.5%wt HF). Acid formulations with lower fluoride concentrationsand higher chloride : fluoride ratio {e.g. 0.5% wt HF and 6% wt HCl} have becomemore popular in recent years, as discussed at the bottom of this section in theparagraphs concerning reprecipitation reactions.
Virtually all the chloride salts formed by the reaction of the formation (damage) withhydrochloric acid have a high solubility. In contrast, some fluoride salts {the simplefluorides (F-) or the fluoro silicates (SiF
6- -)} of sodium (Na), potassium (K), calcium
(Ca) etc. are very insoluble e.g.
CaCO3 + 2HF → CaF
2 ↓ + CO
2 + H
2O
SiO2 + 6HF → H
2SiF
6 + 2H
2O
H2SiF
6 + 2K+ → K
2SiF
6 ↓ + 2H+
The insolubility of these salts implies that mud acid should:
(i) never be diluted with sea water (since it contains calcium and sodium ions).The hydrochloric acid, which forms the basis of the mud acid, is normallydelivered to the wellsite in a concentrated form and diluted on site to thedesignated concentration.
(ii) never be used to acidise a carbonate formation (they contain calcium),
(iii) always be used with a preflush of hydrochloric acid. This preflush should beof sufficient size to dissolve these cations so that they are removed before theformation is contacted by the mud acid. Typically, the volume of the preflushis half that of the mud acid flush. For convenience, the preflush normally usesthe same hydrochloric acid concentration as the main, mud acid flush.
(iv) always be overflushed with a dilute (3%wt or less) solution of hydrochloricacid (HCl) or ammonium chloride (NH4Cl) (ammonium salts have a highsolubility).
The clays, micas etc that are dissolved by the mud acid undergo a series of reactionsthat result in precipitation of silica gel (Si(OH
4) - a hydrated form of silica). This
reaction can be described in simplified form as:
1
18
slowHF + SiO
2 (silica or quartz) → H
2SiF
6 (dissolved)
fast HF + clay → Si, Al in solution
slow H
2SiF
6 + clay → Al (dissolved) + Si(OH)
4 ↓
NB: SiO2 represents quartz or other forms of silica; while Al and Si represents
aluminium and silicon respectively. They are the constituents of clays and micas.
These reprecipitation reactions can not be avoided but their effect can be minimised:
(i) For some formations, an increase in the HCl : HF ratio of the mud acidformulation will reduce the amount of (re) precipitation.
(ii) Since the (re)precipitation reaction occurs slowly, its damaging effects can beavoided by producing the spent acid back rapidly (returning the well toproduction immediately after the treatment has been finished) or overflushingthe mud acid deep into the formation where the effects of the reprecipitation isminimised.
This latter option and the effects of treatment size are illustrated in figure 7.
rd
rd Damaged Zone Radiusk Virgin Rock Permeabiltykd Damaged Zone Permeabilty
k > kd
kacid Fully Acidised Zone, AllSoluble Minerals Dissolved
kprecip Partically Acidised Zonewith Precipitated Acid / RockReaction Products
kprecip < k > kacid
k
rd
kprecip kd kkacid
rd
kprecipkacid k
kprecipkacid k k
1 The Damaged Well
2 A Small Acid Treatment
3 A Large Acid Treatment
4 A Large Acid Treatment Followed by an Overflush
kd
Figure 7
Permeability changes
during a mud acid
treatment
55
Department of Petroleum Engineering, Heriot-Watt University 19
Acidising and Other Matrix Treatments
8.4. Selection of Acid Composition
ACID (HCL/HF)
Impairment
Pore liningclay
Pore fillingclay
Remainingpore space
Increased porosityand permeability andimpairment removal
Secondary reactionproducts ACIDISED
ROCK
NATIVE ROCK
xx x x
xxxx
x xxxxx
xxx
Quartz
Cement (carbonate)
The chemistry of a mud acid treatment is pictured in figure 8. It illustrates how theimpairment, formation clays and inter-granular cements are removed by the mud acidand partially replaced by secondary reaction products. However, there is an overallincrease in porosity and permeability, leading to stimulation of the well.
Formation Characteristics Acid Recommendation Comment
HCl solubility > 15% HCl only HF causes CaF2 precipitation.Organic acids or EDTA/ surfactanttreatments an option.
10% < HCl solubility > 15% Increase HCl preflushvolume
Ensure all carbonate minerals dissolvedprior to mud acid
HIGH PERMEABILITY (> 100 MD)
High quartz (80%); low clay (5%) 12% HCl - 3% HF Regular mud acid
High feld spar/illite (> 15%) 13.5% HCl - 1.5% HF Avoid fluosilicate precipitation
High clay (> 10%) 6.5% HCl - 1% HF Reduce potential migratory fines dueto partial dissolution
High (Fe) chlorite clay 3% HCl - 0.5% HF Avoid iron dissolution. HCl preflush withFe sequestrant an option
LOW PERMEABILITY (< 100 MD)
High quartz (80%); low clay (5%)
High feld spar/illite
High clay (> 10%)
High (Fe) chlorite clay
6% HCl - 1.5%HF
6% HCl - 0.5% HF
6% HCl - 0.5% HF
3% HCl - 0.5% HF
Reduce migrating fines potential
Avoid fluosilicate precipitation
Reduce migrating fines potential
Avoid iron dissolution. HCl preflush withFe sequestrant an option
Replace at least half of the HCl by HCOOH in the region 120…C - 170…C bottom hole temperatureReplace all the HCl by HCOOH for temperatures greater than 170…CNB 1% wt HCl = 1.3% wt HCOOH
Figure 8
A mud acid treatment
Table 6
Acid selection guidelines
1
20
The ideas presented in the previous section, combined with laboratory core floodingand field experience, have shown the need to adjust the mud acid formulation to thetype of formation mineralogy and damage that is being treated. This experience isconsolidated in table 6. Salient points of note are:
(i) Formations with a high (>10%wt), solubility in hydrochloric acid require alarger (hydrochloric acid) preflush. Mud acid should not be used once thissolubility level increases to 15%wt or higher.
(ii) Higher permeability formations with a low clay content can be stimulated withfluids containing a higher HF concentration. Increasing clay contents and thepresence of certain minerals require lower HF acid concentrations and higherHCl : HF ratios.
(iii) A similar approach is followed for lower permeability formations - except thatlower acid concentrations are employed even for the “cleanest” (low claycontent) formations.
This approach can be made even more specific to the particular formation by using achemical thermodynamic simulator to calculate the amounts of rock minerals dissolvedand secondary precipitates formed when a given volume of acid is injected into aspecified formation volume. Coupling of the thermodynamic simulator to a simplifiedreservoir simulation of the injection process allows the effect of the injection of furtherquantities of fresh acid into the partially acidised formation to be simulated. Changesin formation porosity can then be equated with changes in permeability using theKozeny Carmen relationship and the increase in production due to the acid treatmentbe calculated. This process can be carried out for many acid formulations - allowingthe identification of the optimum mud acid formulation of 10%wt HCL and 0.7%wtHF for that particular formation mineralogy (Figure 9).
Insu
ffici
ent H
CI
Con
cent
ratio
n H
F (
% W
t)
Concentration HCI (% Wt)
5
4
3
2
1
0
0 5 10 15 20 25
1.45
1.45
1.401.35
OptimumAcid Formations(10% wt HCl 0.7% wt HF)
Acidized Porosityφ < 35 %
35 % < φ < 40 %
φ > 40 %
Contours With Constant Production Increase (Q acid / Q original)
Composition of Rock
Quartz 74.9 % wtK-Feldspar 7.0 % wtIllite 6.0 % wtKaolinite 10.0 % wtDolomite 0.1 % wtAnhydrite 0.5 % wtAlbite 1.0 % wtSiderite 0.5 % wt
Porosity (%) : 20No Drilling Impairment
This Figure illustrates many of the key aspects of acid formulation selection:
Figure 9
Selection of optimum acid
formulation
55
Department of Petroleum Engineering, Heriot-Watt University 21
Acidising and Other Matrix Treatments
(i) obtain the highest possible Production Increase (Qacid
/ Qoriginal
) consistentwith:
(ii) minimisation of the increase in porosity (i.e. chance of sand production bydeconsolidation of the formation).
It can be seen from figure 9 that this calculation did not include any formation damage.An alternative formulation, 9%wt HCL and 1%wt HF, also meets the above criteria.This second formulation has a higher HF acid content, which would be capable ofremoving greater amounts of, for example, clays from drilling fluids than the 0.7%wtHF formulation. Both these formulations have a high HCL : HF ratio, which is suitablefor treating formations with a significant clay content.
8.5. Selection of Treatment VolumeThe next stage in treatment design is the selection of the treatment volume. This canbe based on:
Formation Temperature <150ºF 150-250ºF >250ºF
Permeability Volume of mud acid (US gal/ft perforations)
< 20 mD 100 50 50
20-100 mD 150 100 100
>100 mD 200 150 100
(i) Field experience when treating wells in the same or similar fields. Oftenbetween 50% and 100% of the volumes suggested in Table 7 are used.
(ii) Practical considerations such as logistics e.g.
(a) how much acid can be delivered to the wellsite? or
(b) how large an acid the volume pumpable during daylight hours?
(iii) Economics (how much acid can we afford based on the expected gain inhydrocarbon production ?)
(iv) Laboratory core flow testing (acid volume required to increase the permeabilityby a target amount). The core may be pre-treated to include damage to thecore inlet face by the suspected form of formation damage.
Laboratory testing in support of matrix stimulation campaigns can also include:
(a) Identification of the formation mineralogy by use of thin sections,X-raydiffraction or Scanning Electron Microscope studies.
(b) Petrographic analysis of formation samples yields information on thetype and location of the minerals, porosity, cementation and clays
Table 7
Typical Mud Acid treatment
volume guidelines
1
22
(c) Fluid-Fluid compatibility tests can be carried out using API RP(Recommended Practice) 42 test methodology. Carrying out these testswill help ensure that all the fluids used are compatible with each otherand the insitu crude oil. Typical tests that can be carried out includechecks to see if:
(i) a sludge is formed by the crude oil* on contacting with acid,
(ii) a “fines” stabilised acid/crude oil* emulsion can be formed,*these two tests require a fresh (non-aged) crude oil sample.
(iii) a viscous oil based mud / acid emulsion is formed.
(v) Theoretical calculation based on the formation mineralogy and formationdamage as used in Figure 9 and described at the end of section 5.8.3. Thistechnique has been applied to a well in which formation damage was presentto a depth of 30 cm. The results summarised in Figure 10. This Figure indicatesthat an acid volume of 16 US gal/ft of perforations is required to achieve theoptimum Productivity Index (PI) for both unimpaired (no formation damage)and impaired (formation damage present) wells. However, this calculationassumes perfect placement of the acid (see section 5.8.8). In practice, a largeracid volume has to be injected to ensure that each perforation receives therequired treatment volume of 16 US gal/ft of perforations.
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
0 16 32 48 64 80 96 112 128 1440.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
Productivity Index (PI) Similarfor Impaired and Unimpaired Well
SteepIncrease
Unimpaired Well Increasing Well Impairment
PI A
cidi
sed
/ PI U
nim
paire
d
Volume of Acid
(m3/m)
(US Gal/ft.)
Notes:
(i) Preflush: 40% of mud acid volume rising to 100% as carbonate contentincreases.
(ii) Main flush: mud acid volume between 50% and 100% of Table 7 values.
Figure 10
Selection of optimum acid
volume
55
Department of Petroleum Engineering, Heriot-Watt University 23
Acidising and Other Matrix Treatments
(iii) Post flush: 10% of mud acid volume if production resumed immediatelyincreasing to 100%-200% if production can not be resumed within 4 hours.
8.6. Selection of Injection RateMatrix treatment fluids have to be injected below the Fracture Propagation Pressure(FPP) (this term is explained in Chapter 6 on Hydraulic Fracturing) to ensure that thefluids are injected radially from the wellbore through the matrix. The maximumallowable injection rate can be calculated from the equation:
qmax = 141.2*10-6 Kav.h (FG.d - ∆ps - pe)
µ {In (rf / rw) + S}
qmax
= maximum injection rate (bpm)h = net treated height (ft)d = well depth (ft)µ = viscosity of injected fluid (cP)p
e= reservoir pressure (psi)
rw
= wellbore radius (ft)K
av= (average) undamaged permeability (mD)
FG = fracture gradient (psi/ft)∆p
s= safety margin* (500 psi)
rf
= radius of injected fluid (ft)S = skin factor
* this saftey margin should be larger if the fracture gradient is not well known.
This version of the radial inflow equation assumes the injected and reservoir fluids donot differ greatly in viscosity and ignores transient flow effects. This equation allowsthe maximum injection rate to be calculated for a constant bottom hole pressure as afunction of skin and injection rate. The Bottom Hole Flowing Pressures can betranslated to Wellhead Pressure by estimation of the hydrostatic head due to thedensity of the fluid in the tubing and any frictional pressure drop. This requiresknowledge of the treatment flow rate & fluid density, depth of the top perforation andinternal diameter roughness of the tubing etc. It can be used to prepare a graph ofWellhead Pressure against the treatment injection rate for a number of different valuesof the well skin can then be prepared Figure 11.
N.B. only injection rates below the Fracture Propagation Pressure need to beconsidered.
1
24
Fracture Propogation Pressure
Maximum allowable pump rate
} 300 psisafety margin
1
2 3 45
6 7
8
9
10
S=340 S=70S=30
S=15
S=7
S=3
S=0
0
0
2 4 6 8 10
500
1000
1500
2000
Treatment Injection Rate (bpm)
Wel
l Hea
d P
ress
ure
(psi
)
kw = 75 mD h = 154 ft rw = 0.35 ft Pe = 3020 psiρ = 8.92 lb/gal FG = 0.74 psi/ft d = 7490 ft Tubing I.D. = 4.56 in
= Pressure and Pumprate measurements taken at various times
This type of plot was first suggested by Paccaloni. He advocated plotting the treatmentpump rate against wellhead pressure so that the reduction in well skin value could bemonitored in real time as the treatment proceeded. The pump rate could also becontinually maximised (which he believed to improve diversion - see section 5.8.9).Figure 11 is used to estimate the current skin from the pump rate / wellhead pressuredata over the wide range of pump rates experienced during the acid treatment. For thisparticular example there is little change in the apparent skin value for the last threemeasurements (points 8, 9 and 10). Paccaloni further proposed that the treatmentshould be stopped once it was observed that the apparent skin was no longerdecreasing.
This Paccaloni approach to injection rate control can not be used if the reservoir isdepleted and the wellhead pressure is zero during the treatment (“the well goes on avacuum”); unless some other means of real-time, bottom hole pressure measurementis available.
8.7. Selection of AdditivesA range of additives to the treatment formulation have been developed to combat oneor more of the forms of formation damage associated with stimulation treatments.They can be expensive, especially those added to acid treatments (e.g. the cost of thecorrosion inhibitor required when acidising a high temperature well can often begreater than that of the acid). Further, many of the additives are incompatible witheach other and may themselves cause formation damage. The use of each additive hasto be justified separately - it should not be just chosen because of the claimedadvantages in the service company sales catalogue!
The more important, frequently used additives are summarised below:
Figure 11
Example of Paccaloni plot
for control of treatment
injection rate
55
Department of Petroleum Engineering, Heriot-Watt University 25
Acidising and Other Matrix Treatments
(i) Corrosion InhibitorsThis additive type is almost always required for acid treatments due to the corrodingreaction of acid on steel:
Fe + 2H+ → Fe++ (dissolved)+H2(gas)
The (acid) corrosion rate increases rapidly with increase in temperature. Further, thecorrosion inhibitor looses its effectiveness as the temperature and treatment timeincrease (due to degredation). A field proven “rule-of-thumb” is that a maximumweight loss of 0.05 lb/ft2 of tubing area (equivalent to the removal of 0.001 in of thetubing wall thickness) during the treatment duration is acceptable . This implies thatthe allowable corrosion rate decreases as the acid treatment time increases. Further,the corrosion should be in the form of a general weight loss type rather than pitting orstress corrosion. The above allows a specification to be developed for the corrosioninhibition of the acid. The type and concentration of corrosion inhibitor chosen willdepend on the acid type, bottom hole temperature, the type of steel contacted and theexpected treatment duration.
N.B. The “spent” acid produced back after the treatment when the well is returned toproduction is often highly acidic and will corrode the tubing at a similar rate to thefresh acid. This is because the corrosion inhibitor has been depleted by reaction withthe acid and absorption on the formation. Injection of extra corrosion inhibitor maybe considered during this phase if, for example, the well is being returned toproduction by nitrogen lifting with a coiled tubing unit.
(ii) Sequestering AgentsBoth the ferrous (Fe++) and ferric (Fe+++) forms of iron will precipitate as the acid“spends” (pH increases). They both form an amorphous, high volume iron hydroxideprecipitate which is highly efficient at creating formation damage. Fe+++ is by far themost insoluble form (see Table 8). Ferric hydroxide already has a low solubility at pHvalues greater than 2, though this value is increased to 6 when mud acid is being usedsince sequestration takes place due to the presence of the fluoride ion (F
-). By contrast,
ferrous hydroxide precipitation is delayed until the pH rises to values greater than 6.
1
2
3
4
5
6
7
8
pH Ferrous Iron Ferric Iron pH Ferrous Iron Ferric Iron
sol
sol
sol
sol
60000 ppm
60 ppm
insol
insol
sol
90000 ppm
900 ppm
insol
insol
insol
insol
insol
NB Iron Sulphides precipitate at any pH when hydrogen sulphide is present
The main source of Fe+++ is the acid reacting with rust in the surface tanks, flowlinesand millscale on the tubing. The Fe++ is mainly (>80%) derived from the formationminerals e.g. chlorite siderite, pyrite etc.
Table 8
Solubility of ferrous and
ferric ions as a function of
pH
1
26
A number of “sequestering” - or solubilising - agents are available to increase thesolubility of iron by forming soluble complexes. The concentration of “sequestrant”required to prevent iron hydroxide precipitation depends on the expected ferric ion(Fe+++) concentration. This is because it is unusual for the pH of the “spent” acid toincrease to a value of 6 when treating clastic formations unless they contain a highpercentage of carbonates i.e. ferrous hydroxide does not normally precipitate underthese circumstances.
The cheapest sequestering agent is citric acid. This has the disadvantage that themaximum concentration allowable - and the maximum amount of iron cations that canbe sequestered - is limited by the solubility of calcium citrate. A more expensivealternative, which can be used at higher concentrations, is EDTA (Ethylene DiamineTetracetic Acid). An alternative approach to preventing ferric hydroxide precipitationis to reduce the Fe+++ to Fe++ by Erythorbic acid or ascorbic acid (vitamin C).
N.B. Fe+++ can catalyse the formation of an asphaltenic sludge when the acid contactssome crude oils.
(iii) Solvents / Mutual Solvents / SurfactantsUse of these materials may reduce emulsion formation but can also be the cause of verystable emulsion formation. Further, they may render the corrosion inhibitor ineffectiveby preventing the absorbtion of the inhibitor onto the steel surface. They can be usefulin some circumstances - but their employment needs to be properly justified and a fullrange of compatibility tests carried out (API RP 42 referred to earlier).
(iv) NitrogenNitrogen gas may be added to the treatment fluid to assist flow back of the spent acidand hence a rapid clean up when treating gas wells or depleted zones.
8.8. Selection of Treatment TypeThe manner of execution of most matrix treatments falls into one of two classes:
(i) “Bullheading” - this term describes a treatment which is pumped down theproduction tubing. The treatment - particularly if it employs acids - can displace rust,scale, pipe dope, etc. present on the tubing’s inner wall into the formation; leadingto formation damage. Sometimes the matrix treatment is carried out prior to runningthe tubing. This can be even worse since dried mud, cement, etc. which are presenton the casing wall can now also be dislodged as well as the damaging materialsreferred to above.
If the formation pressure is sufficiently high, or artificial lift is installed, it is possibleto clean the tubing with a pre treatment (known as “pickling”). This involves injectinginto the tubing a volume of acid - usually equal to 20% of the tubing contents - anddisplacing it until the leading front of the acid is a safe distance above the topperforation. The acid, along with the dislodged, potentially impairing particles, is thenproduced back to the surface.
Bullheading of (cold) fluids from the surface also causes the tubing to contract inlength. If this contraction generates too large tensile stresses in the tubing, the tubingmay part or cause unseating of the packer.
55
Department of Petroleum Engineering, Heriot-Watt University 27
Acidising and Other Matrix Treatments
Both theses problems may be avoided by:
(ii) Pumping the treatment through a coiled tubing (CT) the end of which ispositioned opposite the perforations. Prior to the treatment being carried out theinside of the CT often needs to be cleaned e.g. by “pickling” with acid.However, this can relatively easily be carried out at the surface, if necessary.
Coiled tubing has a smaller diameter than production tubing - so the maximum pumprate is limited (due to friction) and the use of ball sealers for diversion (see Section5.8.9) is only practical if an unusually large diameter CT is employed.
8.9. Selection of Diversion TechniqueMost formations are not homogeneous - in practice the perforated interval will containa number of formation layers with a range of permeabilities and, most likely, differinglevels of skin damage. The treatment fluid injection rate into each layer will begoverned by the radial flow equation since the pressure in the wellbore is in allprobability very similar for all the layers. This situation is depicted in figure 12 whichshows that by far the highest proportion of the acid is injected into the middle, highpermeability layer (Zone B). The natural tendency for the acid to be injected into thislayer was accentuated because this layer also has the lowest skin value. This was dueto the depth of penetration of the formation damage being the least. The Figure showsthat the formation damage in zone B alone was removed by the treatment employingV
1, a small treatment volume. Even doubling the treatment volume to V
2 does not
allow the acid to successfully remove the formation damage in Zone A.
V1
V1
V1
V2
V2
V2
AcidPenetrationFronts
Formation Damage
AcidPenetrationFronts
AcidPenetrationFronts
Zone A
K = 100 mD s = 1.3
Zone B
K = 200 mD s = 0.2
Zone C
K = 50 mD s = 22
Acid thus takes the path of least resistance
Treatment one (V1) employs a smaller volume than treatment two (V2)
Unaltered Reservoir
N.B. The injection rate into a particular layer, relative to the other layers, increasesrapidly once the formation damage has been removed by the treatment. Obtaining aneven distribution of the (acid) treatment is further complicated if the layers havediffering pore pressures.
Figure 12
Matrix (acid) treatment
profiles for a
heterogenuous formation
with varying levels of
formation damage
1
28
Flow Rate
Depth
Flow Meter Surveys
FormationDamage Remains
The Flowing bottomhole pressure was the same for both flow meter surveys
PreStimulation
PostStimulation
PerforatedIntervals
An extra factor to be taken into account when considering acid placement is that, ina bullhead treatment, the treatment fluid reaches the top perforation first i.e. the firstopportunity for the acid to flow into the formation and remove of formation damagewill occur in this top layer. The effect of this process is illustrated in Figure 13. Thisillustrates a flow meter survey made in a vertical gas well before and after a “bullhead”matrix acid treatment. The flow rates were adjusted so that the flowing bottom holepressure i.e. the drawdown, was the same in both cases. The post stimulation surveyshows that the acid job was successful from an economic point of view (50% increasein production at the same drawdown); but that the acid mainly stimulated the topperforations. The reserves from the bottom three quarters of the top perforated intervalare possibly not being produced, unless crossflow is occurring within the reservoir.This differential pressure depletion also could lead to large pressure differencesdeveloping between the various formation layers, causing drilling problems for futurewells.
In practice, many matrix treatments are “bullheaded” into the well but employ one ofthe diversion techniques which have been developed to aid the more even fluiddistribution between the various formation layers. The more frequently used diversiontechniques are described below:
Figure 13
Flow meter surveys of a
vertical gas well made
before and after a
"Bullhead" acid treatment
55
Department of Petroleum Engineering, Heriot-Watt University 29
Acidising and Other Matrix Treatments
(A) Zones A and B to be selectively treated,
without treating zone C.
(B) Install packer betweenZones B and C using wireline
(C) Isolate zones A and B fromthe rest of the well using a
retrievable bridge plug
(D) Run work stringand set inflatable packerbetween zones A and B.Treat zone B by pumping
down work string
(E) Treat zone A by pumping down workstring/
casing annulus
(F) Unset packer and recoverworkstring. Recover retrievable
bridge plug. Return well to production
ZONE A
ZONE B
ZONE C
ZONE A
ZONE B
ZONE C
(i) Mechanical separation using conventional techniques.These techniques include control of the point of fluid injection by use of retrievablebridge plugs placed in packers set between completion zones (see Figure 14), dualpackers on a work string (equivalent to the Selective Placement Tool (SPT), Figure 15),sequential perforation etc.
(ii) Coiled Tubing (CT)Conventional CT may be used with the fluid exit ports at right angles to the tubing sothat the perforations are sprayed with a high pressure jet of treating fluid. Alternatively,the CT may be modified with a single or dual packers to form the SPT (Figure 15 showsa dual packer SPT). These packers are capable of being expanded and deflated manytimes so that all the zones can be treated during one run into the well.N.B. The spacing between the packers is constant while the CT is in the well.
Figure 14
Use of retreivable packers
and bridge plugs to
separately treat the upper
two zones of a multi zone
well
1
30
(A) Running In (B) Performing Treatment
Tubing End Locater
Coiled Tubing Treatment Fluid
UnexpandedPackers
ExpandedPackers
Nozzle
ZONE A
ZONE B
ZONE A
ZONE B
Zone Abeing
treated
Zone Bto be
treatednext
Packers not expanded Packers expanded
(iii) Ball SealersBall sealers are nylon covered balls sized so that they can seal off the perforationsDifferent sized balls are required depending on the perforation diameter e.g. when“big hole” or “deep penetrating” (narrow hole) perforating charges are used. The ballsare pumped whenever it is desired to change fluid injection from one zone to another.Their mode of action is illustrated in Figure 16. It is found in practice necessary topump a 30% - 100% excess of balls above the number of perforations to be sealed. Thedensity of the ball may be chosen so that it:
(i) is buoyant (floats upwards in the treatment fluid due to having a slightly lowerdensity) or
(ii) sinks into the rat hole (i.e. denser than the treatment fluid). The ball densityrelative to the produced fluid controls whether the ball is produced back to thesurface after the treatment is finished.
A
Treatment fluidbeing injectedinto bottom zone
B
Ball sealersbeing carriedby the treatmentfluid to the bottomperforators
C
Bottom zone blocked by lastball sealers,fluid injectiondiverted totop zone
D
Fluid injectionintotop zone
E
Welll returned to production-pressuredifferencial into the wellreleases ball sealers whichare caught at the surface ordrop into rat hole (if densityball > density fluid)
Figure 16
Ball sealer diversion
Figure 15
Selective placement tool
55
Department of Petroleum Engineering, Heriot-Watt University 31
Acidising and Other Matrix Treatments
(iv) Viscous Fluids (gels and foams)The idea behind the use of viscous fluids is that they increase the flow resistance inthe layer taking excessive amounts of treatment fluid so that the fluid is diverted intoa new layer. The skin (S
vis) due to this viscous fluid can be estimated by an equation
similar to the Hawkins formula discussed earlier:
Svis = µvisc -1 In rvisµform re
where µvisc
= viscosity of the viscous fluidµ
form= viscosity of the formation fluid
rvisc
= depth of the invasion of the viscous fluidr
e= well drainage radius
The diversion process is only effective if the viscous fluid is highly shear thinning i.e.its viscosity increases rapidly as its flow velocity decreases at greater depths ofinjection. This allows it to form a viscous “plug”.
(v) Pack the perforation tunnel with a granular particulate (typically ± 200 mD)
(vi) Form a low permeability film, on the wall of the perforation. (typically < 1 mD)
Both techniques are illustrated in Figure 17. The permeability of the diverting agentis mainly dictated by its particle size. Its presence in a perforation will reduce theinjectivity into that perforation, reducing the rate of treatment fluid injection andincreasing the flowing wellbore injection pressure. This will divert the injectionstream into a new zone. Further quantities of the diverter material may be addedcontinuously or intermittently in batches.
Formation Formation
CasingCement
Thin, impermeable (< 1mD)film on perforation wall
Granular Diverter(± 200mD)packs perforation
Diversion with"Film Forming"chemicals
Diversion withGranular particulates
The concept behind the choice of materials used for these two diverting techniques isthat they must be capable of:
(a) being prepared in the required range of particle sizes (hence filtercake permeability),
(b) be stable in the treatment fluid,
Figure 17
Diversion with granular
particulates or film formers
1
32
(c) disappear (dissolve or sublime in the produced fluids) from theperforation so that it becomes fully open to flow once the wellis returned to production and
(d) be non-toxic, cheaply and readily available.
Clearly, a higher concentration of the granular particulate material is required thanwhen a film forming chemical is used . Treatment design is difficult when using eitherof these concepts - the injection profile can be made more uneven (“anti-diversion”)if an inappropriate treatment design is used. In the extreme case, the excessive use ofthe film forming chemicals can result in a complete loss in well injectivity before allthe treatment has been pumped (very embarrassing when acid is in the tubing!).
Typical chemicals used for both particulate and film forming diversion agents, bypreparing them in the correct particle size ranges, are:
(a) Benzoic acid (water / oil / gas),
(b) Sodium Chloride (NaCl or rock salt) crystals (water) and
(c) Oil soluble resin (oil) particles.
N.B. These particles are soluble, or sublime in the case of gas, in the phases indicatedin brackets after the name of the chemical.
9. MATRIX STIMULATION FIELD CAMPAIGNS
From a technical point of view, the preferred way to carry out matrix treatments in anoil or gas field is to treat a number of wells in the form of an organised campaign, ratherthan carrying out the work on an ad hoc basis. This allows the design methodologyintroduced in Figure 5 to be applied, i.e. that experienced gained in EVALUATINGearly treatments is used to improve the efficiency of later treatments.
Figure 18 compares the pre and post stimulation well productivity of a number of wellsin a field both before and after matrix stimulation treatments were carried out. Thedecrease in average well skin achieved by the matrix stimulation is clear. In fact, thesegains show that matrix stimulation is often a high reward activity. In this case, despitesuffering a 20% treatment failure rate, the stimulation campaign delivered anincreased production capacity at 10% of the cost of generating the same capacitythrough infill drilling.
55
Department of Petroleum Engineering, Heriot-Watt University 33
Acidising and Other Matrix Treatments
No
. of
Wel
ls
AveragePre / Post Matrix Stimulation
Well ProductivityDistribution
Prior to MatrixStimulation
PI actual / PI undamaged
00
3
6
9
12
15
0.2 0.4 0.6 0.8 10.35 0.65
The operator followed the Matrix Stimulation Design Methodology set out in figure5 and by evaluating the treatment's performance and was able to halve the acid volume(per net unit perforated interval length) during the course of the campaign; while stillmaintaining the same, favourable, well response to stimulation.
A second example of the results of following this stimulation cycle methodology isshown in Figure 19. Careful candidate selection ensured that the wells selected earlyin the campaign showed higher levels of formation damage - hence higher productiongains were achieved compared to later in the campaign. However, cost savingmeasures were so effective that, despite the average production gain nearly halvingduring this five year period, the cost per unit of production gain (bopd or barrel oil perday) was also halved, i.e. later treatments were more economic.
'87 '88 '89 '90 '91 '92
0
200
400
600
800
1000
1200
150 80
Cost (US$/bopd)
Time
Productionafter MatrixAcid Stimulation (bopd)
Example Stimulation Campaign
In a third case history and operator achieved and average technical cost of US$ 0.40/bbl oil gained from their stimulation activities. This cost represents a high return on
Figure 18
A stimulation campaign
Figure 19
A second example of a
stimulation campaign
1
34
the capital employed. It is obviously greatly influenced by the candidate selectioncriteria used.
10. STIMULATION OF CARBONATE FORMATIONS
Acidising of carbonate formations is fundamentally different from the acidising ofclastic formations. This is due to their differing physical nature and chemistry:
(i) Carbonates consist of very fine grains exhibiting a vugular or fracture porosityrather than the intergranular porosity shown by sandstones.
(ii) Carbonates react much more rapidly with hydrochloric acid than sandstones,for the same formation temperature. Also, the use of mud acid is prohibited dueto the limited solubility of calcium fluoride.
Carbonates are normally found as massive deposits of chalk, limestone or dolomite.Their constituent particles are much smaller than the typical sand grains found inclastic formations. They will have undergone large porosity and permeabilityreductions during burial and diagenesis. Although they are often pure (>95% wtcarbonate), they can also include iron minerals, clays and silicaceous materials givingthem a very variable composition.
The many possible diagenetic processes can lead to formations with similar chemicalcompositions having a strength that varies from very strong to behaving similar totoothpaste. Strong and weak layers can be present a small distance apart. Thiscomplicates the planning of well completion - and stimulation - procedures.
10.1. Acid Composition SelectionHydrochloric acid is used to:
(i) bypass drilling or completion damage by dissolving the rock matrix;
(ii) widen natural fractures or secondary porosity so as to improve fluid conductivityto the wellbore (see section 5.11.3);
(iii) increase the effective wellbore radius by wormhole formation.
Dolomite reacts much more slowly with Hydrochloric Acid than chalk or limestone- the optimum reaction rate is achieved with a concentration of 28% wt HCl acid forall dolomitic reservoirs. 15% wt HCl is used with the other carbonate formation types.
The amount of rock dissolved by the acid is determined by:
constant {volume acid * concentration acid * reaction stochiometry}
The constant depends on the units employed e.g. 1m3 of 14% wt HCl will dissolve 206Kg of limestone with a volume of 0.073m3 assuming a porosity of 5% vol. Thecorresponding amounts dissolved when dolomite is treated are some 7.5% smaller.
55
Department of Petroleum Engineering, Heriot-Watt University 35
Acidising and Other Matrix Treatments
The injected acid does not dissolve the rock uniformly, instead it forms “wormholes”- see Figure 20a and 20b. “Wormholes” consist of a main channel from which manyhighly branched structures are formed. The number and extent of the wormholesdepend on:
(i) the carbonate formation’s reactivity (high reaction rates promote few, longwormholes);
(ii) The acid leak-off rate into the matrix (controlled by formation permeability,acid and formation fluid viscosities and the injection pressure overbalance);
(iii) The presence of higher permeability streaks, fractures, vugs etc will determinethe preferred direction of wormhole growth.
������������
yyyyyyyyyyyyCasing
Cement
Carbonate Formation
�������������������������
yyyyyyyyyyyyyyyyyyyyyyyyy
Borehole Wall
Pre-Perforated Liner
Figure 20a
Wormhole formation from
a perforation during matrix
acidising
Figure 20b
Wormhole formation from
an open hole completion in
a carbonate formation
1
36
10.2. Treatment Types for Carbonate Rock Acidising
10.2.1. Matrix TreatmentsWormhole formation during matrix treatments improves the well inflow performanceby providing a high conductivity channel at depth from the wellbore. They are createdusing either a:
(a) low rate, low volume, low acid concentration treatment.(Typical values are 0.004m3/min/m, 0.3m3/m and 14% wt HCl for the injectionrate, injection concentration and acid concentration respectively). The low rateand long contact time encourages wormhole formation and the bypassing ofshallow formation damage. This type of treatment is most suitable for shortintervals (< 12m).
(b) high(er) rate, large volume, high concentration treatment.(Typical values are 0.025m3/min/m, 1.6m3/m and 14-28% wt HCl respectively).The larger acid volume compensates for the reduced wormhole formationcaused by the use of the higher pump rate. Ball sealers are more effective in highrate treatments - making them more suitable for treating longer perforated zones.
Both types of treatments have been applied with success - the preferred methodprobably depends on the local situation with regard to formation damage, presence ofnatural fractures & vugs etc.
10.2.2. Acid Wash (or Soak) Type TreatmentsWormhole formation is undesirable if the treatment objective is to remove near wellbore damage (e.g. perforations plugged with drilling mud, cement etc) present in a newcompletion or after a workover. This is because forming the wormhole will consumea large part of the available acid. Wormhole formation is avoided by keeping theinjection rate very low (<0.0002m3/min/m). The treatment is now called an acid washor soak. Treatment guidelines are:
(a) use the highest possible HCl acid concentration (max 28%) permitted bycorrosion considerations for the planned treatment time (10 hours or longer);
(b) use a coiled tubing (CT) to place a volume of acid, equal to the casing volumeof the perforated interval. If CT is not available the tubing should be “pickled”(see section 5.8.7);
(c) The acid should contain an iron sequestrant (see section 5.8.6) and, if an oilbased mud cake is to be removed, a mutual solvent or dispersing surfactant (seealso section 5.11.3).
55
Department of Petroleum Engineering, Heriot-Watt University 37
Acidising and Other Matrix Treatments
11. ACIDISING OF SPECIAL WELL TYPES
Some examples of well types requiring special treatment techniques.
11.1. Gravel Packed Wells
This well type poses a number of areas where fluid inflow can be impeded viz thescreen, perforations, gravel, grave/sand interface and the near wellbore formation.Dedicated treatments require identification of the type and location of the formationdamage eg:
(i) the screen: removal of material plugging the screen requires placing (or“spotting”) the acid across the entire screen length while minimising inflowinto the gravel pack. This is best done by use of a CT which is moved across thecompletion while pumping acid. A small acid volume is used e.g. 120% of thewellbore completion volume.
(ii) the gravel: removal of residue left from the viscosifier. Viscosifier residuesare normally best removed with solutions of enzymes (for formation temperature <65˚C), hypochlorite (a bleach) or 2% wt hydrochloric acid (at highertemperatures). A treatment volume equal to 120% of the gravel pack volumeshould be used together with a diverting agent which passes through the gravelpack sand and filters out on the formation.
(iii) near wellbore formation damage: this should be treated in a similar mannerto perforated completions, except that the available diversion techniques aremuch more limited due to the presence of the gravel pack.
11.2. Horizontal WellsTreatment of horizontal wells is no different from conventional wells with respect tothe candidate and fluid selection criteria concerned. The most common, horizontalwell completion techniques use a slotted or perforated liner or a wire wrapped screen.Cased, cemented and perforated completions are less frequent because of the greatercosts. Many of the conventional diversion treatments (ball sealers / chemicaldiverters) would not be successful in the horizontal orientation since the treatmentfluid / diverter density differences will result in shutting off either the top or bottomof the completion section. The SPT (Figure 15) is not effective in the uncemented lineror screen completions because of the presence of the open annulus. Diversion witha viscous fluid (foams or gels) is a possibility, but the very large volumes required maymake them operationally impractical.
1
38
�������������������������������������������������������
yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyFORMATION
FORMATION
Mud
cake
Rem
aini
ng
on O
pen
Hole
Sect
ion
Cemen
t
Wor
kstri
ng o
r
Coile
d Tu
bing
Line
r
Exte
rnal
Cas
ing
Pack
er
Polis
hed
Bore
s
Spac
ed F
ar A
part
Polis
hed
Bore
s
Spac
ed C
lose
Tog
ethe
r
Mud
cake
Bei
ng
Remov
ed b
y Tr
eatm
ent
Flui
d
Polis
hed
Bore
s Reg
ular
ly
Spac
ed B
etwee
n th
e EC
P's
Slot
ted
or P
erfo
rate
d
Line
r or W
ire
Wra
pped
Scr
een
Rubbe
r Sea
lsInjected
Treatment Fluid
Fluid Returns to Surface
Mechanical diversion, using the set up shown in Figure 21, is a possibility. Thehorizontal well has been divided into a number of sections using external casingpackers. Rubber sealing elements installed on the outside of the CT seal against apolished bore installed at regular intervals inside the liner or screen. This arrangementallows for the completion zone between the external casing packers (ECP) to beselectively treated in two manners:
(i) circulation of the treatment fluid passed the mud cake, when the flow returnsto the surface via the CT / tubing annulus. This will aid removal of the mud cake;
(ii) injection of the treatment fluid into the formation by closing the wellhead valvewhich allowed this annular flow.
The regular arrangement of ECPs and polished bores allow the treatment to berepeated along the length of the completion interval.
This method requires the installation of a large number of ECPs and polished boresduring the initial completion of the well. This is expensive (increasing the well cost)and also presents an increased risk of failure during the initial completion. It is alsooften not known in advance whether - and at which point in the horizonal wellbore -such treatments will be required.
Figure 21
Removal of mudcake from
behind slotted liner using
washpipe and seals
55
Department of Petroleum Engineering, Heriot-Watt University 39
Acidising and Other Matrix Treatments
Open Fracture
Blocked Channel
11.3. Naturally Fractured Formations(Natural) fractures often have a conductivity many times that of the formation. In fact,in many carbonate formations the matrix has a very low permeability and all theproduction comes from the fracture. Stimulation may have two possible objectives:
(i) removal of damage to the fracture conductivity by drilling mud or cement;
(iii) enhancing the (natural) fracture conductivity by dissolving the cementingmaterials deposited in the fracture.
As usual, fluid acid selection depends on the treatment objectives. A key point in thefluid selection procedure is whether the fracture contains materials insoluble inhydrochloric acid - if so these insoluble materials will remain in the fracture and blockthe created permeability. This is illustrated in Figure 22. The top picture shows a cleanopen channel created by acidising a pure, fractured limestone. It appeance iscontrasted with that of a similar experiment carried out on an impure, silty limestone(lower picture). In the latter case the channel has been blocked by insoluble particlesremaining after the acidisation is completed. These particles can be mobilised by theuse of a silt suspending agent (a surfactant) so that they are flushed out of the fracturewhen the well is returned to production.
Similar effects are observed when creating wormholes during matrix acidisingtreatments of impure carbonate formations (see point c at the end of section 5.10.2).
If more than one fracture is connected to the wellbore, then ball sealers can be usedto divert the acid treatment from one fracture to another. In addition, the pump ratecan also be used to aid acid placement within the fracture.
(i) low pump rates (and hence long contact times) enhance removal of mud cake/cement etc. in the near wellbore region;
(ii) high pump rates - even exceeding the FPP - encourage removal of fracturefilling materials from at depth from the wellbore.
Figure 22
(Acid) insolubles remaining
after the acid frac of impure
carbonate formation will
block channel if not
removed
1
40
12. ALTERNATIVE ACID FORMULATIONS
Proprietary acids based on other chemistry than that described here has been found tobe useful in specific circumstances e.g. fluoroboric acid has been shown to be usefulin gravel packed, gas production wells producing from clay rich formation prone topermeability damage due to fines movement. Conventional mud acid was shown tobe effective at removing this damage (figure 23), but a rapid decrease in wellproductivity over the next year resulted from further formation damage as shown bythe increase in turbulence factor during this period (figure24).
2,800
2,300
1,800F
TP,
psi
Gas
Pro
duct
ion,
MM
scfd
Trun
klin
eR
estr
ictio
n
ProprietaryAcid Treatment
Time, Years
Declineafter
Mud Acid
18
16
14
12
10
8
6
4
210 2 3
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.00.0 5 10
One year afterMud acid treatment
Immediately aftermud acid
Rate, MMscfd
Squares measured> four years afterproprietary acid
Line measuredimmediately following
proprietary acid
(P2 r-P
wf2 )
/q, p
si2 /
scfd
Use of the proprietary acid, by contrast, lead to a sustained increase in well productionwith no increase in turbulence over a four year production period. N.B. carefulselection of the mud acid compaction (table 6) might have achieved a similar long termbeneficial effect to the Proprietary acid.
Figure 24
Turbulence analysis plot
for the gas well shown in
figure 23
Figure 23
Production history for a gas
well prone to formation
damage due to fines
movement
55
Department of Petroleum Engineering, Heriot-Watt University 41
Acidising and Other Matrix Treatments
Table A-1
Matrix fluid selection
13. Appendix A :
Matrix fluid selection chart depending formation damage type to beremoved
TypeDamage
Symptom Cause Solution
Oil wet formation Reduced oilproduction
Corrosion inhibitors.Oil based mud.
Mutual solvent* / waterwetting surface
Water block Reduced gasproduction in lowpermeability ordepleted formations
Increased near wellbore,water saturationFluid loss duringdrilling / work over
Surfactant / alcoholsolution injectedwith nitrogen
Emulsion Viscous emulsionnear wellbore
Emulsion stabilised bysurfactant or oil wetsolids e.g. asphaltenes
Inject mutualsolvent / demulsifier
Wax Deposit on tubingor in formation
Oil cooled belowcloud point
Dissolve in heated oil /solvent or disperse insurfactant solutionMechanical removed (scraping)Crystal growth inhibitors
Asphaltene Deposit in facilities,tubing or information
Pressure reduction(precipitation oftenstarts near bubble point)
Mechanical removalDissolution and inhibitionnot very effective
BacterialSlime orDeposit
Deposits (mainlyin well) and corrosion
Bacterial Infestation Bactericide prevents infestationMechanical removed andoxidising agents (e.g. bleach,chlorine dioxide removesdeposits)
Silts and Clays Reduced wellinflow
Clay swelling and fines migration
Clastics - dissolve damagewith HCI/HFCarbonates - HCI to dissolveformation / bypass damage
InorganicScale
Deposit on wellequipment or information
Formation brinesbecoming super saturateddue to T and P reductionor mixing of incompatiblebrines
Mechanical removalin tubingInhibition (inject inhibitor in tubing or squeezeinto formation)Dissolution (see table A-2)
Matrix Fluid Selection
* Mutual solvents dissolve both oil and water phases
1
42
Type of Scale Usual Manifestation Treating Fluid Comments
Carbonates CaCO3 HCl very common treatment
Sulphates CaSO4 / CaSO4.2H2O
BaSO4 / SrSO4
"Conversion Treatment" (Na2CO3 solution followedby HCl) or EDTA
EDTA + Synergist
common treatment
very expensive
Chlorides NaCl H2O gas well only
gas well onlySulphur S liquid at most downholetemperatures
Iron FeS / Fe2O3 HCl + chelating agent,also sometimesreducing agent
encountered in anacidic environment(CO2 / H2S)
Silica /Aluminosilicate
SiO2 / NaAlSiO4 HCl / HF these scales causeFormation Damage as well as deposit in tubing
NB Scales are inorganic deposits formed by supersaturation of formation brines or mixing of two incompatible brines. Mechanical removal is always an option in the tubing.
Fluid Selection for removal of Inorganic Scale
MATRIX STIMULATION TUTORIAL
Question 1
1. List three important factors each from drilling and production operations thatreduce well inflow that can be changed by matrix stimulation. Comment onconsiderations that will guide the choice of the stimulation fluid.
Answer 1
Drilling and Completion Fluids1. Invasion of solid particles from the drilling mud lead to blocked pore throats andhence reduced permeability
2. These solids have many sources e.g. drilled solids, insoluble contaminants withinthe base chemical components that were used to make up the drilling or completionfluid, corrosion products eroded from the walls of the casing or work string etc.
3. Fluid loss from drilling and completion fluids into the formation increases thenear wellbore liquid saturation and hence reduces the (relative) permeability to oilflow. This is often called a “Water block” when water based fluids are lost to theformation.
4. Polymers absorbing from the drilling & completion fluid reduce the (relative)permeability to oil flow.
5. Fluid loss into an formation incompatible with the this fluid leading to clayswelling and reduced permeability
Table A-2
Fluid selection for removal
of inorganic scale
55
Department of Petroleum Engineering, Heriot-Watt University 43
Acidising and Other Matrix Treatments
Production Operations:1. Inorganic scales formed by mixing of incompatible brines within the well or dueto precipitation induced by temperature and pressure changes as the brine isproduced up the well.
2. Wax precipitated from cooling of the oil as it is produced up the well.
3. Asphaltene solids precipitated the oil phase due to pressure decreases undergoneby the oil as it passes through the production system.
4. “Fines” (very small diameter particles) that are dislodged by the flow of theproduced fluid followed by their migration through the pore spaces. They tend tolodge in the pore throats in the near wellbore area. This is because their concentrationis highest here due to the nature of the radial flow process towards the well.
Matrix stimulation fluid selectionThe Matrix stimulation fluid should remove (dissolve, solubilise or at least mobilisethe source of the formation damage, e.g.
wax - hot, organic fluidscalcium carbonate scale - hydrochloric acidclay particles - mud acid“water block” in a oil well - mutual solvent that solubilises both oil &
water phases
Bonus marks: choice of fluid has to maintain the materials in solution (avoidreprecipitation of the dissolved solids) and be compatible with the formation.
Question 2.
You are required to rank the 5 wells in the PetEng field in the order that they shouldbe acidised during the planned well stimulation campaign. You are required to justifyyour selection in terms of the potential production increase and treatment costefficiency (US$ per additional BPD oil production capacity). The artificial liftinstallations in each well and the surface production facilities have sufficient capacityto cope with the production increases.
The completions installed for all the 5 wells is similar with a wellbore radius (rw) of0.3542 ft and a drainage radius (re) of 800ft. The full formation height is fullyperforated with an interval height (h) of 100ft. The reservoir oil has a viscosity (Uo)of 0.92cp under downhole conditions and the volume shrinkage value during production(B
o) is 1.35 bbl/STB. The formation porosity is 33% and is independent of permeability.
The wells were all damaged during the completion phase due to lack of fluid losscontrol - greater volumes of (damaging) completion fluid were lost into the higherpermeability wells. The fluid loss caused a smaller (percentage) damage in the higherpermeability, more productive wells. One pore volume of acid to be injected into eachwell to remove this formation damage.
1
44
Well No. Parameter Units 1 2 3 4 5Undamaged permeability (k) mD 10 50 100 250 500Depth of damage in 4 5 8 10 12Percentage permeabilty damage % 80 75 50 30 25Producing drawdown "DP" psi 120 120 120 120 120
You have negotiated a stimulation contract with the service company in which the costis related to the volume of acid pumped. The charge is 1500 US$/ft3 of stimulationfluid pumped.
Two useful equations are the Hawkins Formula to calculate the skin (S) and that forcalculating the well’s productivity ratio without and with formation impairment (Jo/Ji):
Skk
Inrr
JoJi
Inrr
Inrr
S
d
d
w
e
w
e
w
= −
=
+
1
while the oil production rate can be calculated from the radial inflow equation:
qk h P P
B In rr S
oe wf
o oe
w
= −
+
141 2
* * ( )
. * * *µ
The net revenue from the produced oil is $10 /bbl and your companies economiccriteria demands a 4 month pay back line.
Answer 2.
The first requirement is to calculate the Skin effect resulting from the formationdamage.
Well No. Units 1 2 3 4 5Damaged permeability (kd) mD 2 12.5 50 175 375Permeabilty ratio (k/kd) 5.00 4.00 2.00 1.43 1.33Damage radius (rd) ft 0.69 0.77 1.02 1.19 1.35Skin (S) 2.66 2.33 1.06 0.52 0.45
Rank Order 1 2 3 4 5
Well No. 1 is the well with the highest skin (most heavily damaged). The apparentstimulation candidate ranking using the Skin value as the ranking criteria is Well No.1, 2, 3, 4, & 5.
55
Department of Petroleum Engineering, Heriot-Watt University 45
Acidising and Other Matrix Treatments
However, this raking by skin factor does not take into account that the wells show awide formation permeability range. The potential production increase from the wellsachieved by removing the damage represents the payback for the stimulation treatment cost:
Well No. Units 1 2 3 4 5Skin (S) 2.65 2.33 1.06 0.52 0.45Productivity ratio (Jo/Ji) 0.74 0.77 0.88 0.94 0.95Production rate with damage (Qi) bpd 66 340 779 2,076 4,188Undamaged production rate (Qo) bpd 89 443 886 2,215 4,430Increase in production (Qo-Qi) bpd 23 103 107 139 242
Rank Order 5 4 3 2 1
It can be seen that the stimulation candidate ranking order is reversed once we basethe ranking criteria on the potential gain in production capacity.
A different picture results gain if we base the ranking criteria on economics, based onthe fact that we know the depth of the formation damage and can thus alter the acidtreatment size depending on the volume of rock to be treated.
Well No. Units 1 2 3 4 5Increase in production (Qo-Qi) bpd 23 103 107 139 242Formation Pore Volume ft3 36 49 96 135 179from which damage removedCost treatment US$ 54,545 73,635 143,997 201,812 268,355or US$ per BPD additional US$/bpd 2,407 716 1,348 1,448 1,107production Rank Order 5 1 3 4 2
Well No. 2 is now the most attractive stimulation candidate since it minimises theinvestment required for the gain in production capacity. This comes about becausewell No. 2 had a shallow depth of impairment combined with a reasonable productionrate response to the removal of the formation damage. The ranking order is quitedifferent from either of the other two quoted above. This criteria is often the best onesince the alternative investment is often to drill a new well or to sidetrack an existing well.
The four month pay back whether time criteria can now be used to determine if thetreatments are economic. The preferred order in which the wells can be treated is alsoshown
Well No. Units 1 2 3 4 5
US$ per BPD additional production US$/bpd 2,407 716 1,348 1,448 1,107
Pay back time days 241 72 135 145 111
Economic to carry out treatment? No yes No No Yes
Treatment order - 1 - - 2
However, note that well 5 is producing more than 60% of the fields total production.It would be unwise to place this well at risk by carrying out an acidisation treatmentfor a potential 5% increase in production. Thus well 2 is the only stimulationcandidate.
1
46
14. FURTHER READING
“Production Operations” Volume 2 (4th edition)by T. Allan and A. Robertspublished by Oil and Gas Consultants IncISBN 0-930972-18-X
“Petroleum Production Systems”by M.J. Economides, A.D. Hill and C. Ehlig-Economidespublished by Prentice HallISBN 0-13-628683-X
“Well Performance” (2nd edition)by M. Golan and C. Whitsonpublished by TapirISBN 0-13-9046609-6
“Reservoir Stimulation” (2nd edition)edited by M.J. Economides and K.G. Notlepublished by Schlumberger Educational ServicesISBN 0-13-775115-X
“Acidising Fundamentals”by B.B. Williams, J.L. Gidley and R.S. Schecterpublished by The Society of Petroleum EngineersISBN 0-89520-205-0
C O N T E N T S
1. INTRODUCTION2. HYDRAULIC FRACTURE TREATMENT
SELECTION GUIDELINES3. FRACTURE STIMULATED WELL INFLOW
PERFORMANCE4. A PROPPED HYDRAULIC FRACTURING
TREATMENT5. TYPES WELL FRACTURING STIMULATION
TREATMENTS6. A BRIEF REVIEW OF ROCK MECHANICAL
ISSUES RELATED TO HYDRAULICFRACTURING6.1. Insitu Sress6.1.1. Effective Stresses6.1.2. Fracture Initiation and Perforation
Programme6.1.3. Data Gathering6.2. Fracture Size6.2.1. Fracture Containment6.2.2. High Insitu Sress Contrasts6.2.3. Fracture Growth into Boundaries6.2.4. Fracture Height Measurement6.2.5. Fracture Azimuth and Well Deviation
Orientation6.3. Modelling the Shape of the Induced
Fracture6.3.1 2D Fracture Models6.3.2. Fracture Width
7. CREATION OF A PROPPED HYDRAULICFRACTURE
8. HYDRAULIC FRACTURING IN PRACTICE9. OPTIMISATION OF HYDRAULIC
FRACTURE DIMENSIONS10. PROPPED FRACTURE CONDUCTIVITY11. THE INFLUENCE OF FRACTURING FLUID
AND THE FRACTURE CONDUCTIVITY12. FRACTURING FLUID13. TIP SCREEN OUT FRACTURING
13.1. Applications of TSO Fracturing14. FURTHER READING
6Hydraulic Fracturing6
2
LEARNING OBJECTIVES:
• List the nomenclature of propped Hydraulic Fracturing (HF)
• Describe the factors which control the Productivity Increase Factor (PIF) achievable by HF
• Relate PIF to Net Present Value economics as a function of treatment size so as tooptimise HF treatment design
• Explain the role of Rock Mechanics in supplying basic design data for an HFtreatment
• Identify the key elements of the Rock Mechanics of Fracture Initiation andPropagation
• Analyse Fracture Propagation Pressure Record to derive basic design data
• Discuss the importance of the perforation programme design to the success of anHF treatment
• Distinguish between the different Fracture Propagation Models
• Explain how to select fracturing materials (fluids/proppants) for an HF treatment
• Discuss the factors influencing Hydraulic Fracture geometry (fracture shape and length)
• Critically describe the Hydraulic Fracture Treatment Design Procedure
• Describe the stages of a Hydraulic Fracturing Treatment operation
Department of Petroleum Engineering, Heriot-Watt University 3
6Hydraulic Fracturing6
1. INTRODUCTION
Propped Hydraulic Fracturing consists of pumping a viscous fluid at a sufficientlyhigh pressure into the completion interval so that a two winged, hydraulic fracture isformed. This fracture is then filled with a high conductivity, proppant which holds thefracture open (maintains a high conductivity path to the wellbore) after the treatmentis finished (Figure 1). The propped fracture can have a width between 5mm and 35mmand a length of 100m or more, depending on the design technique employed and thesize of the treatment.
An Acid Fracture
Sid
e V
iew
Cro
ss S
ectio
nAcid EtchedChannels
Top Boundary
Confining Shale
Propped Hydraulic Fracture
Sid
e V
iew
Cro
ss S
ectio
n
Confining Boundary
Confining Boundary
Carbonate orSandstoneFormation
Confining Shale
A strong, inhomogeneous carbonate formation
Both types of fracturing treatments create highly conductive paths from deep in thereservoir to the wellbore
Propped hydraulic fracturing is aimed at raising the well productivity by increasingthe effective wellbore radius for wells completed in low permeability carbonate orclastic formations. The radial well inflow equation:
Q = = kh(Pe - Pwf) kh(Pe - Pwf)
µB0 In +S141.2 141.2rerw r'
wµB0 In
re
shows that the well production rate (Q) can be increased by:
Figure 1
Fracturing Concepts
4
(i) increasing the formation flow capacity (k.h) {the fracture may increase theeffective formation height (h) or connect with a formation zone with a higherpermeability (k)};
(ii) bypassing flow effects that increase the skin (s) e.g. near wellbore formationdamage;
(iii) increasing the wellbore radius (rw) to an effective wellbore radius (r'
w) where
r'w is a function of the conductive fracture length Lf (see Figure 2a).
pe
Lf
Lf
kf
re
r'w
2rw
w
2rw
h
ImpermeableFormationBoundary
ImpermeableFormationBoundary
Fracture
Fracture
w
Wellbore
Formation Permeability (k)
k = 0.1 mdk = 1.0 md
k = 10,000 md
k = 1,000 md
k = 100 md
k = 10 md
100
10
1
0.01 0.1 1.0 10 100 months
Pro
duct
ivity
hyd
raul
ical
ly fr
actu
red
wel
lP
rodu
ctiv
ity o
f uns
timul
ated
wel
l.
Time
Figure 2a
Propped hydraulic fracture
geometry
Figure 2b
Production increase due to
150 ft long hydraulic
fracture with a flow
conductivity of 8,000 mD ft
Department of Petroleum Engineering, Heriot-Watt University 5
6Hydraulic Fracturing6
If the hydraulic fracture has infinite conductivity i.e. the pressure drop along its lengthdue to flow is negligible, then:
rw' = L
f/2
Thus high conductivity fractures allow fluids to flow to the well whose effectiveradius has been enlarged to a value equal to half the single wing fracture length.Alternatively, if the actual wellbore radius is used, this improved inflow can beexpressed as a negative skin.
The relative increase in production achievable by placement of a hydraulic fractureis much greater in the case of low permeability formations (see figure 2b).
2. HYDRAULIC FRACTURE TREATMENT SELECTION GUIDELINES
Hydraulic fracture stimulation is required for the economic development of lowpermeability reservoirs. This is because a highly conductive fracture results in anegative skin. The wellbore flowing pressure (P
1) has been increased, at a given flow
rate, compared to an unimpaired (P2) or impaired (P
3) well (see Figure 3):
Well Boundary
Pwellhead Pseparator
re
Kd
Pr
P1
P2
P3
P1 = -ve SkinFlowing bottom hole pressure
P2 = Zero SkinP3 = +ve Skin
Gas
GASOil to Tank
Skin (Zone of damaged
permeability)
Reservoir
Choke
Reservoir
Permeability(K)
(i) the pressure observed (P2) for the same flow rate for a well with an ideal
(S = 0) completion or
(ii) the even lower pressure (P3) measured for the equivalent well showing a
positive skin due to formation damage.
The hydraulically fractured well with the negative skin will have the greatestproduction rate. Propped hydraulic fracture well stimulation should only be consideredwhen the:
Figure 3
The Producing System
6
(i) well is connected to adequate produceable reserves;
(ii) reservoir pressure is high enough to maintain flow when producing thesereserves (or it is economically justifiable to install artificial lift);
(iii) production system can process the extra production.
These minimum criteria are equivalent to those used for matrix treatments and aresummarised in table 1. There is, however, one extra, unique requirement for proppedhydraulic fracturing:
Parameter Oil Reservoir Gas Reservoir
Hydrocarbon Saturation >40% >50%
Water Cut <30% <200 bbls/MMscf**
Permeability † 1-50 mD * 0.01-10 mD
Reservoir Pressure <70% depleted twice abandonment pressure
Gross Reservoir Height >10 m >10 m
Production System 20% spare capacity ø
* Frac and Pacs may be applied to higher pereability formationsø Lower value if several wells are manifolded together† see also criteria for matrix treatments (Table 3 of Chapter 5)** evaluate potential for well killing itself due to liquid loading in the tubing
(iv) professional, experienced personnel are available for treatment design,execution and supervision along together with high quality pumping, mixingand blending equipment.
This latter requirement arises because a propped hydraulic fracturing treatment has acomplexity and difficulty an order of magnitude greater than that associated withmatrix or acid fracturing treatments. This arises because the ability to complete thetreatment to the specified design requires numerous, on-site adjustments during thetreatment. The first hydraulic fracturing treatments in a new area often experienceearly screen out (see section 6.11) resulting in premature stopping of the treatment i.e.the "learning curve" has to be climbed. The steepness of this "learning curve" can beincreased by employing personnel who have gained experience in successful fracturingtreatments in another area.
3. FRACTURE STIMULATED WELL INFLOW PERFORMANCE
The Inflow Performance of a Fracture Stimulated well is controlled by the dimensionlessFracture Conductivity (F
cd):
Fcd =
kf*wk*Lf
Table 1
Minimum hydraulic
fracturing candidate well
selection screening criteria
Department of Petroleum Engineering, Heriot-Watt University 7
6Hydraulic Fracturing6
where:
kf*w = fracture permeability (k
f) * conductive fracture width (w)
≈ ability of the hydraulic FRACTURE to conduct fluid to the wellbore= the fracture conductivity.
k*Lf
= formation permeability (k) * conductive fracture single wing length (Lf)
≈ ability of the FORMATION to deliver fluid to the hydraulic fracture.
These parameters are illustrated in Figure 4. It can be easily visualised that theobjective of the propped hydraulic fracture treatment design process is to ensure thatthe pressure drop down the length of the fracture is low compared to the pressure dropacross the formation. Thus, as much as possible of the well drawdown should be takenacross the reservoir with the pressure drop within the fracture making a negligiblecontribution to the total value of the well drawdown. We will now consider these twofactors in turn:
��������������yyyyyyyyyyyyyy
Cement
Casing
Lf
w
Formation Permeability- k
Fracture Permeability-kf
the fracture conductivity is increased by:
(a) an increased fracture width (w),
(b) an increased proppant permeability (large, more spherical, proppant grainshave a higher permeability) and
(c) minimising the permeability damage to the proppant pack from the fracturingfluid.
Frequently the increased production achieved by carrying out a hydraulic fracturingtreatment is represented by the "Folds of Increase" or FOI:
FOI = Qf/Q
o
where Qo = k.h (Pe - Pwf)
µ.Bo.In(re/rw)
and Qf = k.h (Pe - Pwf)
µ.Bo.In(re/r'w)
Figure 4
Factors contributions to the
dimensionless fracture
conductivity
8
Qo and Q
f are the well production under equivalent conditions before and after carrying
out the hydraulic fracturing treatment. Thus:
FOI = In(re/r'w)In(re/rw)
There have been several studies of the composite effect of fracture length, fractureconductivity and formation permeability on the well inflow performance (see Cinco-Ley, SPE 10043, 1982 for a review). A widely used correlation is that published byCinco-Ley and Samaniego (JPT, 1981, 1749-1766) in which {r'
w/L
f}, or effective
wellbore radius divided by conductive fracture length, is plotted against thedimensionless fracture conductivity (F
CD). This is illustrated in Figure 5. This Figure
shows that a FCD value of 15 is required to ensure that the well inflow is not beinglimited by the fracture conductivity. An alternative presentation (Figure 6) allows the(negative) skin effect due to the propped hydraulic fracture (S
f) to be calculated from
the dimensionless fracture conductivity.
.01.1
.1
1
1
10 100
Dimensionless Fracture Conductivity: Fcd = kfw / (kLf)
r’ w
/ L f
Maximum value for Fcd > 15
Figure 5
Cinco-Ley and Samaniego's
1981 correlation between
effective wellbore radius
and fracture conductivity
Department of Petroleum Engineering, Heriot-Watt University 9
6Hydraulic Fracturing6
Sf +
In (
xf /
rw
)
FCD
Constant value for Fcd > 15
0.1
1
1 10 100 1000
2
0
Dimensionless Fracture Conductivity: Fcd = kfw / (kLf)
The above correlations and equations can be used to quantify the relationship betweenthe increased production (FOI) as a function of the fracture length (L
f), formation
permeability (k) and the fracture conductivity (kf*w) - see Figure 7. Figure 7 shows
that for wells in low permeability (0.1mD) formations:
100
1000
10000
Fracture Conductivity(kf.w, mD.ft)
Fracture Half Length (Lf, ft)
Fol
ds In
crea
se in
Pro
duct
ion
FormationPermeability
FormationPermeability
= 0.1mD
= 10 mD
0
0
2
4
6
8
10
200 400 600 800 1000
(i) high values of the FOI are possible;
(ii) FOI is related to fracture half length, while the fracture conductivity has alimited effect, providing its value is greater than a certain minimum.
Figure 6
Fracture skin effect varies
with fracture conductivity
Figure 7
Well productivity response
to hydraulic fracturing
10
The (low) formation permeability is controlling the well inflow and increased fractureconductivity does not improve well performance.
An increase in the formation permeability to 10 mD results in a different picture:
(i) a fracture with a low conductivity (100 mD.ft) has essentially no effect on thewell production;
(ii) increasing the fracture conductivity by a factor 10 (to 1,000 mD.ft) increases thewell production (or FOI); but the FOI is still independent of fracture length forvalues greater than 100 ft;
(iii) a further increase to 10,000 mD.ft is required before the inflow performancebecomes sensitive to created fracture length i.e. the fracture conductivity is nolonger the only limiting factor in well inflow.
Inflow from the formation into the fracture is no longer the controlling factor for thishigher permeability reservoir. The above considerations will control the hydraulicfracture treatment design process since long and highly conductive fractures are moredifficult to make and have a greater cost.
4. A PROPPED HYDRAULIC FRACTURING TREATMENT
The major steps needed to carry out a propped, hydraulic fracture treatment aresummarised as follows:
(i) Pumping the fracturing fluid at a sufficiently high pressure to overcome therock stresses i.e. initiate and propagate a fracture.
(ii) The fluid properties are adjusted to ensure efficient fracture creation - low fluidloss and tubing head pressure values are frequently achieved by use of aviscous, shear thinning, water based, cross-linked gel.
(iii) The created fracture is then filled with proppant to "hold it open" or provideconductivity for fluid flow when fluid pumping is halted.
(iv) The viscous fracturing fluid is degraded after the treatment to a viscositysimilar to that of water by incorporation of a chemical breaker into thefracturing fluid formulation. This will allow it to be produced back after thetreatment, followed by the initiation of hydrocarbon production.
The surface and well set up required to achieve the above is schematically illustratedin Figure 8.
Department of Petroleum Engineering, Heriot-Watt University 11
6Hydraulic Fracturing6
Proppant (Quality Control)
ProppantProppant Transport and Conductivity)
Blender
Viscous Fracturing Fluid
(Quality Control)
Pressure casing/Tubing Annulus to Reduce
Tubing Burst Stress
Pump (Pressure Rating) Spare Pump Units Available in Case of Breakdown
Fracture (Fracture
Length and Width)
Pay Zone
Fracturing Fluid(Viscosity Degradation After Completion of Treatment)
(Fracture Containment ?)
Packer(Unseating Forces)
Casing (Integrity)
Tubing (Tensile Strength
Burst Pressure, condition?)
Wellhead (Pressure Rating,Tree Saver Required?
If Yes , Erosion?)
Figure 8 shows the (viscous) fracturing fluid being combined in a blender withproppant (e.g. sand grains) used to keep (or prop) the fracture open once the treatmenthas been completed. The quality of both the fluid and the proppant need to besubjected to proper quality control measures.
The proppant/fluid slurry is then passed to a high pressure pump where the fluidpressure is increased to a value that a hydraulic fracture can be created in the pay zone.Hydraulic fracturing has to be done as a continuous process - spare pumps have to beIMMEDIATELY available if pump breakdowns are experienced. The well head musthave a sufficiently high pressure rating. This pressure rating of the wellhead can betemporarily increased by the installation of a "tree saver" at the wellhead. This isessentially a length of smaller diameter, thick walled tubing installed inside thechristmas tree. It has seals installed at the top and bottom to ensure that the wellheadcomponents with a lower pressure rating are protected from the high pressuresexperienced during the hydraulic fracturing treatment. The tree saver's smallerdiameter leads to increased pressure losses and to the possibility of erosion of thetubing if the proppant slurry exits the tree saver at too high a velocity.
The production tubing will be subject to burst forces due to the high pressures requiredfor fracturing. They can be reduced somewhat by pressurising the casing/tubingannulus. The tubing will also contract due to the pumping of the cold fracturing fluids.This may lead to tensile failure (tubing parting) or the packer unseating. In practice,the design value for the tubing strength should be reduced appropriately to allow forany corrosion if the well has been on production for a number of years. Fieldexperience indicates that it is often impossible to mechanically carry out a hydraulicfracturing treatment in a well unless this was included in the original well designspecifications.
Figure 8
Practical issues during a
propped hydraulic
fracturing treatment
12
The created length of the fracture is considerably longer than the propped length sincethe fracturing treatment is still in progress. Issues to be evaluated during the designinclude:
(i) transport of the proppant to the fracture tip;
(ii) settling of proppant due to inadequate fracturing fluid viscosity;
(iii) creation of the required proppant pack width and degradation of the fracturingfluid to minimise permeability damage to the proppant pack and formation and;
(iv) containment of the hydraulic fracture to the pay zone.
These will all be discussed in the following sections.
5. TYPES WELL FRACTURING STIMULATION TREATMENTS
Some of the main variants in fracturing technology are depicted for vertical wells inFigure 9. Their areas of application have been summarised in table 2.
Reservoir Type TreatmentCandidate Well
Skin (S)
Permeability (k)
Sandstone orCarbonate(Figure 9 a)
Naturally FracturedReservoirs filled with
Calcite Cement(Figure 9 b)
InhomogenousCarbonates(Figure 9 c)
Homogenous Carbonates(Figure 9 d)
Conventional Proped Fracture
Skinfrac*
Matrix Acid
None
HCl Acid pumpednear or above Fracture
Propogation Pressure (FPP)
HCl Acid pumpedabove FPP
WISPER+- Pump VisiousPad follow by HCl Acid.
BOTH above FPP
CFA∅- Pump Visious Padabove FPP, follow byHCl Acid just BELOW
FPP.
High / Low
High / Low
High / Low
High / Low
High
High High
HighVery Low
Low
Low
Low
Low
Low(Matrix)
Low
Medium
* Skinfrac - creation of a short, highly conductive fracture for treating high permeabilty formations+ WISPER - Wide Spaced Etched Ridges∅ CFA - Closed Fracture Acidising
Table 2
Treatment selection
guidelines
Department of Petroleum Engineering, Heriot-Watt University 13
6Hydraulic Fracturing6
hProppant
(settling due to insufficient viscosity)Hydraulic Fracture
Containment?
Sand LadenFluid
(i) Figure 9a illustrates a propped hydraulic fracture treatment. Propped hydraulicfracturing is applicable to both sandstone and carbonate formations. Proppanttransport and hydraulic fracture containment within the pay zone are key issuesto be addressed during the treatment design. The proppant will form therequired highly conductive channel from at depth in the reservoir to thewellbore. Skin fracing (the creation of short but highly conductive fractures)is applied to medium permeability reservoirs (k >100 mD).
Prior to Acidisation(Limited fracture conductivity)
After Acidisation(Acidised channel is highly
conductive)
Calcitecement
filling natural
fracture
Calcite cementednatural fracture
Low matrix permeabilty
Limitedopenchannels,probablynot connected
Width ofnatural fractureincreasedfor carbonate formations
Open channel created by acid dissolvingcalcite cement
Acid insolubleresidue
Figure 9b illustrates the increased conductivity achieved by pumping hydrochloricacid into a naturally fractured carbonate formation. see also chapter 5.11.3
Figure 9a
Propped hydraulic
fracturing for sandstones
and carbonates
Figure 9b
Acid treatment of a low
permeability fomation with
natural fractures filled with
a calcite cementing material
14
Lf
h wHCl Acid EtchedFracture Walls
deeper channel etchedwhere rock is more reactive
Figure 9c illustrates an (HCl) acid fracture treatment of an inhomogeneous, carbonateformation. These naturally occurring inhomogeneities will ensure that some parts ofrock will react more quickly with the acid than others - resulting in a deeper etchingof the fracture wall at this point. Providing the formation is strong and inhomogeneousenough, the (deeper) etched channels will remain open after the treatment is finished,forming a conductive flow path to the wellbore.
N.B. the carbonate acid formulation selection guidelines (chapter 5 on "Acidising"and other Matrix treatments, section 5.10,1) should be followed for both (9b) and (9c);
h w
Lf
ViscousPreflush
HCl AcidFingers
}
Widely spacedof Groups ofPerforations
CFA (Closed Fracture Acidising) gives the similar result,and allows higher perforation density
Figure 9d illustrates the results of a treatment designed to create a number of separate,acid etched fingers. Soft, homogeneous carbonates (e.g. chalks) require the artificialcreation of the necessary inhomogeneities. This is done either through:
(i) the low viscosity acid fingering through the high viscosity gel preflush atpressures greater that the Fracture Propagation Pressure (FPP) (the WISPER orWidely SPaced acid Etched Ridges process) or
(ii) by pumping the acid at just below the FPP (CFA or Closed Fracture Acidisingprocess). The special perforating procedure (widely spaced groups ofperforations) illustrated for WISPER is not required for CFA.
Figure 9d
WISPER (Wide Spread
Etched Ridges) process for
acidising homogenous
chalks
Figure 9c
Acid fracturing for
inhomogenous carbonates
Department of Petroleum Engineering, Heriot-Watt University 15
6Hydraulic Fracturing6
Table 2 shows that, with the exception of "skinfrac", all the fracturing basedtechnologies deliver an increased well inflow both by bypassing any near wellboreformation damage (skin) and increasing the effective wellbore radius. "Skinfrac" typetreatments, as their name implies, are intended to bypass skin due to formationdamage. The increased well inflow due to an increased effective wellbore radius is notso important because they are carried out in formations with a higher formationpermeability.
6. A BRIEF REVIEW OF ROCK MECHANICAL ISSUES RELATED TOHYDRAULIC FRACTURING
6.1 Insitu StressIt is well known (see Rock Mechanics appendix of the Petroleum Geoscience Module)that there are three principle earth stresses oriented at right angles to one another(Figure 10). Below about 500 m, in a relaxed tectonic environment, the vertical stress(σ
v) is normally the greatest. It can be quantified by integrating the density log from
the point of measurement to the surface. An average value of 1.0 to 1.1 psi/ft ismeasured for wells at reasonable depth - though lower values are encountered inshallow, particularly offshore environments subject to rapid deposition.
��������
��������
��������
��������yyyyyyyy
zzzzzzzz
{{{{{{{{
||||||||��������
����������������
��������yyyyyyyy
zzzzzzzz{{{{{{{{
||||||||��������
��������
��������
��������yyyyyyyy
zzzzzzzz
{{{{{{{{
||||||||��������
����������������yyyyyyyy
zzzzzzzz{{{{||||�y�z�z�y
��y{��y{��z|��z|
Overburden Stress (σv)
Minimum Horizontal Stress (σh)
Maximum Horizontal Stress (σH)
Typically σv > σH >σhσH > σv >σh orσH > σh >σv
at shallow depths (<500m)
Figure 10
Insitu stresses in the
subsurface
16
The other two stresses - the maximum and minimum horizontal stresses are orientedat right angles to each other and at right angles to the vertical stress. The maximumhorizontal stress is also referred to as the intermediate stress.
N.B. In tectonically active areas e.g. in the foothills of mountain chains wheremountain building is occurring; the maximum insitu stress may no longer be vertical.The orientation between the three insitu stresses remains the same - hence theminimum and intermediate stresses are also no longer horizontal.
Propagation of a Hydraulic Fracture, involving the actual splitting apart the fabric ofthe formation. The split will propogate in the direction of least resistance i.e. will occurmost easily in the direction perpendicular to the minimum insitu stress. Thus, intectonically relaxed environments, we can assume that a hydraulic fracture will havea vertical orientation and will grow (propagate) in the direction intermediate (ormaximum horizontal) of the insitu stress {at right angles to the minimum insitustress}.
As discussed earlier, the vertical stress (σv) can be measured or assumed with
reasonable accuracy. The important rock property for predicting the other two stressesfrom the vertical stress is called Poisson's Ratio (v), the ratio between Lateral Strain(ε
y) and the Longitudinal Strain (ε
x)
i.e. v = = εyεx
Lateral StrainLongitudinal Strain
{the negative sign is included because (convention) states expansion should be treatedas negative and we wish Poisson's Ratio to be a positive number}.
Figure 11 explains how subjecting a rock sample to a vertical (the overburden) stressresults in it shortening in the vertical direction and expansion in the horizontaldirection. A similar effect occurs in a reservoir rock deposited in a sedimentary basin.The magnitude of the vertical stress at any depth is related to the weight of theoverlying rock mass. This can often be estimated by integrating the density log fromthe depth under study to the surface (a default value of 1.0 - 1.1 psi/ft can be used ifthis log is not available).
Department of Petroleum Engineering, Heriot-Watt University 17
6Hydraulic Fracturing6
εx = dxl
εy = dy
d
l
d
d
Deformed GeometryDue to Overburden Stress
OriginalUndeformed
Geometry
Overburden Stress (σv)
dy /2
dx
Longitudinal Strain:
Lateral Strain:
If we assume that the horizontal stress depends only on the elastic behaviour of therock, the overburden load can be related to the horizontal stresses via the abovePoisson effect together with assumptions about the lateral boundary conditions. In atectonically relaxed area, the two horizontal stress components will have the samevalue within a specific lithology:
σh = σH = v1−v
σv ≈ σv if v = 0.25 13
Variations in lithology, and hence variations in Poisson's Ratio, thus lead to abruptchanges in horizontal stresses with depth.
The above is a simplified picture of the behaviour of rock formations but it does allowthe derivation of a first estimate of the potential horizontal stress changes betweenlayers. Figure 12 illustrates the resulting stresses from deposition of a formation withconstant lithology. Here tectonic forces resulting from movement of the earth's crusthave induced a stress component (σ
tec) so that the two horizontal components are no
longer equal.
Figure 11
Measurement of Rock
properties:
Poissons's Ratio (v)
18
Insitu Stress Magnitude
Dep
th
(
d crit
)
Original Level Ground Surface
Current Ground Level
σH > σh >σv
σH > σv > σh
σv > σH > σh
HorizontalInducedFractures
VerticalInducedFracture
σtec
Overburden Stress Gradientσv = ∫ρ.g.h ≈ 1 - 1.1 psi/ft
σH, max= σh + σtec
σh= υ σv ( ≈ 0.33 σv ) 1−υ
Original Stress
Today's Stress after erosion lowers ground level
Surface elevations change over geologic time. Figure 12 shows how surface erosionhas changed the vertical stress profile while the horizontal stresses have remained"locked-in" due to, for example, inelastic rock deformation. This explains why thevertical stress is normally not the largest one at shallow depths. Other geologicalprocesses, such as burial and uplift, also lead to similar anomalous stress patterns.
The presence of a minimum horizontal stress implies that there will be a preferreddirection of fracture propagation. The fracture will grow in "easiest" direction(requires the least amount of energy) i.e. as a vertical fracture perpendicular to theminimum horizontal stress below the critical depth (d
crit in Figure 12) and as a
horizontal fracture at shallower depths.
6.1.1 Effective StressesThe pore fluids present within the rock matrix will support a proportion of the totalapplied stress. This means that effective stress (σ') carried by the rock matrix grainsis smaller than the total stress. This was quantified by Terzaghi as:
σ' = σ − P
where σ is the total stress, p is the pore pressure and σ' the effective stress which willgovern the failure of the material. This is illustrated in figure 13.
Figure 12
Magnitude of insitu stress
Department of Petroleum Engineering, Heriot-Watt University 19
6Hydraulic Fracturing6
Figure 13
Effective stress acting on a
propped fracture.
Stress
Insitu Stress
Pores Insitu
��yy��yy��yy
�y ��yy�y��yy��yy����yyyy�y��
��yyyy����yyyy��yy����yyyy��yy��yy
��yy��yy��yy��yy�y
PorePressure
PorePressure
PorePressure
Proppant orGrains�y ��yy
��yy��yy��yy�y�y
It was later recognised by Handin that the intergrain cementation does not allow thepore pressure to completely counteract the applied load. A correction factor, the poro-elastic constant α, was introduced:
σ' = σ α − P
where α can vary between 0 and 1 but has a typical value of 0.7 for petroleumreservoirs.
One important conclusion from these equations is that the values of the stresses whichcontrol fracture propagation can change as the reservoir pressure depletes during thelife of a petroleum reservoir. Hence the stress profile measured early in a field'slifetime may become invalid as the field matures. E.g. A hydraulic fracture createdlater in the life of a field will tend to be more confined to the pay zone than a similartreatment carried out early in field life. This occurs because the pay zone reservoirpressure will have decreased due to oil or gas production, while the pressure and hencethe insitu stress in the bounding shale will be unchanged.
6.1.2 Fracture Initiation and Perforation ProgrammeAs discussed, the induced hydraulic fracture propagates at right angles to theminimum insitu stress. From a conceptual point of view, it can be seen that the:
Fracture Propagation Pressure (FPP) ≈minimum rock effective stress + pore pressure + fracture toughness
since the propagating fracture has to overcome the forces that are preventing furthersplitting of the rock i.e. the minimum rock insitu stress and its fracture toughness (thecohesion between the formation grains which needs to be overcome to allow thefracture to propagate).
The Fracture Initiation Pressure (FIP) i.e. the pressure needed to start the fracturepropagating from the perforation will normally be greater (see section 6.6.4) then theFPP. This is because fracture initiation requires additional energy to overcome thetensile stresses present around the borehole plus any extra pressure required toovercome the fact that the perforation is not oriented in the preferred direction for
20
fracture propagation. Such a case of inefficient perforating leading to an increase inthe FIP is illustrated in Figure 14. This illustrates how it is unlikely that inline (00
phasing) perforations will be aligned with this preferred direction of fracture propagation(In the case illustrated the perforation is oriented at right angles to the preferredfracture criteria).
σh σH
00 Phasing (or in line) perforations
600 Phasing
Perforation Tunnel
Maximum Width
Liner
Cement
Restriction
Restriction
Hydraulic Fracture
Formation
Figure 14 shows how the fracture has to initiate from the perforation and then travelaround the well until it achieves the preferred orientation, after which it will propagateaway from the well. The width of the induced fracture is related to the differencebetween the fluid pressure in the fracture and the insitu stress. This means that, in thiscase, the induced fracture will be much narrower at the point of initiation since themaximum rather than the minimum, horizontal stress is acting on the fracture. Thisarea of restricted width may not be wide enough to allow the passage of proppantduring the later stages of the fracturing treatment - resulting in a premature screen out.(The relative importance of this restriction will depend on the difference between thetwo horizontal stresses).
Field experience (backed up by simple geometric considerations) indicates that theFIP can be minimised by perforating the well with 600 phasing - the maximumdifference between the perforation and induced fracture orientation will now bereduced to 300.
Figure 14
Fracture initiation
Department of Petroleum Engineering, Heriot-Watt University 21
6Hydraulic Fracturing6
6.1.3 Data GatheringPrior to carrying out the minimum hydraulic fracturing treatment it is advisable tocarry out a smaller data gathering fracture treatment to measure the formation andfluid properties. Depending on the volumes pumped, this is called a microfrac (<5 m3)or a mini frac (<50 m3). Figure 15 illustrates the set up for such a treatment and someof the important pressure reference points. The fluid is pumped at a constant rate forthe required time and the treatment pressure measured. Bottom Hole Pressuremeasurements are much more accurate than surface measurements, since the hydrostatichead and friction pressure drop components no longer have to be estimated.
Ground Surface
Tubing head treatment pressure (THTP)
Fluid Storage Tanks
High Pressure Pumps
X-Mas Tree
SC-SSSV
Minimum in-situ stress acting on fracture walls
Packer
Downhole pressure guage
Tubulars∆Pfriction∆Phead
∆Pperforation ∆Pfracture propagation
∆P fracture friction
Bottom hole treatingpressure(BHTP)
N.B. Proppant is not used during these treatments. Also only a short interval is oftenperforated.
An example pressure/time recording is shown in Figure 16 and is described as follows:
Figure 15
Arrangement for a mini or
microfrac treatment
22
FractureInitiationPressure (FIP)
FractureReopeningPressure (FRP)
Fracture PropogationPressure (FPP)
Instantaneous Shut - In Pressure (ISIP)
FractureClosurePressure (FCP)
˘P Propagation
˘PPerforation
˘PFracture Friction
+
Pump Cycle 1 Pump Cycle 2
Pre
ssur
e
PumpRate
Time
0
(i) The bottom hole pressure begins to rise as soon as the pumps are started. Thisincrease continues until the Fracture Initiation Pressure (FIP) is reached, afterwhich it drops rapidly to the Fracture Propagation Pressure (FPP).
(ii) The pumps are stopped when the desired fluid volume has been pumped andfracture propagation ceases. This pressure drops rapidly to the InstantaneousShut-In Pressure (ISIP):
ISIP = FPP - ∆Pperforations
- ∆Pfracture
friction
Where ∆Pperforation
= Pressure drop across the perforations and∆P
fracture friction= Frictional pressure drop from the perforations to the tip of the fracture
(iii) The fracture is still open at the ISIP. Leak off continues at a high rate from theopen fracture. The pressure drops until the Fracture Closure Pressure (FCP), equalto the minimum insitu stress (σ
h), is reached.
ISIP = FCP + ∆Pfracture propagation
Where ∆Pfracture propagation
is the pressure required to overcome the fracture toughness
The FCP is recognised as a change in slope in the pressure decay curve. Fluid isleaking off into the formation from the whole fracture surface when the fracture isopen.The fluid loss rate decreases to a low value after the fracture closes, since the areaavailable for fluid loss (perforations rather than an open fracture) is so much smaller.
N.B. If observed long enough, the pressure will eventually equalise at the reservoirpressure. This may take a long time due to the low reservoir permeability.
Figure 16
Bottom hole pressure
recorded during micro or
mini frac
Department of Petroleum Engineering, Heriot-Watt University 23
6Hydraulic Fracturing6
(iv) Reopening of the fracture during a second pump cycle will normally occur ata lower value than the FIP - often the pressure “hump” does not occur andthe FPP is observed immediately.
Table 3 records typical values for the various parameters discussed above for a mini/microfrac treatment in a well perforated at 6000 ft. These values are specific to the wellto be treated. They are required when designing the full scale, hydraulic fracturingtreatment. In addition one can derive further, valuable information.
Tubulars psi
THTP 6400∆Pfriction -6200∆Phead +4500BHTP 4700
Fracture
∆Pperforation -100∆Pfriction -100∆Ppropagation -300
Minimum in-situ stress 4200
Example
(i) The longer the fracture takes to close after the cessation of pumping, the lowerthe leak off coefficient and the greater the fracture volume created (higherfracture fluid efficiency). A volume balance can be performed to quantify thisfluid loss coefficient which can then be used as input to fracture treatmentdesign programs.
(ii) The fracture height can be determined using a temperature log. A temperaturelog is run across the perforated interval before and immediately after completion of the minifrac treatment. This production log, which incorporates a highresolution thermometer, will record a (cool/lower temperature) zoneacross the created fracture due to the injection of the cold fracturing fluid.Figure 17 is a schematic example of such a temperature log, the fracture beingobserved to have grown upwards from the perforated interval.
Table 3
Typical pressure values for
a mini/micro frac treatment
at 6,000ft
24
Geotherm
al Gradient (recorded
beforem
ini frac)
PerforatedInterval
Created Fracture Height
Zone of cooling recordedimmediately aftermini frac
Temperature Log
Before mini frac
After mini frac
Temperature
Dep
th
Upward Growthof Fracture
6.2 Fracture SizeGreater volumes of fracturing fluid will create larger fractures - with higher treatmentcosts but also potentially more productive. However, often uncontrolled growth offractures is not desirable from a production point of view e.g. when the target oil zoneis overlain by gas with water underneath. Figure 18 shows how the maximum fracturesize is limited for this situation. It assumes that:
������������
yyyyyyyyyyyy
��������yyyyyyyy
���������
yyyyyyyyy
��������yyyyyyyy
Initial fracture geometry Maximum size fracture allowableassuming radial propagation
GAS
OIL
WATER
Lf
(i) the fracture is initiated from perforations at the mid point of the oil zone.
(ii) the fracture propagates radially (in practice, this implies that the formation ishomogenous with a stress gradient equal to the hydrostatic head of thefracturing fluid - see section 6.6).
The resulting maximum allowable fracture half length (Lf) is slightly less than half the
height of the oil column.
Figure 18
Fracture size limited by
geometry and fluid contacts
Figure 17
Fracture height
measurement using a
temperature log
Department of Petroleum Engineering, Heriot-Watt University 25
6Hydraulic Fracturing6
6.2.1 Fracture ContainmentThe hydraulic fracture should thus be designed so that it does not contact unwantedfluids within a single formation layer. It must also be consider whether the hydraulicfracture is contained within the pay zone i.e. whether upward and/or downwardfracture growth is retarded by changes in the formation property contrast between thetwo layers. Important formation properties include:
������������������������
yyyyyyyyyyyyyyyyyyyyyyyy
������������yyyyyyyyyyyy������������yyyyyyyyyyyy
Formation Minimum insitu stress (σh)
Pay Zone
Poisson'sRatio (υ)
Shale 0.25
Sand 0.15
Shale 0.25
(i) (Minimum) insitu stress: permeable formations e.g. sand typically have a lowerPoisson’s Ratio compared to the bounding shale layers. The resulting lowerinsitu stress will aid hydraulic fracture containment (see section 6.1 and Figure 19).
(ii) Fracture toughness: increased values of the fracture toughness imply that it ismore difficult for the fracture to propagate in that zone.
(iii) Leak off: high fluid loss rates will retard fracture propagation through the zone.
3-D fracture propagation models are available to predict the simultaneous lateral andvertical growth of the hydraulic fracture. These complex numerical codes have beenused to generate the following examples which illustrate some of the important factorsconcerning fracture containment.
Figure 19
Insitu stress contrasts
26
6.2.2 High Insitu Stress Contrasts and Fracture Shapes
Frac length
Cap
Roc
k
(Sal
t)
Pay
Zon
e
(San
d)
Frac length
Cap
Roc
k
(Sal
t)
Pay
Zon
e
(San
d)
Frac length
Cap
Roc
k
(Sal
t)
Pay
Zon
e
(San
d)
Perforations
Frac length
Cap
Roc
k
(Sal
t)
Pay
Zon
e
(San
d)
10 MPa 10 MPa
10 MPa 10 MPa
N.B. Fracture shape indicated by shading, volume fracture fluid pumped is the same in all cases
(i) Constant stress in pay zone (ii) Hydrostatic Stress Gradient (0.45 psi/ft)in pay zone
(iii) Extensional Stress Gradient (0.7 psi/ft) in pay zone
(ii) Overburden (or Maximum) Stress Gradient (1.1 psi/ft) in pay zone
Perforations Perforations
Perforations
Minimum insituStress
Minimum insituStress
Minimum insituStress
Minimum insituStress
Dep
th
Dep
th
Dep
th
Dep
th
Figure 20 shows a case where a massive, homogeneous, gas bearing sandstone isoverlain by a (sodium chloride) salt zone (the cap rock). The well is perforated nearthe top of the pay zone. Due to the plastic nature of salt:
σh = σ
H = σ
V
There is thus a very large insitu stress contrast (estimated as 10 MPa) at the salt/sandstone boundary resulting in upward fracture growth being immediately halted.Due to the homogeneous nature of the sandstone, it is expected that there is a constantstress gradient. The effect on fracture shape and containment of differing values ofthe stress gradient is illustrated in figure 20
(i) constant stress (or zero stress gradient): the fracture grows downwards due tothe density of the fracturing fluid giving rise to an increased pressure at thebottom surface of the fracture compared to the top surface.
(ii) hydrostatic stress gradient (0.45 psi/ft): the stress gradient in the formation isnow the same as that of the fracturing fluid - resulting in radial propagation of thefracture, apart from the top surface where upward growth is stopped due to the large(10 MPa) stress contrast at the salt/sandstone boundary.
(iii) extensional stress gradient (0.7 psi/ft) is commonly found in relaxed tectonicenvironments. A longer fracture results since the fracture downward growthbecoming limited due to the minimum insitu stress with increasing at a greaterrate (0.7 psi/ft) than the hydrostatic head of the fracturing fluid (0.45 psi/ft)
.
Figure 20
Effect of stress gradient on
a fracture shape and
containment
Department of Petroleum Engineering, Heriot-Watt University 27
6Hydraulic Fracturing6
(iv) overburden (or maximum) stress gradient (1.1 psi/ft) . This scenario yields thelongest fracture with the least downward growth. In fact, the fracture is tryingto grow upwards, the reverse of scenario (i), but is constrained by the high stresscontrast of the salt/sand boundary.
6.2.3 Fracture Growth into BoundariesWhether a pay zone boundary is capable of containing a fracture will depend on themagnitude of the fracture containment mechanism e.g. minimum insitu stress contrastand the thickness of the boundary. Figure 21 schematically illustrates fracturecontainment for 3 different values of the stress contrast. Initially the fracturepropagates radially in the pay zone until the boundary layer is reached; after which isbecomes more elongated - with greater stress contrasts giving rise to the moreelongated shapes.
N.B. The consecutive lines growing from the left hand side refer to the fracture shapeat increasing times/volume of hydraulic fracturing fluid pumped.
0 120 240 360
500 psiStress Contrast
Formation B
5,590'
6,000'
6,050'Formation A
200 psiStress Contrast
50 psiStress Contrast
0 120 240 360
Formation B
5,590'
6,000'
6,050'Formation A
0 120 240 360
Formation B
5,590'
6,000'
6,050'Formation A
Dep
thD
epth
Dep
th
Fracture length
Fracture length
Fracture length
Upward fracture growth stopped at formation boundary
Limited upward fracture growth at formation boundary
Almost unimpeded upward fracture growth
Figure 21
Vertical growth of fracture
for various stress contrasts
28
Figure 22 schematically illustrates what occurs when fracture containment is nolonger effective due to the height of the upper barrier and the available insitu stresscontrast being insufficient to prevent the fracture breaking through into a shallower(water bearing) zone.
(a) The fracture is initiated at the centre of the pay zone (time 1) and initially growsradially (time 2 and 3) since a hydrostatic stress gradient is present. This resultsin the FPP decreasing as the fracture becomes longer.
(b) The stress contrast at the upper and lower boundary cause upward anddownward fracture growth to be retarded. Upward fracture growth is somewhat easier as the stress contrast in this direction is less. The effect of this partialfracture containment is to increase the FPP (times 4, 5 and 6).
7
6
5
4
43
2
15 6
7
321
∆σh
Water Zone(Sand)
Barrier(Shale)
Pay Zone(Sand)
Barrier(Shale)
Fra
ctur
e P
ropa
gatio
n P
ress
ure
Time
Time 7 - Pressure drop shows fracture breakthough into the low stress water sand
Dep
th
Insitu Stress
Figure 22
Fracture containment no
longer achieved
Department of Petroleum Engineering, Heriot-Watt University 29
6Hydraulic Fracturing6
(c) At time 7 the upward fracture growth allows the fracture to break through intothe upper water bearing zone. Since this has a very low, constant fracturegradient, the fracture will grow rapidly upwards, resulting in a drop in the FPP.
Note that the fracture length in the payzone has actually decreased due to thefracture's rapid expansion into the upper, water sand. Real time measurement of theFPP thus allows monitoring of fracture containment.
6.2.4 Fracture Height MeasurementKnowledge of the fracture height is clearly important when designing and executinga hydraulic fracturing treatment. There are a number of possible measurementtechniques which can be used to measure this:
(i) run a temperature log immediately after the fracture treatment to measurecooled zone denoting fracture fluid entry {see figure 7}.
(ii) the depth at which fluid is entering into the well from the fracture can bemeasured by running a production log across the perforated interval to measure theflow profile (spinner a flow meter log) or the flow induced noise (noise log) ortemperature changes.
(iii) the proppant can be given a lightly radioactive coating. Running a gamma raylog after the excess proppant has been cleaned out of the well will measure thepropped fracture height.
(iv) the fracture can be physically observed in open hole completions using :
(a) a formation Microscanner (a resistivity log with many, closely spaced,measurement pads which can differentiate between the fracture and formation) or:
(b) a borehole camera (these cameras generate a picture of the boreholewall using video or acoustic signals).
(v) making passive seismic measurements. This involves triangulation ofseismic events emitted from the propagating fracture tip. These seismicevents are measured with geophones installed at the surface or in the well.
(vi) use of a tiltmeter at the surface (for land wells only). This very sensitive devicemeasures (very small) changes in the surface topography due to propagation ofthe hydraulic fracture. These changes indicate the length and orientation of thehydraulic fracture.
It is apparent that each of the above techniques {apart from (v)} makes measurementsat the wellbore only. Further, they do not measure the same fracture attribute sincethe created {(i) and (v)}, propped {iii)} and permeable {(i), (ii) and (iii)} heights, evenat the wellbore, are not necessarily the same.
30
6.2.5 Fracture Azimuth and Well Deviation and OrientationEarlier discussions (6.6.1) discussed how the induced hydraulic fracture wouldorientate itself at right angles to the direction of the minimum insitu stress. Thisdiscussion implicitly assumed that the well was vertical. A more complicatedsituation arises when the well is deviated from the vertical - since the fracture and wellazimuth are unlikely to be orientated in the same direction.
σhσHσhσH
(a) Well deviation in preferred fracture plane.,Good communication
(b) Well deviation not in preferred fracture plane.,Poor communication
(i) Figure 23a illustrates the case when the well was deviated in the plane of thefracture. Good fracture/well communication results once the well has beenefficiently perforated {Section 6.6.1.2 and figure 14}.
(ii) Figure 23b shows how the fracture only has a short length of contact with thewell when the fracture and well orientations are not coincident. This can(potentially) lead to poor fracture/well communications and reduced wellproductivity.
The propped, hydraulic fracturing of deviated wells can require special measures,especially when the two insitu stresses have very different magnitudes. The drillingof S shaped wells which are vertical across the pay zone is one option to be considered.
6.3 Modelling the Shape of the Induced FractureSeveral (commercial) programmes are available to predict the shape (height, lengthand width) of the induced hydraulic fracture. They fall into one of three classes:
(i) 2D: these models use two dimensional, analytical equations where the fractureheight is required to be input.
This class is used extensively in engineering programs, requiring a reduced data setto the others, as well as limited computing requirements.
(ii) P3D: or pseudo three dimensional programs. These combination of analyticaland numerical routines will predict the fracture height and width to varysomewhat independently.
Figure 23
Hydraulic fracturing of
deviated wells
Department of Petroleum Engineering, Heriot-Watt University 31
6Hydraulic Fracturing6
This class is used by fracturing specialists and has moderate computing requirements.
(iii) Fully 3D: these fully three dimensional programs are complex numericalmodelling programs with extensive input data and high end computingrequirements. The fracture height, width, length and shape can all varycompletely independently.
This class of program is normally only found in the research laboratories. Realisticpredictions form any of these programs depends on accurate input data beingavailable. This includes:
(i) Geological data (formation boundaries etc.)
(ii) Rock mechanical properties from sonic and density logs and core measurements (Young’s modulus, Poisson’s Ratio etc.)
The extent of data provision required for specific case studies often limits theapplication of fully 3D programs to more generic studies evaluating the effects ofchanges in a given treatment or formation parameter.
6.3.1 2D Fracture ModelsThe complexity of hydraulic fracturing models derives from the need to simultaneouslysatisfy two sets of laws:
(i) conservation of momentum, mass and energy;
(ii) a fracture propagation criteria that controls the advance of the fracture tip.
32
H
2R
H
Elliptical cross section
Rectangular cross section
Elliptical cross section
Area of largest flow resistance
xf = L/2
xf = L/2
(a) Kristanovic, Geertsmaand De Klerk (KGD)Assumes: H > L
(b) Perkins,Kern and Nodgren(PKN) Assumes: L > H
(c) Radial Model
Assumes: L = H = 2R
R
KeyH (or 2R) is fracture height at wellbore.xf (or L/2 or R) is fracture half length.(A i d f llb )
We will look at the class (i) or 2D models in greater detail:
(i) The simplest solution available is when the fracture height is greater than thefracture length and free slippage occurs at the upper and lower boundary givinga rectangular shape at the wellbore. The fracture shape will not depend on thevertical position in the fracture (see Figure 24a). This model was introducedby Kristianovic, Geerstma and de Klerk (KGD).
(ii) A second situation is when the fracture is confined by boundary layers-it hasan elliptical shape at the wellbore and the length is much greater than the height.
Figure 24
Fracture shapes predicted
by two dimensional
hydraulic fracturing models
Department of Petroleum Engineering, Heriot-Watt University 33
6Hydraulic Fracturing6
Figure 24b sketches this model. It was introduced by Perkins, Kern andNordgren. The FPP decreases as the fracture grows longer.
(iii) A limiting case between the two is the radial one when the fracture height equalsthe fracture length (Figure 24c).
As mentioned earlier, they require a value for the fracture height to be input. Withinthis constraint and when used in the correct application area (i.e. L/H ratio), they giveresults which agree well with the more complex models.
6.3.2 Fracture WidthThe preceding sections introduced the concept of fracture containment. However, tobe of practical use, the created fracture also has to be wide enough to admit proppantthat will provide the permeable flow path from the reservoir to the wellbore after thefracturing treatment has finished. Pumping of the proppant when the fracture is notwide enough will lead to an immediate premature screen out (bridging of the proppantparticles at the mouth of the fracture). The 2D fracture models discussed above areused to predict this width value:
ωmax ≈viscosity*pumprate
Youngs modulus
14
(ISIP -FCP)
where ωmax
is the maximum fracture width at the wellbore.
The quarter power means that changes in the viscosity or treatment pump rates overpractical ranges have limited effect on the fracture width e.g. increasing the viscosityor treatment pump rate by a factor three only increases the width by 32%. On the otherhand, rock mechanical properties of the formation have a much larger impact:
(i) The Young’s modulus can vary by a factor 100, from 1 x 105 psi in softformations (soft chalks, diatomite, coals etc) to 1 x 107 psi in strong sandstonesas found in deep, low permeability formations.
(ii) The fracture toughness - which controls the ease of (or excess pressure requiredfor) fracture propagation and hence the value of the term (ISIP - FCP).
34
7. CREATION OF A PROPPED HYDRAULIC FRACTURE
(a) Pad creates fracture. Fluid loss limits rate of fracture length creation.
(b) Fracture closed on proppant. Well ready for production after fracturing fluid's viscosity degrades
Fracture Leak-off
(b) Inject Proppant fluid slurry. Prop displaced towards fracture tip. Fract length growth continues.
Fracture Leak-off
(c) Stop injection when proppant reaches fracture tip. Fluid leak-off continues.
Fracture Leak-off
Figure 25 summarises the main stages in the process involved in creation of a proppedhydraulic fracture:
(i) An initial fracture of appropriate length and width is created by pumpingfracture fluid called the pad. The most common fracturing fluids are waterbased, cross-linked, polymer solutions (or gels) - which exhibit highly non-Newtonian rheological properties (see figure 40 chapter 7 and figure 15,chapter 1) and appropriate fluid loss properties. This is often preceded by asacrificial pre-pad (a low viscosity fluid which satisfied part of the fluid lossfrom the fracture at a reduced cost). Typically, about 50-80% of the totalfluid volume pumped leaks off to the formation while only 20-50% createsuseful fracture volume.
(ii) Proppant and Gravel Pack Sand are of similar size and the same material maybe used for both applications (see the section 7.5.4. on "Gravel Pack SandSelection"). Proppant particles are added at low concentration to thefracturing fluid once:
(a) the fracture width is sufficient to admit the proppant without causing ascreen out and
(b) the created fracture is nearing its design length.
Figure 25
Creation of a propped
hydraulic facture
Department of Petroleum Engineering, Heriot-Watt University 35
6Hydraulic Fracturing6
Pad Creates Fracture Slurry Stage0
1
2
3
4
5
6
7
8
9
Start Treatment Time
Pro
ppan
t Con
cent
ratio
n (p
pg)
End
Pad1 lb/galproppantconcentration
4 lb/gal 2 lb/galto 3 lb/gal
Pad1 lb/galconcentratedto 4 lb/gal dueto leak off
1 lb/galconcentrationto 8 lb/gal
6 lb/galto 8 lb/gal
Pad4 lb/galto 8 lb/gal
2 lb/galto 8 lb/gal
First proppantstage reachesperforation
Proppant concertration sche dul eincreases proppant concentration in jection into fract ure at perforationin jected Proppant concentration in fract ure increases du ring job due to leak off .
Proppant concentration profi le at end of treat ment
(i)
(ii)
(iii)
Settling of dense proppantcontinues until fracture closure complete.
8 lb/gal
Fracture containing boundary
Proppant settle downwards due to particle density greater than that of fracturing fluid.
Figure 26(a)
Hydraulic fracturing
proppant concentration
schedule during an
hydraulic fracturing
treatment
Figure 26(b)
Proppant profile
development during a
hydraulic fracture
treatment.
36
The concentration of proppant is increased towards the end of the job with proppantconcentrations as high as 40% vol being pumped {(Figure (26a)}. This ensures thata more uniform, final proppant concentration in the fracture is achieved. This occursbecause the low proppant concentration in the slurry pumped initially will becomemore concentrated as it is displaced towards the fracture tip as the fracturing fluid leaksoff into the formation. "Slumping" of the denser proppant will also occur due to theinfluence of gravity. This process will continue until fracture closure is complete.These processes are illustrated in figure 26(b).
(iii) The proppant slurry in the wellbore is displaced to the perforations and fluidinjection halted. This normally occurs at about the same time as the firstproppant reaches the fracture tip. There will be minimal stimulation from thefracturing treatment if the final proppant slurry volumes are overdisplacedaway from the wellbore i.e. the well looses (direct) communication with thehydraulic fracture. However, excessive under displacement of the slurry willleave large amounts of proppant in the wellbore at the end of the treatment. Thiswould then have to be removed by a special cleanout trip made with a coiledtubing or work over unit prior to returning the well to production.
(iv) Leak-off continues and the fracture closes on the proppant. Viscosity degradationfrom the action of the chemical breaker (see section 6.12) added to thefracturing fluid aids in the back production of the degraded fracturing fluidfollowed by hydrocarbon production.
A complex computer program is required for treatment design since, in addition to thecomplexities of fracture shape prediction described earlier, it must also optimise thetransport of the proppant within the fracture.
8. HYDRAULIC FRACTURING IN PRACTICE
The equipment required for a Massive Hydraulic Fracturing (MHF) treatment on landis schematically illustrated in Figure 27.
Department of Petroleum Engineering, Heriot-Watt University 37
6Hydraulic Fracturing6
Base Fluid Tanks
Base Fluid Tanks
Base Fluid Tanks
Base Fluid Tanks
Base Fluid Tanks
Base Fluid Tanks
Base Fluid Tanks
Low P
ressure Pum
p
Base F
luid
Well
Low P
ressure Pum
p
High P
ressure Pum
p
Pum
p
High Pressure Pumps
Proppant Bins
Sand Bins
Suction M
anifold
Gel Tanks
Gel Tanks
LiquidGel Concentrate
Additives e.g.Buffer,
Cross Linker,Breaker
Blender 1
Blender
Discharge Manifold
Fuel
ConveyerConveyerConveyerBlender 2
Blender 2
Production F
low Line
Clean Up Pit
Spare
Spare
Control V
ans
}
(Spare)
(i) The base fracturing fluid (a brine compatible with the formation e.g. 1% wt KClsolution) is placed in a series of clean tanks which are manifolded together.
(ii) A low pressure pump transfers the fluid to blender No 1 where it is mixed witha concentrated polymer gel solution to give a typical polymer concentration of40-100 lbs/1000 gal. Sufficient residence time combined with intense agitationwithin the blender ensure the gel is properly hydrated (dissolved) in the basefluid.
Figure 27
Equipment for a massive
hydraulic fracturing
treatment on land
38
(iii) The base gel is transferred to blender No 2 where additives such as crosslinkingagents, buffer, breaker surfactants etc are added. Quality control checks arecarried out to ensure the (cross-linked) gel has the desired properties i.e. theadditives are being added in the correct amounts and no contaminants arepresent. Proppant is also added in blender No 2.
(iv) The proppant was loaded into bottom discharge bins prior to the commencement of the fracturing treatment. It is transported to the blender using a movingband conveyor. Measurement of the proppant slurry density ensures that thecorrect addition rate of proppant is being achieved.
(v) Blender No 2 transfers the prepared fracturing fluid or slurry to a number ofhigh pressure pumps manifolded in parallel. The pumps must besufficiently powerful to be able to pump to the well itself the fluids/slurries atthe required pressure and rate. A dual, high pressure flow line is oftenprovided between the pumps and the well to minimise frictional pressure losses.
(vi) The instrumented control van allows the “frac master” to monitor the treatmentprogress, to order adjustments to be made as required and to cope withoperational problems e.g. faulty pumps, conveyors etc. as they occur.
(vii) Spare or standby pumps, blenders, conveyors, etc should be hooked up andoperational before the treatment commences. This ensures failures have aminimal impact on the treatment progress.
(viii)A fuel tanker is required to ensure that all the engines driving pumps, blendersetc have sufficient fuel for the treatment duration.
(ix) The proppant slurry is displaced with base gel to just above the perforations andthe pumps stopped. The well is shut in for sufficient time to allow the fractureto close and the fracturing fluid. viscosity to degrade The well may now beplaced on production. Initially only degraded fracturing fluid will be produced- this can be directed to a clean up pit. Once hydrocarbons are being producedthe well production should be directed to the facilities via the production flow line.
The process for fracturing an offshore well is essentially the same as the above,although all the equipment is installed in a specially designed stimulation boat (figure28). Sea water is often used as the base fluid, reducing the fluid storage requirements.
Carrying out a hydraulic fracturing treatment is an expensive, complex undertaking.The costs, volumes of material used and the dimensions for a typical MHF treatmentare summarised in table 4.
Department of Petroleum Engineering, Heriot-Watt University 39
6Hydraulic Fracturing6
Crew's Quarters
and Wheelhouse
Observation Deckand
Control Room
Pumps High Pressure
Pumps High Pressure
Pumps High Pressure
Pumps High Pressure
Pumps High Pressure
Pumps High Pressure
Pumps High Pressure
Pumps High Pressure
ProportioningBlender
Base fluid, gel and additive
feed lines from below deck tanks
Proppant holding tank,main proppant
storage in lower decks
Safety, high pressurepop-off valve
Hyraulic quick disconnect
Flexible high-pressure line to platform
Platform Arrangements
Well
Valve
Flow Meter
RadioactiveDensometer
Check Valve
ElectronicPressure
Transducers
N.B. Heavy loads (base fluids, proppant etc.) stored on the lower boat decks to maximise vessel stability
Figure 28
Layout of an offshore
stimulation vessel
40
Base Fluid
Proppant
Pump Rate
Wellhead Treating Pressure
Cost
Fracture Length
Height
Width
Dimensionless FractureConductivity
Production ImprovementFactor
6000 bbl
1,000,000 lb
30 - 50 bbl / min
10,000 psi
1,000,000 US$ (offshore, UK Southern North Sea)350,000 US$ (land, USA)
500 ft
200 ft
0.5 - 2 inch
3 - 100
2 - 5
9. OPTIMISATION OF HYDRAULIC FRACTURE DIMENSIONS
The expense of carrying out a hydraulic fracturing treatment dictates that the proppedhydraulic fracture dimensions are carefully optimised to generate the maximumdiscounted net present value (NPV) for the stimulation project. Part of this processis illustrated in Figure 29. It consists of two strands, (A) and (B), which are combinedto generate a plot of NPV against fracture length.
A well performance prediction computer program is used to generate a plot ofwell production rate against time for various fracture lengths (L
f) (see Figure29a).
We assume a high {CFD
> 15} dimensionless fracture conductivity at this stage. Asdiscussed in section (6.3) longer fractures will give higher production rates. Thesefigures can be translated into a plot of cumulative production against time, for eachfracture length (Figure 29b), and then further into a plot of discounted revenue againstfracture length (Figure 29c). Longer fractures generate more revenue, but the rate ofrevenue increase decreases as the fracture length increases.
The second strand of the process (B) uses a computer program for hydraulic fracturetreatment design to generate estimates of the fluid volume required for a number offracture lengths (Figure 29d). These volumes may be equated with treatment cost(Figure 29e) - as expected, longer fractures require more volume and are more costlyto produce. Further the rate of cost increase becomes larger as the fracture lengthincreases.
Finally, the data from Figures 29c and 29e are combined in the fracture optimisationstep (c) to give a plot of project net present value. The trends in the later two figuresensure that there is a fracture length which gives the maximum project profitability(Figure 29f).
Table 4
Typical massive hydraulic
fracture treatment statistics
Department of Petroleum Engineering, Heriot-Watt University 41
6Hydraulic Fracturing6
The above only describes part of the treatment optimisation process since the fractureconductivity (a combination of proppant type, size and impairment from residues leftbehind by the fracturing fluid) and fracturing fluid selection all impact on thetreatment cost. Some of the considerations involved in selecting these aspects arediscussed below.
Well PerformanceComputer Program
Hydraulic FractureDesign
Computer Program
A
Fracture ValueOptimisation
C
B
Pro
duct
ion
Rat
e
Time (days)
11
10
100
1000
10 100 1000 10000
Lf = 3000 (ft)
Lf = 500 (ft)
Lf = 0 (ft)
(a)
Fracture Length, Lf (ft)
(c)
0 1000 2000 3000Dis
coun
ted
Rev
enue
Net
Pre
sent
Val
ue (
$)
Trea
tmen
t Vol
ume
Fracture Length, Lf (ft)0 1000 2000 3000
(d)
Fracture Length, Lf (ft)
Trea
tmen
t Cos
t($)
0 1000 2000 3000
(e)
Time (years)
Cum
ulat
ive
Pro
duct
ion
Lf = 3000 (ft)
Lf = 1000 (ft)
Lf = 500 (ft)
Lf = 0 (ft)
(b)
Fracture Length, Lf (ft)
Net
Pre
sent
V
alue
($)
0 1000 2000 3000
(f)
Figure 29
part of the process to
optimise the dimensions of
an hydraulic fracture
42
10. PROPPED FRACTURE CONDUCTIVITY
The proppant placed within the fracture is stressed as the fracturing fluid leaks awayand the fracture closes. This fracture closing stress (FCS) is equal to the minimuminsitu effective stress (σ'
h)where:
FCS = σ'h= σ
h - P
frac
where Pfrac
is the fluid pressure within the fracture.
Sand Grains Soft Proppants Hard Proppants
Crushing Deformation Little Embedment
Sand Grains Soft Proppants Hard Proppants
Crushingand Embedment
Deformation Much Embedmentand Embedment
No ClosurePressure
Closure PressureApplied
No ClosurePressure
Closure PressureApplied
Hard Rock
Soft Rock
Depending on the properties of the proppant and the strength of the formation, figure30 illustrates how this fracture closing stress will result in:
(i) crushing of the proppant grains leading to reduced proppant permeability, andhence reduced fracture conductivity;
(ii) deformation of (soft) proppants which leads to reduced fracture width, andhence reduced fracture conductivity;
(iii) embedment of the proppant in the fracture wall, leading to a further reductionin fracture conductivity.
The size range for commercial proppants are specified in a similar manner as gravelpack sand (section 7.5.4). Figure 30 shows how the resulting fracture conductivity isa result of a combination of the proppant type (quality) and the formation properties.The available proppant types are listed in Table 5 while Figure 31 schematicallyillustrates the variation in permeability as a function of closure stress (steel platenswere substituted for the fracture wall so that negligible proppant embedment occurred).It can be seen that the more well rounded a proppant is the higher its
Figure 30
Behaviour of proppant
under stress
Department of Petroleum Engineering, Heriot-Watt University 43
6Hydraulic Fracturing6
Proppant Type Resistanceto Crushing Cost
Low Quality Sand
High Quality (Ottawa) Sand
Resin Coated Sand
Intermediate Strength Proppant (Ceramic)
High Strength Proppant (Bauxite)
Low
High
Low
High
permeability will be for the same proppant size and the greater its strength since thefracture closing stress will be spread more evenly over the proppant grain's surface.Thus.
(i) the low quality sand {with its multi-crystalline, angular (or not well-rounded)grains} began to crush at low stresses (< 200 psi).
(ii) the high quality sand (well rounded, mono-crystalline grains) shows a muchgreater stress resistance.
(iii) bauxite, a high strength proppant, showed negligible grain crushing and a smallamount of deformation; even at the highest closure stresses (10,000 psi).
NB. It must be remembered that the highest Fracture Closure Stress will be experiencedby the proppant in the fracture when the reservoir is depleted and the well is underproduction i.e. equate P
frac to the minimum expected flowing bottomhole pressure
when calculating the FCS to be used for proppant selection.
High Strength Proppant (e.g Bauxite)
High Quality Sand
Low Quality Sand
Fracture Closing Stress (psi)
% F
ract
ure
Con
duct
ivity
Rem
aini
ng
0
0 2000 4000 6000 8000 10000
20
40
60
80
100
Table 5
Proppant types
Figure 31
Fracture conductivity
variation with stress
44
11 . THE FRACTURING FLUID AND THE FRACTURE CONDUCTIVITY
There is another process that reduces the fracture conductivity. Most modernfracturing fluid consist of a low concentration of a polymer dissolved in a brine. Thedilute polymer solution's viscosity is increased by joining the polymer moleculestogether with a crosslinking agent. During the fracture treatment this dilute polymersolution is pumped into the fracture at a pressure much greater than the reservoirpressure - resulting in a high percentage of the base brine “leaking off” into theformation. The large polymer molecules are too large to be able to flow through thepore throat and hence form an external filter cake on the fracture surface. This isparticularly true for the low permeability formations where fracturing is often applied.The chemical breaker is normally dissolved in the fracturing fluid. There is sufficientchemical breaker present in fracturing fluid within the closed fracture to degrade thefluid viscosity. However, it is not capable of destroying the filter cake, which remainsin the fracture. This further degrades the fracture conductivity (see Figure 32).
The efficiency of the removal of the remnants of the fracturing fluid from the proppantpack itself is measured by the Retained fracture conductivity where:
Retained fracture conductivity = (conductivity after exposure to fracturing fluid x 100%)
conductivity prior exposure to fracturing fluid
(Soft) Formation Fracture Face
(Soft) Formation Fracture Face
Embedment in formation
Filter cake
Fracture closure stress
N.B. Proppant grain crushing and deformation not illustrated (see Fig 30)
Original propped width fracture
Effective fracture width for flow
It can be imagined that the type and concentration of polymer used to prepare thefracturing fluid effects the fracture conductivity. This is illustrated in Figure 33 fora number of different fracturing fluids - retained fracture conductivities of between50% and 100% are observed.N.B. The skin effect due to the filter cake or fracturing fluid - reservoir rockincompatibility has little effect on the reservoir inflow performance providing thepermeability damage is not excessive e.g. (k
damage / k
original < 0.5). This is due to the large
inflow area of the hydraulic fracture surface. The effective fracture width open to flowand the proppant permeability are the key parameters controlling the fracture flowcapacity.
Figure 32
Actual hydraulic fracture
conductivity
Department of Petroleum Engineering, Heriot-Watt University 45
6Hydraulic Fracturing6
CMHPG withAluminium Crosslinker
HPG Solution withTitanium Crosslinker
Emulsion (67% Diesel,33% HEC Solution)
HPG Solution withBorate Crosslinker
Oil Gel
70% vol Nitrogen GasFoam / 30% HEC Gel
0 10 20 30 40 50 60 70 80 90 100
Fra
ctur
ing
Flu
id T
ype
Retained Conductivity (%)
N.B. Proppant Loading - 2 lbs/ft2 Polymer Concentration - 40 lb/1000 gal Ammonium Persulphate Breaker - 2 lb/1000 gal
Key to Polymer Types
HPG - Hydroxy Propyl Guar
HEC - Hydroxy Ethyl Cellulose
CMHPG - Carboxy Methyl Hydroxy Propyl Guar
It can be surmised from Figure 32 that the values measured for this retainedconductivity are dependent on the fracture width or the number of layers of proppant(there are three proppant layers in Figure 32). This is because embedment of theproppant layers in the fracturing fluid filter cake and in the fracture face itself becomesprogressively more important to the final fracture conductivity as the number of layersof proppant grains decreases.
The number of proppant layers can also be expressed as a proppant loading (weightof proppant per unit fracture area). Figure 34 summarises experiments performed atvarying proppant loadings. They clearly show that the filter cake effectively destroysthe fracture conductivity at low proppant concentrations.
Experimental Conditions
Fracturing Fluid
Cross Linker
Chemical Breaker
Proppant Loading
Proppant Type
40lb / 1000 gal HPG
Titanium
2lb / 1000 gal(ammonium persulphate)
0.5 - 2 lb / ft2
Ottawa Sand
0 10 20 30 40 50 60
0.5
1
2
Pro
ppan
t Con
cent
ratio
n (lb
/ sq
ft)
Retained Fracture Conductivity (%)
Figure 33
Typical values for fluid type
and retained fracture
conductivity
Figure 34
Typical values for proppant
loading and retained
fracture conductivity
46
The chemical breaker added to degrade the fluid viscosity is capable of partiallydestroying the filter cake. This is made clearer in Figure 35 where an increase in thebreaker concentration, in this case ammonium persulphate, results in greatly increasedretained fracture conductivity. However, addition of large concentrations of breakerto the fracturing fluid is not a viable approach since this will result in the fluid viscositybeing degraded during, rather than after, the hydraulic fracturing treatment. Thisearly decrease in fracturing fluid viscosity will prevent the (denser) proppant beingtransported to the tip of the fracture. Instead, the proppant will sink to the bottom ofthe fracture under the influence of gravity and a premature screen out can result.
0
50
0 1 2 3 4 5
Ammonium Persulphate Breaker Concentration (lb / 1000 gal)
Ret
aine
d F
ract
ure
Per
mea
bilit
y (%
)
Fracture Fluid Composition 40 lb / 1000 gal HPG Polymer Solution with Borate Cross Linker
PROPPANT FRACTURING FLUID FRACTURE GEOMETRY& PRODUCED FLUID
Size(Average Grain Size andGrain Distribution)
Grain Roundness& Sphericity
Proppant Crush Resistance&
Fracture Closure Stress
Temperature
Time
Polymer Type andConcentration
Fluid Loss Additive
Crosslinker Type
Breaker Type &Concentration
Temperature
Fracture Width(especially at wellbore)
Reduced Permeability-Non-Darcy or turbuentflow effects at high flowrates and multiphaseflow
The factors which influence the effective proppant permeability when placed in thefracture are summarised in table 6. These include, in addition to the above, the:
(i) fracture width - particularly at the wellbore since:
fracture conductivity = constant*(fracture width)*(slurry proppant concentration).Kprop
Figure 35
Chemical breaker
concentration and retained
fracture conductivity
Table 6
A summary of factors
affecting proppant
conductivity
Department of Petroleum Engineering, Heriot-Watt University 47
6Hydraulic Fracturing6
where Kprop
is the permeability of the proppant. Larger proppant grains will have agreater permeabiltiy {see chapter7 (Sand Control) section 7.5.4}.
(ii) reduced effective permeability due to:
(a) high rate (non-Darcy or turbulent), flow effects. These can be particularlyimportant when stimulating gas wells.
(b) multiphase flow effects (as observed during oil production).
(iii) greater activity from the chemical breaker at higher bottom hole temperaturesleading to greater retained fracture conductivity.
(iv) gradual “clean up” (reduction in water saturation from the fracturing fluid in thenear-fracture formation). This leads to a long term increase in the effectivepermeability of this formation situated next to the hydraulic fracture. Thisprocess can take many months after hydraulically fracturing a low permeabilityformation.
12. FRACTURING FLUID
The properties and function of the fracturing fluid have been discussed in many of theprevious sections. These are all summarised in this section and a more in-depthdiscussion provided on the aspects of fluid loss.
Fracturing Fluid Functions
Characteristics (Required to achieve the above)
(1) Initiate and propogate the fracture(2) Develop fracture width(3) Transport proppant thoughout the length of the fracture(4) Easily produced back to the surface after the fracture treatment is finished, leaving a fracture with the maximum permeability
(a) Stable, predictable rheology under surface and downhole treating conditions and treatment duration(b) Low friction pressures drop at high pump rates in tubing and flow lines(c) Provide fluid loss control(d) Clean and easily degradable to minimise formation damage to propped fracture(e) Compatible with reservoir formation and fluids(f) Economical / low cost
The functions of a fracturing fluid are listed in table 7. The fluid initiates and thenpropagates the hydraulic fracture; creating the required fracture width so that proppantcan be admitted during the slurry stage. Gravity settling of the denser proppant in theless dense fracturing fluid will result in a higher concentration of proppant at thebottom of the fracture than at the top. This settling has to be controlled so that proppantsince we require that it be transported to the tip of the fracture. This is to ensure thatthe required (but not necessarily complete) fracture height is propped. (Rememberthat massive hydraulic fracturing treatments can take many hours). Finally, the
Table 7
Fracturing fluid summary
48
viscosity of the fluid has to reduce to a value similar to that of water so that the basefluid can be easily produced back to surface leaving a fracture with the maximumpossible retained conductivity.
These fracturing fluid functions can be translated into a series of characteristics whichare needed to achieve the above.
(a) Points (1) to (3) in table 7 require that the fluid has a stable and predictablerheology under both surface and downhole (treating) conditions. Thus theaction of any cross linker used has to be stable at high temperature and highshear conditions, while the chemical breaker should not become effectiveduring the time taken to pump the treatment.
(b) Fluids with a highly shear thinning rheology (a low power law n value - seefigure 40, chapter 7 and figure 15, chapter 1) combine the requirements of:
(i) low frictional pressure drop in the surface flow lines and tubing and
(ii) good proppant suspension properties (limited settling of the denserproppant) in the fracture.
This is because the fluid’s rheological properties give it a:
(i) low apparent viscosity in the high shear rate regime present in the tubing and
(ii) an high apparent viscosity in the low shear rate regime present in thefracture.
Obtaining a low pressure loss in the tubing is aided by a second effect due to the fluid’slow power law "n" value. This increases the pump rate at which the transition toturbulent flow occurs - this transition is associated with a large increase in the pressuredrop. In fact, the wellhead treating pressure often decreases when changing, atconstant pump rate, from pumping water (power law n = 1, viscosity = 1 cp) topumping a cross-linked fracturing fluid which has the consistency of a very thick, hairgel when placed in a bottle .
(c) The fracture volume is created by that proportion of the fracturing fluid (pad)which does NOT leak-off into the formation (see Figure 36). In practice, this“useful” portion of the pad, or its efficiency (see below), corresponds to 20%-60% of the total volume pumped. Figure 36 also shows that the fluid leak-offoccurs in a linear manner from the fracture face and can be described by afluid loss coefficient with the units:
volume / (area * time) ≡ distance / time.
Department of Petroleum Engineering, Heriot-Watt University 49
6Hydraulic Fracturing6
Fracture Face
Fracture Height
Fracture Length
Well Fluid leakoff from fracture face
Fluid Efficiency15%
30%
45%
70%
0.005 ft/min0.004 ft/min0.003 ft/min0.002 ft/min0.001 ft/min
Fluid LossCoefficient
Fra
ctur
e F
luid
Vol
ume
Pum
ped
(100
0 ga
l)
0
0
200
400
500 1000 1500 2000
Created Fracture Length (ft)
Figure 37 schematically illustrates the effect of the value of this fluid loss coefficienton the created fracture length as a function of fluid volume pumped. This fluid losscoefficient determines the fluid efficiency, where:
fluid efficiency = volume of fracture created
total fracture fluid volume pumped
The best estimate of the fluid efficiency when designing a specific well hydraulicfracturing candidate is to measure the fluid efficiency during a mini frac treatment.
fluid efficiency = constant*(fluid loss between ISIP and FCP)
total fluid loss to FCP or total volume pumped
Figure 36
Fluid leak-off during an
hydraulic fracturing
treatment
Figure 37
The fluid loss coefficient
controls the created
fracture length
50
This field calibration method of measuring the fluid loss coefficient for use as inputinto the fracture treatment design program produces value which takes into accountvariations in the formation geology (the fracture may contact formation zones ofhigher and lower permeability compared to that observed at the wellbore).
(d) The need for a fracturing fluid that was easy to degrade to leave a maximumpermeability propped fracture was discussed in detail in Section 6.11 (thefracturing fluid and fracture conductivity).
(e) Choosing a fracturing fluid which is compatible with the reservoir formationand fluids ensures that the leaked off fracturing fluid does not produce a zoneof formation damage on the fracture face. Due to the large inflow area of thefracture face, such a damage zone will have a much more limited impact on theformation inflow than similar skin values in the radial flow case (see chapter4 (Formation Damage), section 4.4).
The preferred option is to choose a compatible fluid and to ensure such formationdamage does not occur.
(f) Low cost solutions are always preferred! As shown in Figure 27, the fracturingfluid consists of a base gel prepared by diluting a concentrate polymer solutionwith the base fluid (normally water) and then adding one or more additivesselected from:
(i) cross linking agents
(ii) temperature stabilisers
(iii) viscosity breakers
(iv) clay stabilisers
(v) surfactants e.g. foaming or anti foaming agents
(vi) fluid loss additives
(vii) bactericide
(vii) buffers etc.
The Service Companies market the many, commercially available fracturing fluidsystems together with a wide choice of additives. In practice, an engineeringcompromise has to be made when selecting a particular fracturing fluid and judgingits performance against the criteria listed above. A preferred choice for use whenfracturing medium temperature wells is a fluid based on a low concentration (40 lb/1000 gal) of Hydroxy Propyl Guar (HPG) polymer dissolved in a dilute brine with theaddition of a borate ion cross linking agent. This fluid meets the table 7 criteria whilethe low polymer concentration minimises the cost and maximises the retained fractureconductivity.
Department of Petroleum Engineering, Heriot-Watt University 51
6Hydraulic Fracturing6
13. TIP SCREEN OUT FRACTURING
Chapter 6 Described the carrying out of a conventional hydraulic fracturing treatmentdesigned to stimulate wells completed in low permeability formations with long, thinfractures. The created fracture width at the wellbore depends on the fracture lengthand the rock compliance (longer fractures are wider) while the propped fracture widthdepends on the slurry proppant concentration (higher proppant concentrations willlead to greater propped width).
The capabilities of hydraulic fracturing technology have been extended into softerformations (which often have a medium to high-permeability). This involves thehydraulic fracture treatment design to deliberately create an early Tip Screen Out(TSO) This process involves the pumping of the proppant slurry early during thefracturing treatment. When the first proppant arrives at the fracture tip it forms a“bridge”, or proppant plug at the tip (the TSO).
A. Fracture intiaties
B First proppant enters
C. Tip screenout
D. Fracture inflation by continued pumping proppant slurry
E. Packed fracture
Proppant reaches fracturetip and screens out
Increasing pressure in fracture
Fluid leak off
The proppant plug has a relatively high pressure drop across it compared to thefracture. This reduces the fluid pressure acting on the fracture tip, so that furtherfracture length growth is halted and fracture width growth can begin to provide theextra fracture volume to accommodate any further volumes of proppant slurryinjected into the fracture. This results in an increase of the FPP since the fracture is nolonger increasing in length.
A typical treatment history is schematically illustrated in Figure 38b and table 8. Thetreatment proceeded as follows:
(i) The fracturing fluid used was a low concentration (40 lb/1000 gal) HPGpolymer solution with a borate cross linker (as recommended above).
(ii) The treatment was started with a pump rate of 5 bbl/min. Fracture re-openingwas confirmed by a peak in the downhole pressure, followed by a slow decreaseas the fracture extended in length.
(iii) After 13 minutes the pump rate was increased to 25 bbl/min - with a correspondingrise in treating pressures. A much greater increase was observed in the surfacemeasurement since the increased pump rate gave rise to some 3000 psi extrafrictional pressure drop across the tubing.
Figure 38(a)
Fracture width inflation
with the tip-screenout
technique.
52
(iv) Pumping of proppant slurry was begun after 23 minutes with the first proppantarriving at the perforations some 3 minutes later. The bottom hole pressurebegan to increase after 28 minutes. The tubing head pressure drops between 28and 30 minutes because the proppant containing slurry is denser than thefracturing gel above.
(v) Pumping of the higher proppant concentrations - stage 3 began at 30 minutes.The bottom hole pressure continued to rise smoothly while the tubing headpressure shows a much steeper increase in pressure after 33 minutes - the denserproppant slurry has an increased viscosity compared to the cross-linked gel alone.
(vi) The final proppant stage was pumped (stage 4) and displaced to the perforations(stage 5).
Figure 38b and table 8 show that the crucial stage of this fracturing treatment occupiesa very short time period - 67% of the proppant is pumped in a only a 5 minute timeperiod. Also, the total treatment took less than 40 minutes.
0
0
1000
2000
3000
4000
5000
6000
10 20 30 40
StartPumps
PumpRate
PumpRate
Tubing HeadPressure
Tubing HeadPressure
Measured BottomHole Pressure
Surface ProppantConcentration
Start PumpingProppant atSurface
First Proppantarrives Downhole Downhole
ProppantConcentration
Stage No. 1 2 3 4 5
Pre
ssur
e (p
si)
Time (min)
Fracture initiation Fracture creation Pack fracture Overflowwith the increasingproppant concentrations
Figure 38(b)
Treatment record for tip
screen out hydraulic
fracture stimulation
Department of Petroleum Engineering, Heriot-Watt University 53
6Hydraulic Fracturing6
Stage Fluid Volume Proppant Cumulative (gals) Consentration Proppant
(lb/gal) ( pumped lb)
1
2
3
4
5
Fracturing GelCreates Fracture
Dilute Slurry
ConcentratedSlurry
Very Consentrated
Slurry
Base FluidOverflush
0
0-4 (increasing)
4-12 (increasing)
12
0
15,000
16,000
36,000
48,000
7,000
3,500
1,000
1,800
These figures should also be compared with the equivalent ones for a massivehydraulic fracture stimulation treatment (table 4). The TSO fracture uses only 5% ofthe amount of proppant and 10% of the fluid volume. These differences between thetwo types of treatment are summarised in table 9 and Figure 39.
Fracture Type
Description
Width (inches)
Length (ft)
Proppant
Concentration
( lb / ft2 )
Conventional
Long and Thin,(Lower Conductivity)
> 0.25
500 - 1500
0.5 - 2.0
Tip Screen Out
Short and Fat,(Higher Conductivity)
0.25 - 1.5
50 - 500
4 - 12
Table 9
Tip screen out and
conventional fracturing
compared
Table 8
Pumping schedule for tip
screen out hydraulic
fracture stimulation
treatment
54
Conventional Propped
Fracture
500 - 1500 ft.
50 -
500 ft.
Tip
Screen-O
ut
Propped
Fracture
0.25 in.
0.25 - 1.5 in.
13.1. Applications of TSO FracturingTSO fracturing is now a field proven technology for use in a number of areas.
(1) Sand control: gravel packing and fracturing are combined into a singleoperation. In addition, a number of successful field trials in which fracturingalone was used have been reported (successful sand control was attributed to thereduced flow velocities associated with the TSO fracture compared to aconventional perforated completion).
(2) Alternative to matrix acidising: using highly conductive fractures to bypassnear wellbore formation damage. Successful matrix acidising often requiresthat the source of formation damage be identified in order to select the optimumacid formulation {see chapter 5 (Acidising and other matrix treatments) section5.2}. In contrast, TSO fracturing has the advantage that it is independent of thetype of formation damage. This type of treatment is generally called a "skinfrac".Figure 40(a) shows how the well productivity increases slowly once the highlyconductive hydraulic fracture has penetrated the damage zone.
(3) Reserve increase in laminated sands: well completion in formations consistingof finely laminated sands is problematic since there is a high chance that theperforation density will not be sufficient for the well will to make contact withall thehydrocarbon bearing zones (Figure 40b). Production via a TSO fracturewill ensure that recovery is achieved from all the zones (Figure 40c).
Figure 39
Tip screen out and
conventional fracturing
compared
Department of Petroleum Engineering, Heriot-Watt University 55
6Hydraulic Fracturing6
(a)
Shale Permeability Barrier
Unconnected Zone
Pay Zone
Pay Zone
Pay Zone
Pay Zone
Pay Zone
Pay Zone
(b)
TSO Fracture
Pay Zone
Pay Zone
Pay Zone
Pay Zone
Pay Zone
Pay Zone
No fracture (damaged well)
Lf = 5 ft
Lf = 15 ft
Lf = 40 ft
Lf = 80 ft
Lf = 150 ft
k = 100 md
rd = 10 ft
kd = 0.05 * k
S = 65
ks.w = 8,000 md-ft
0.01 0.1 1.0 10 100 Months
106
105
104
103
Fracture partially penetrates damage zone
Cum
ulat
ive
prod
uctio
n
(c)
Production from undamaged well
Time
Figure 40c
Cumulative production for
different fracture lengths
piercing a 10 ft radius
damage zone around the
wellbore.
Figure 40a
Lost reserves in sands not
connected by perforations
Figure 40b
Tip screen out connects
extra reserves. All
productive sand bodies
being produced after
connection via hydraulic
fracture
56
(4) Production of medium viscosity oil from lower permeability formations: theincreased production rates achievable using TSO fracturing can sufficientlyimprove the project cash flows that economic development of fields becomespractical. A possible alternative development scenario is to use horizontalwells to increase the well rates or even combine the two technologies. This isschematically illustrated in Figure 41 where the large number of closely spacedpropped TSO hydraulic fracturing treatments have to be placed along a 1200mhorizontal well. This type of completion has been used on a wide scale for theoil bearing chalk fields in the Danish sector of the North Sea.
Figure 42 shows that the potential production rate increases as the number of hydraulicfractures increases, but that the rate of increase for each extra fracture is continuallydecreasing because they interfere with each other in the terms of the reservoir inflowas well as frictional pressure losses along the length of the horizontal well becomingprogressively more important.
(5) Fines migration: is triggered by high flow velocities due to radial inflow intoa perforated well. The flow velocities are reduced, and this form of formationdamage avoided, by the linear flow associated with a fractural well. Similarly,sources of formation damage which are caused by pressure reductions may beavoided since the improved well inflow will increase the flowing bottom holepressures (at a given production rate).
��������������������yyyyyyyyyyyyyyyyyyyy
��������������������yyyyyyyyyyyyyyyyyyyy�
�yy
��yy
��yy
��yy��yy��yy�
�yy
OIL
MassiveChalk
Formation
HorizontalWell
145m GAS140m
7thfrac
6thfrac
5thfrac
4thfrac
3rdfrac
2ndfrac
1stfrac
WATER
10
0
0
20
30
40
2 4 6 108
Number of Fractures
Flo
w R
ate
(10
00 m
3 / d
ay )
Figure 41
Completion employing
combined horizontal well
and multiple fracturing
technologies
Figure 42
Example of increased gas
production achieved from a
multiply fractured
horizontal well in a low
permeability formation
Department of Petroleum Engineering, Heriot-Watt University 57
6Hydraulic Fracturing6
HYDRAULIC FRACTURING TUTORIAL
Question 1
Contrast the application areas of Matrix Stimulation and Propped Hydraulic Fracturing
Answer 1
Matrix (often Acidising) - removal of near wellbore damagePropped Hydraulic Fracturing - improving well inflow performance by creating a
high permeability channel with a large surface area for fluid inflow
Stimulation treatments are designed to increase the well’s Productivity Index (PI).
PIKh
B rr Se
w
=
+
µ ln
Matrix acidising aims to increase PI by reducing S by dissolving formation damageand rock in the near wellbore region.
Fracturing increases PI by increasing the effective wellbore radius. Creates a highconductivity fracture or channel from the wellbore. This extends deeper into reservoirthan can be achieved by acidising. It can also bypass the near-wellbore formationdamage zone and is occasionally used for this purpose alone.
Skin Permeability TreatmentHigh High Matrix treatmentHigh Medium Matrix/"Frac and Pack"High Low Fracturing (matrix may be possible if perm. not too low)Low Low Hydraulic fracturingLow High Treatment economic?
Question 2.
You are a service company representative who has designed the following fracturingtreatment:
Wellbore radius (rw): 0.328 ftReservoir height: 200 ft; bounded by competent shalesReservoir Permeability (k): 1.0 mDProppant to be used: 500,000 lbDesign Fracture Conductivity (kf*w): 1,500 mD*ft at 4 lb/ft2 proppant loading
• Use the Cinco-Ley and Samaniego type graph (below) to advise your client on theexpected well negative skin (the fracture treatment covers the full formation height)
• Do you think this is the optimum fracturing treatment for this well?
58
2
1
0
0.1 1 10 100 1000
Sf +
In(L
f/rw
)
FCD = kf.w k.(Lf)
Lf is fracture half length
Answer 2.
Proppant loading * frac. heightFracture length (both wings) =
=500,000 / (4*200)=625 ft
Fracture half length (Lf) = 312.5 ft
FCD = kf*w/k*L
f = 1,500 / (1*312.5) = 4.8
From graphIn (L
f/r
w) + S
f = 0.91
∴ Sf = 0.91 - In (312.5/0.328)
Sf = - 5.9
∴ Fracturing treatment will achieve a negative skin of - 5.9.
Fracture treatment design will produce a well with a negative skin of -5.9, a reasonablevalue, However, a propped fracture with a higher FCD might be even more beneficialto ensure that well productivity will not be impeded by lack of fracture conductivity.An FCD of 10 - 15 is the standard recommendation, but the optimum in a particularcase depends on a detailed well production / fracturing treatment cost optimisation forthe case under study.
Department of Petroleum Engineering, Heriot-Watt University 59
6Hydraulic Fracturing6
Question 3
The operator you are working for notices the above and queries whether this is anoptimum fracturing treatment design. However, he says that he is already over budgetso you are not allowed to increase the size of the hydraulic fracturing treatment butcommissions laboratory tests that show a fracture of Fracture Conductivity (kf*w) of2,500 mD*ft can be expected at proppant loading of 5 lb/ft2 and 3,500 mD*ft atproppant loading of 6 lb/ft2. You have to advise the operator as to which fracturedesign is optimum and why.
Answer 3
The limitation on the cost of the fracturing treatment means that the only way toincrease the well productivity is to alter the manner in which we deploy the proppant.We need to increase the dimensionless fracture conductivity (FCD), implying that weneed to produce a shorter, more conductive (fatter) fracture.
Repeating the above calculation scheme gives:
Fracture length (both wings) = 500,000 / (5*200) or 500,000 / (6*200)= 500 ft j = 416 ft
Fracture half length (Lf) = 250 ft = 208 ft
FCD = kf*w/k*L
f= 2,500 / (1*250) = 3500 / (1*208)= 10 or = 16.8
From graph In (L
f/rw) + S
f= 0.81 or = 0.76
∴ Sf
= 0.81 - In(250/0.328) = 0.76 - In(208/0.328)S
f= - 5.8 = -5.6
The original fracture design thus appears to be optimum, giving the most negativeSkin value. This arises from the possible combinations of proppant loading (whichcontrols fracture length) and the created fracture conductivity.
60
14. FURTHER READING
“Production Operations” Volume 2 (4th edition)by T. Allan and A. Robertspublished by Oil and Gas Consultants IncISBN 0-930972-18-X
“Petroleum Production Systems”by M.J. Economides, A.D. Hill and C. Ehlig-Economidespublished by Prentice HallISBN 0-13-628683-X
“Well Performance” (2nd edition)by M. Golan and C. Whitsonpublished by TapirISBN 0-13-9046609-6
“Reservoir Stimulation” (2nd edition)edited by M.J. Economides and K.G. Notlepublished by Schlumberger Educational ServicesISBN 0-13-775115-X
ELY J WStimulation Treatment Handbook: An Engineer's Guild to Quality ControlPenn Well 1985ISBN 0-87814-284-3
2
Well Control 11
C O N T E N T S
1. INTRODUCTION2. TYPES OF SAND PRODUCTION
2.1. A Southern North Sea Case History2.1.1. "Continuous" Sand Production2.1.2. “Clean-up” sand production2.1.3. “Fines” Production2.2. Sand Cementation2.3. When is Sand Production a Problem?2.3.1. The Consequences of Sand Production2.3.2. Living With Sand Production2.3.3. Monitoring of Sand Production
3. PREDICTION OF SAND FAILURE3.1. Field Experience3.2. Petrophysical Analysis3.3. Rock Strength Measurement3.3.1. On Site Strength Estimation3.3.2. Rock Mechanical Strength
Measurements3.3.2.1. Unconfined Compressive Strength3.3.2.2. Brinell Hardness Number3.3.2.3. Thick Wall Cylinder Collapse Strength3.3.2.4. Trixial Rock Strength Measurement3.4. Prediction of Downhole Rock Failure3.5. A Final Word
4. COST OF SAND CONTROL5. SAND EXCLUSION
5.1. Introduction5.2. Different Types of Mechanical
SandExclusion5.2.1. Open Hole Completions5.2.1.1. Slotted Pipe5.2.1.2. A Wire Wrapped Screen5.2.1.3. Resin Coated Sand Pre-Packed Screen5.2.1.4. Application of Open Hole Completions5.2.1.5. Enhanced Drilling Fluid Requirements
for Open Hole Completion’s5.2.2. External Gravel Packs5.2.3. Internal Gravel Pack5.2.4. Special Gravel Packs5.3. Advantages / Disadvantages of
GravelPacking5.4. Gravel Pack Sand Selection5.4.1. Operational Considerations5.5. Gravel Packing - Surface Operations5.6. Fluids for Gravel Packing5.6.1. Properties of Viscous Gravel Packing
Fluids
2
7Unstable Formations and Sand Control7
5.6.2. Other Base Brines5.7. Fluid Loss Control5.8. The Gravel Pack Operation5.9. Gravel Placement with Low Viscosity
Fluids5.10. Gravel Placement with High Viscosity
Fluids5.11. New Technology5.11.1. Gravel Pack Evaluation5.11.2. "Frac and Pack"5.11.3. New Screen Technology5.12. Chemical Sand Consolidation
6. FURTHER READING
1
2
LEARNING OBJECTIVES:
Having worked through this chapter the student will be able to:
• Describe the impact of an incorrect decision of whether or not to allow for theinstallation of sand control during the original completion design
• Explain the reasons for the lack of definition of “what constitutes a sand problemin the field”
• State measures taken to modify completion/surface facilities to “live with” sandproduction
• Construct a methodology for predicting sand failure
• Discuss the various rock mechanical measurement that may help with this prediction
• Identify and contrast the advantages/disadvantages of the various sand controloptions (none, sand consolidation, bare screen, internal and external gravel packs,“Frac and Pack”)
• Identify the key elements of liner/screen design
• Discuss openhole completions design and drill-in fluid selection
• Discuss the characteristics and functions of the fluid and proppant used in a gravelpack
• Design a gravel pack be selecting the optimum gravel pack size and fluid for agravel pack operation
• Calculate the pressure drops associated with flow through a gravel pack completion
• Describe fluid and proppant placement for the various types of gravel pack options
2
Department of Petroleum Engineering, Heriot-Watt University 3
27Unstable Formations and Sand Control7
1. INTRODUCTION
Sand production from (relatively) unconsolidated reservoirs is a frequently encountered,costly operational problem which has a significant impact on the case of welloperation and the economics of oil or gas production. It is usually associated withshallow, young formations, but has also been encountered at depths greater than4000m. Other types of unstable formations also show similar production problems -i.e. they allow “pieces” of the formation to break away and enter the perforation orwellbore. Soft chalks, shales, siltstones and rubble zones can all flow particles,undergo plastic failure or slough particles due to mechanical formation failure thestresses imposed by well production.
This chapter will:
(i) describe the rock mechanical process that lead to sand production,
(ii) discuss what institutes a sand production problem from a production operatorspoint of view,
(iii) indicate the operational measures taken to identify a sand problem well,
(iv) describe the various types of sand control.
These include sand consolidation and the mechanical methods:
screens / prepacks / gravel packing / frac packing and finally:
(v) discuss the completion operations required for the installation of sand control.
Identification of which wells require the inclusion of sand control equipment,especially when during (initial) phase of a new field development, is a difficult butkey decision.
Incorrect omission of sand control measures during the completion causes Loss ofProduction due to:
• Sand bridges forming in the tubing, sand filling up the casing so that the perforationsare covered and production from the lower intervals is lost
• Damage to the well’s integrity e.g. casing/liner collapse due to loss of lateralsupport in the areas where a cavity has formed due to sand production
• Valves and other downhole equipment become stuck due to the presence of solidparticles preventing their mechanism working properly
• Erosion (sand blasting) of holes in the surface pipework leading to the pollutionassociated with an oil or gas spill and the consequent loss of well control.
• The produced sand needs to be removed from production equipment and disposedof in an environmentally acceptable manner.
1
4
The unnecessary inclusion of sand control measures results in:
• Extra completion cost (higher initial CAPEX) and loss of future workover flexibilityfor the well
• Reduced well production from the impairment brought about by gravel packing
• Increased artificial lift costs due to this reduction in well Productivity Index
• Loss of reserves since the well is taken off production due to the Well’s MinimumEconomic Production Rate, at which the well operating expenses equal incomefrom the oil production, being reached after less total cumulative production.
The technology developed by the industry for efficient sand exclusion along with amore detailed discussion of the above will be covered in the following chapters.
2. TYPES OF SAND PRODUCTION
The determination of the sand production potential of a given completion is difficultbecause there is no clear picture of the underlying (scientific) mechanism(s) that resultin formation failure and sand production. Formulation of the problem is difficult. Thepay zone or sand body is heterogeneous, displaying a wide range of mechanical andchemical properties. Drilling of the well and shooting of the perforations will alter thenatural stress regime. Naturally strong formations will accommodate this redistributedstress; while for weaker sand this stress may exceed the formation strength. To furthercomplicate the issue, the increased stresses due to fluid flow towards the wellbore,(frictional forces) as well as pore pressure changes (draw down and/or depletion) mayall contribute to this exceeding the formation strength. Sand production can commenceonce the sum of all these forces exceeds the formation strength (figure 1).
the problem area
sand control may be requiredsand control required sand productionnot expected
Rock StrengthLow High
The fact that the forces initiating sand production increase with a rising wellproduction rate, implies that there can be a critical flow rate below which sandproduction will not occur. This has been observed in the field - beaning the well back
Figure 1
Is/Will sand control be a
problem ?
2
Department of Petroleum Engineering, Heriot-Watt University 5
27Unstable Formations and Sand Control7
to reduce the production rate is a field proven method to temporarily alleviateproblems caused by wells suddenly producing sand (figure 2).
0
0
0
0
sand
pro
duct
ion
w
ater
cut
sand
pro
duct
ion
maximumsand-freerate
simultaneoussand & waterproduction
production rate
background level
background level
time
Wat
er
2.1 A Southern North Sea Case HistoryGas production was initiated from this Rotligindes gas reservoir in 1971. All the wellson the platform started producing sand in 1976 at a FBHP of around 2,600 psi.However, it was noted that the sand production diminished with time at a constantFBHP and stopped altogether when the FBHP was increased. The following sandcontrol strategy was developed over the next 8 years:
• Avoid perforating the weakest rock
Figure 2
Sand production triggered
by high production rates or
water production
1
6
• Ensure wells are beaned up slowly
• “Pre-condition” all sand prone wells prior to the (high rate) winter productionseason by producing them to an acceptably low sand production rate at a lowerFBHP than is expected to be reached during the forthcoming winter season.
• The sand production will then stop once the production rate is decreased (FBHPis increased) below the “Pre-conditioning” level.
Field experience showed that routine gas production has been virtually sand free formany years since 1984. This strategy has the advantage of being cheap does notrequire the installation of downhole sand exclusion and has allowed an economic wellproduction rate. All wells have been produced at the maximum rate during the winterperiod and the absence of formation damage associated with types of sand exclusionhas maximised recoverable reserves.
N.B. In common with many Rotligendes gas fields; there is no, or only very limited,water production at the perforations. The question that the production engineer oftenhas to answer is: "when is water production expected and will it trigger sandproduction?"
2.1.1 "Continuous" Sand ProductionVery soft formations may exhibit sand production at the time of initial discovery - theso called “sloughing” sands can lead to great difficulties during completionoperations when hole collapse can become a major problem. Many other, normallysomewhat more consolidated formations, begin to show sand production after aconsiderable period of production - due to reservoir pressure depletion (decrease inminimum in-situ stress), water production, increased fluid velocities etc. Propercompletion practices are often critical in marginal situations where sand productioncan be created by poor completion practices e.g.
increased drawdown + plugged perforations↓
increased local fluid velocities + potential for water coning
One possible solution is to repeat the perforation process so as to minimise the wellproducing drawdown.
Continuous sand production - as discussed above - has to be distinguished from:
2.1.2 “Clean-up” sand production:This refers to (relatively minor amounts) of sand that are produced from a new wellduring its first few days of production. This form of sand production is attributed tothe production of material (partially) de-consolidated by the shock wave producedduring the perforating process. Further, local rock failure may alter the shape of theinitially formed perforation tunnel so that it changes to a more stable shape as aresponse to the local stresses. In extreme cases, the change in the perforation shapemay take place to such an extent that the individual perforations merge together toform a “cavity”. Such “cavity” formation results in the casing becoming unsupported,leading to casing failure in some cases.
2
Department of Petroleum Engineering, Heriot-Watt University 7
27Unstable Formations and Sand Control7
The key point about transient sand production is that it is temporary i.e. it reduces toa low level after a certain period of time when the well is kept on production at aconstant rate. The time required for stabilisation to occur is dependent on the type ofwell - gas wells responding more quickly than light oil wells, which, in turn, respondfaster than heavy oil wells (figure 3).
Time (hrs)
Time since startup (days)
Time since startup (days)
300
0
0
San
d P
rodu
ctio
n (c
ount
s)S
and
Cut
(pp
tb)
San
d C
ut (
g/m
3)
Bean-up
a. Gas wellSand volume ~ 1litreInterval 20m
b. Light oil wellsand volume ~ 35 litreInterval 10m
c. Heavy oil wellsand volume ~ 200litreInterval 20m afterbeam pump artificial installed
10 40 50
10
20
30
100
200
300
400
500
Pro
duct
ion
Rat
e
200
A new burst of sand production is observed each time the well's production isincreased (beaned up). This is thought to be due to formation of a stable arch aroundthe entrance to a perforation cavity (figure 4). This arch remains stable as long as flowrate and drawdown are constant. If these are altered, the arch may collapse and a newone forms once flow stabilises again. To encourage such stable arch formation andhence prevent sand production, sand-prone wells should be opened up slowly(possibly over a period of hours or days). Well production conditions should bemaintained as constant as possible.
Figure 3
Typical transient sand
production for different
types of wells
1
8
I I
I II
I I
I II
II
I
II
II
II I
II
II I
I II I
I
I
II
II
I
II II
II
I
II
II
I
III I
IIII
I
II
I
I I
I II
I I
I II
II
I
II
II
II I
II
II I
I I
I II
I
II
II
I
II II
II
I
II
II
I
III I
IIII
I
II
I
Sand Grain
Cement
Fluid Flow
Fluid Flow
Fluid Flow
Fluid Flow
Formation
Perforation Tunnel
Cas
ing
Cas
ing
Enlarged perforations
Cavity
Enlarged perforationsmerge to form cavityincreased wellbore radius
Fluid Flow
Fluid Flow
Fluid FlowDebris
Fluid Flow
Fluid Flow
Fluid Flow
The individual perforation cavities may merge into a larger cavity, which may alsostabilise (figure 5).The formation of a large cavity means that the well is no longersupported by the formation. Complete loss of the well due to casing failure thenbecomes a very real possibility. A residual, constant level of solids production maycontinue once a stable arch has been formed. This is discussed in the next section.
Figure 4
Cavity stabilisation by Arch
Formation
Figure 5
Sand production can lead
to large cavities
2
Department of Petroleum Engineering, Heriot-Watt University 9
27Unstable Formations and Sand Control7
2.1.3 “Fines” ProductionSand production relates to the production of load bearing formation solids while"fines" production involves the production of mobile, very small, solids which are notpart of the mechanical structure of the formation. These clay, feldspar or silica "fines"are much smaller than the formation sand grains since they can pass through the porethroats. Attempting to stop production of these “fines” would drastically impair thewell’s production. Allowing them to pass into the well with the producing fluidsprevents them causing permeability damage at the interface of the gravel pack and theformation. As a practical “rule of thumb”, the smallest 5% of solids found duringsieve analysis of (soft) cores are probably interstitial solids and can be considered tobe potentially produceable “fines” in the above context.
2.2 Sand CementationSandstone formations were originally laid down as a bed of loose sand grains at thebottom of a river, or as a beach at the sea shore. Over geological time these individual,loose grains became cemented or consolidated together - a process which resists sandproduction. The individual sand grains making up most sandstone formations arebonded together by clay, quartz, calcite, mineral growth or precipitate bonding. The“overburden load” (or weight of all the sediments on top of the formation) is resistedby the strength of the individual sand grains, the pressure of the fluids within the porespaces and the strength imparted to the consolidated formation by these intergranularcements (figure 6).
0 00 0
00
0
000
0 00 0
00
0
000 0
0 00
00
00
0 00 0
0
00
00
0 00 0
00
00
0 0 00 0
00
0
00
00
00
00
0 0 00 0
00
0
00
0 00 0
00
0 00
000
0 00
00
0
0 0
00
0
00
0 00 0
00
0
00
00
00
0 00
000
0 00
00
0
0
0
0
0
0 00 0
00
0
00
00
00
0
0 00
00
00 0
Sand Deposit
Burial + Pressure Soloution Cementation
Overgrowth: Quartz, Feldspar
Grain Coating Clays: Chlorite, SmectiteIlite, Siderite, etc.
Dolomite
Calcite
Pore filing cements
Siderite
Pyrite
Kaolinite
Unconsolidated
ConsolidatedWeak Strong
Figure 6
Sand burial and
cementation process
1
10
Cementation is developed in several ways but is most commonly due to precipitationof minerals from the water phase within the pore space between the sand grains. Thelocalisation of the cementing precipitate at the grain-to-grain contact is caused by apressure dissolution and precipitation cycle. This is brought on by the application of(high) overburden load on the small area of the point-to-point contacts of theindividual sand grains. At the high pressures associated with the point-to-pointloading, dissolution of the contacting surfaces of the sand grains takes place since thesand (quartz) is more soluble at the point of loading. The water surrounding the activesites becomes over-saturated in quartz, compared to the unloaded part of the sandgrain. As time passes, the contact points of the grains become flatter, spreading theload over a wider area of total contact i.e. the local point loading decreases.Eventually, the increased contact area created by this method is sufficient to withstandthe overburden and the form of the individual sand grains is stabilised. Precipitationof the super- saturated matrix minerals in the water, occurs on the sand grain surfaces.The result is that the grains are “cemented” together. This pressure solution driven"dissolution and precipitation" process at the grain contact is only one method of sandgrained cementation. Additional amounts of the same or different cements may beprecipitated from flowing hydro thermal fluids. Mineral cements may be precipitatedas the water in the pore space comes to equilibrium with the local conditions of thetemperature, pressure and mineral composition. Quartz overgrowth around calciumcementation or clay development at the grain boundaries is a sign of secondarycementation.
Formations may not consolidate for a number of reasons, e.g. reduced compactionloading caused by shallow burial or very large, load supporting arches above the pay(e.g. grabens). Faults may block the necessary stresses and leave high porosity andpoor intergrain bonding. If the sealing structure is filled with hydrocarbon shortlyafter formation, mineral solution or precipitation reactions from water cannot takeplace. High pore pressure (geopressures) also reduces the compaction loading; in thiscase the pore fluid carries a greater percentage of the overburden supporting load.Other types of chemical reactions such as ground water leaching of the matrix grainsor dolomitization, may heavily modify the cement or the size of the sand grain.
2.3 When is Sand Production a Problem?The well’s sand production tendency may change during the lifetime of the well. Inparticular, what was a negligible or marginal problem may become much more severelate in field life due to:
(i) Water Production. This results in:
(a) higher drawdowns (increased rock stress) for some production rate due torelative permeability effects.
(b) loss of capillary pressure between the sand grains. It can be shown theoretically,or from experimental measurements, that this does not significantly reduce therock strength. However, it does allow the already failed (loose) rock particlesto be produced by the forces associated with fluid flow.
2
Department of Petroleum Engineering, Heriot-Watt University 11
27Unstable Formations and Sand Control7
(ii) Production Plans that call for maintaining the net oil production. This resultsin a continually increasing gross fluid production as the water cut increases. i.e.production occurs at an ever increasing drawdown and fluid flow velocities.
(iii) Reservoir Pressure Depletion results in further increases in stress on the rock.
Avoiding perforating the weakest rock may reduce the well’s sanding tendency.However, a decision still has to be made whether to control sand production (the safeoption) or live with it and risk the (later) consequences of sand production, asdescribed in the next chapter.
2.3.1 The Consequences of Sand ProductionSand production has numerous technical, environmental, operational and economicconsequences:
The operationally acceptable level of sand production will vary with the location, thewell and production facility design and local conditions. The effects of sand onproduction operations and the potential safety, financial and environmentalconsequences will influence whether sand production limits are set and their level.
Potential Consequences of Sand Production
Location Problem Effect
Wellbore fill from bottom • Restricted access to production interval
• Reduced work over success since Completion interval Cavity formation zonal isolation no longer practical
• Increased well productivity Index
• Loss of full diameter access to Casing buckling completion interval
• Casing failure in extreme case
Tubing Sand bridge • Well production halted • Loss of productivity and reserves
• Sub surface, surface controlled, Sand deposition safety valve (SSSV) not operating -
Subsurface equipment in tubing and accessories a safety issue • Difficult Wireline operations
Erosion • Equipment replacement and failure
• Control equipment malfunction • Reduction in separator residence time
Sand accumulation • Unscheduled shut-downs/deferred production
Surface installations • Sand separation from hydrocarbons and disposal
• Failed equipment replacement Erosion • Oil/gas spill (environmental/safety
issue)
Table 1
Potential Consequences of
Sand Production
1
12
Unconsolidated formations are rarely homogeneous and apparent formation strengthis affected by changes in depletion, water saturation and pressures over the life of thefield. This results in one of the biggest problems for predicting sand movement andfor designing control methods for new field developments The flow rate that willproduce sand in one interval may cause no damage only a foot away and the controlprocess that will halt the flow of sand in a zone with large formation sand grains maynot work in zones made up of smaller grains. The whole formation must often be giventhe same treatment as the worst-case zone. Remedial Sand control measures arerequired in formations where large sections of the hydrocarbon bearing sand are weak.However, in formations where only a small percentage of the zone presents a risk ofsand production, selective perforating can often avoid the need for installation of sandcontrol. The answer is to determine which sections of the pay zone sand will fail.
2.3.2 Living With Sand ProductionA sand management system has to be installed if the “living with sand” option ischosen. Further, the production system may be changed so that it becomes moretolerant of the volumes of sand that are produced. The measures taken could include:
(i) measurement of the amounts of sand produced
(ii) installation of hard faced chokes in the “bean” box
(iii) installation of appropriate artificial lift methods e.g. gas lift with no movingparts through which the oil flows - is more sand tolerant than electric submersiblepumps (esp) with their rapidly rotating impellers.
(iv) monitor flow line wall thickness. An X-ray or sonic measurement device canbe used for this purpose. The same techniques may be used to monitor sand buildup in surface vessels (figure 7)
SAND
Radioactivesource
Ray Path
MovableDetector
high velocity water jetsfluidises sand atbottom of vessel
(v) design flow line so that large changes in velocity and direction do not occur e.g.remove 90° pipe bends.
(vi) install surface sand collection and disposal facilities e.g. water jets installed atthe bottom of separators. The high velocity water stream fluidises the sand andtransports it to a collection and disposal vessel. Here, the adhering oil may beremoved by vigorous agitation with water (and surfactants) to allow easier subsequentclearing and disposal.
Figure 7
Monitoring sand level in
vessel through increased x-ray
absorbtion and removal of
settled sand.
2
Department of Petroleum Engineering, Heriot-Watt University 13
27Unstable Formations and Sand Control7
(vii) increase the well production in a number of steps (slow bean up) and avoidingcycling the downhole pressure.
(viii) ensure that fluid velocities are sufficient to transport sand to the surface andtransport it to an operationally convenient collection point. This is done bycalculating the terminal settling velocity (Stokes law) and comparing this with theminimum (upward) flow velocity. Sand particles will concentrate at points at whichthe settling velocity is greater than the flow velocity. The settling velocity increaseswith increasing particle size and decreases with increasing fluid viscosity (figure 8).N.B. The settling velocity will change as the fluid composition changes e.g. it willdecrease when (viscous) emulsions are formed or increase in the presence of watercontinuous oil-in-water dispersions formed at high (~>50%) water cuts.
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.05
0.1
0.15
0.2
0.25
Sand Particle Size (mm)
Term
inal
Set
tling
Vel
ocity
(m
/s)
50 API
SuspendingFluid(Crude Oil)
0
35 API0
20 API0
It is not practical to live with unlimited amounts of sand production. Typicaloperationally allowable levels are summarised in table 2.
Typical Allowable Sand Production Levels
Produced Fluid Production Rate Allowable Sand Level
Gas (<50 Mscf/d) 1 lbs/MMscf (>50 Mscf/d) 0.5 lbs/MMscf
Light Oil <5000 bopd 30 lbs/1000 bbls* 5000-15000 10 lbs/1000 bbls>15000 5 lbs/1000 bbls
* reduced by 50% for high GOR
Heavy Oil 200 lbs/1000 bbls or even higher
Table 2
Typical Allowable Sand
Production Levels
Figure 8
Terminal settling velocity of
sand particles in different
crude oils
1
14
It can be seen that the allowable sand content is very dependent on the fluid velocity(production rate) and fluid viscosity.
2.3.3 Monitoring of Sand ProductionThe various techniques used for monitoring sand production are summarised in table3. This table indicates the measurement principle employed, its key points and theassociated drawbacks.
MONITORING OF SAND PRODUCTION
Technique Principle Measurement Drawbacks
Wellhead Sampling Centrifuge wellhead Simple Not continuous; produced oil sample Non-representative
Manual In-Line Sand Trap Sand settling into pots Simple Allows some estimation of sand
located in flow lines at flowline production rate Manual
Erosion Probes Hollow probe inserted into Erosion measurement. Alarm system only; install where flowline: penetration of probe wall Depends on sand production severe erosion expectedby sand erosion activates alarm rate,flow rate, type of since flow line pressure appears produced fluids, inside probe probe position etc.
Sonic Detectors Acoustic noise of solids impact Detects mass rate of sand Interference from flow noise gas measured by a piezo electric crystal produced i.e. larger particles bubbles and small solid particles which is mounted on a make bigger impact (fines) solid probe which is mounted in No erosion measurement
the flowline
Flowline Erosion Measure wall thickness by ultra Erosion detection only Not on-line or continuous Sonic / x-ray techniques Expensive
3. PREDICTION OF SAND FAILURE
A number of techniques have been employed when deciding whether to install sandexclusion techniques. These include:
(i) Field experience
(ii) Petrophysical log and core analysis
(iii) Wellsite rock strength estimation
(iv) Rock mechanical measurements and calculation
Unfortunately, none of these techniques yields a perfect answer, as will becomeapparent in the following sections.
3.1 Field ExperienceThe production history of wells producing from the same formation in the field wherethe new well is planned is the best guide as to whether sand control should be installed.Alternatively the same formation may be found in a nearby field with the samegeological history.
Table 3
Monitoring of sand
production
2
Department of Petroleum Engineering, Heriot-Watt University 15
27Unstable Formations and Sand Control7
3.2 Petrophysical AnalysisLaboratory measurements on core material from the same geological section typicallyshow that:
(i) The sonic travel time (∆T, µs / ft) is proportional to the porosity i.e. the higherthe porosity, the longer the travel time (figure 9)
porosity
xxx x
xx
xx
x
x xx
xx
x
xx
xx
xxxx
xx
x x
x
xxx
xx
x
xx
xx
thic
k w
all c
ylin
der
stre
ngth
(ba
r)or
unc
onfin
ed c
ompr
essi
ve s
tren
gth
(bar
)
porosity
xx x
x
xx
xx
xx xxxx
xx
xxx
x
xx
xx
xxx
xx xx
x
xx
xx
x
xx
xx x
x
xx
xx
soni
c tr
avel
tim
e (
µs )
Cloud of experimental measurements
Cloud of experimental measurements
Average correlation line
Average correlation line
(ii) The rock strength (as measured by an unconfined compressive strength test orthick wall cylinder collapse test - see rock mechanical tests, section 7.3.3) isinversely related to porosity (Figure 9), i.e. the lower the porosity, the greater therock strength.
Figure 9
Porosity is linearly related
to sonic travel time and
thick wall cylinder strength
1
16
N.B. These correlations are in the form of a trend line drawn through a cloud of datapoints. This inexactness is due to the heterogeneity of the formation properties andthe errors in the measurements themselves.
A’
A
B’
B
SAFE RISKTota
l wel
l pre
ssur
e dr
awdo
wn
sinc
e di
scov
ery
A,B producing conditions forwells A and B on initialcompletion
A’, Abandonment conditions forwell A, NO sand failure predicted
B’, Abandonment conditions forwell B, sand failure possible
Sonic travel time (µs/ft)
no failuresand failure.
NB Position of boundary between "safe" and "risk" regions varies from field to field
Data points observed from field experience
The well's sonic travel time log can thus be processed to derive a continuous estimateof the formation strength. This allows the identification of the weakest sandstonewhich can be left (selectively) unperforated. The method can be extended using thefield observation that the vertical stress near the wellbore i.e. overburden minusflowing total drawdown (reservoir depletion + near wellbore pressure drop due to wellproduction) at which continuous sand production is first observed is related to theacoustic travel time (Figure 10). It uses the acoustic velocity as a strength indicatorcoupled with the laboratory finding that the increase in insitu rock stress due topressure depletion of the reservoir as a whole and due to the near wellbore producingdrawdown are equivalent in terms of contributing to sand failure.
The boundary between the “safe region” (where sand production is not expected)and the (risk region) where it may occur - is based on field experience. It varies fromone field to another. It can also be used to predict:
Figure 10
Field experience rock based
approach to sand failure
prediction.
2
Department of Petroleum Engineering, Heriot-Watt University 17
27Unstable Formations and Sand Control7
(i) whether sand production is expected at initial completion
(ii) the earliest time that sand failure can be expected
i.e. when a well's position crosses into the “risk” region. This uses a combination ofthe results from the field reservoir engineering model together with production inflowmodelling of the well itself.
It can be seen that this approach requires a lengthy production history together withactual sand failure in one or more wells. If this is not available; a high drawdown/highproduction rate test can be designed to simulate future producing conditions.
3.3 Rock Strength Measurement
3.3.1 On Site Strength EstimationThe simplest approach is attributed to D. Sparlin who stated that “a potential sandproblem can be expected if the core is friable (finger nail makes groove) or weaker”.This judgement can be extended using table 4 which relates the rock classification tosimple strength observations made on the core to the approximate sonic travel timeand the core recovery. The two strength measurements also included in the table willbe discussed in the next section.
3.3.2 Rock Mechanical Strength MeasurementsA number of rock mechanical tests are used to characterise the rock. They include:3.3.2.1 Unconfined Compressive StrengthAn unsupported cylinder of rock with a length to diameter ratio of at least 2:1 and aplan parallel end is loaded axially at a steady rate of between 0.5 - 5 MPa/min to failure(Figure 11). The compressive failure is in the form of diagonal fractures across thebody of the sample. The maximum stress reached is called the Unconfined or UniaxialCompressive Strength (UCS).
σA
σA σUCSLoad Plattern
End Platten
Unsupported cylindrical rock sample
Time
Load sample at 0.5-5MPa/min
Failure
Failur
e M
ode
Figure 11
Unconfined Compressive
Strength measurement
1
18
3.3.2.2 Brinell Hardness NumberThe Brinell Hardness Number (BHN) is the load required to press a standard sphericalindenter a constant distance into a slabbed core face (Figure 12). The stronger the rock,the greater the load required and the higher the BHN. The Brinell Harness Number(BHN) measurement is a straightforward method of classifying the strength ofsandstone as shown in table 4.
σBHN
Slabbed Core
ONSITE ROCK STRENGTH ESTIMATION
Uniaxial Brinnel Rock Classification Core Observation Approximate Core RecoveryCompressive Hardness Sonic Travel Strength (psi) Number Time
(kg / mm2) (µs / ft)
0 0 Quicksand Hole Slumps >150 zero
<1000 <2 Unconsolidated No apparent cement >145 sleeve1 (poor3)between sand grains conv.2 (Zero)
1000-2500 2-5 Semi-Consolidated Easily Crushed >130 sleeve1 (good) conv.2 (poor3)
2500-3500 5-10 Friable Rub-off Grains 105-130 conv.2 (good)
3500-7500 10-30 Consolidated Crushable with forceps 105-175 excellent
7500-12000 30-50 Moderate Hard Cannot Crush 65-75 excellent
12000-20000 50-125 Hard Cannot Crush 40 / 65 excellent
1 sleeve refers to use of rubber sleeve core barrel to support cored material2 conv. refers to use of conventional steel core barrel (no support for core)3 great care required during coring - prevent core barrel jamming, avoid excessive mud velocities etc.
Apart from the unconsolidated sands (BHN < 2kg/mm2), all other sands in theclassification may exhibit post-failure stabilisation, following the onset of initial sandproduction. The phenomenon will be dependent on the frictional strength characteristicsas well as the degree of cementation. In this context, it is important to note that theBHN is dominated by the degree of grain-to-grain cementation, and by the intergrainfrictional strength to a much lesser extent.
BHN is used because it is linearly related to, but much quicker and easier to measurethan the Thick Wall Cylinder collapse strength (TWC) discussed in the next section.Thus rapid, point measurements can be made at many depths on a slabbed core anda hardness (strength) profile created.
Table 4
On site rock strength
estimation
Figure 12Brinell hardnessmeasurement
2
Department of Petroleum Engineering, Heriot-Watt University 19
27Unstable Formations and Sand Control7
3.3.2.3 Thick Wall Cylinder Collapse StrengthA hollow, thick wall cylinder is created by drilling a narrow hole in the middle of arock cylinder similar to that described in the unconfined compressive strength test. Arubber sleve is mounted around the outside of the cylinder which is then mountedbetween two end caps (figure 13). The hole in the centre of the cylinder is maintainedat atmospheric pressure. The stress applied to the solid end cap and the outside of thesleeve are kept the same and increased at a steady rate of 0.5 - 5 MPa/min. The stressat which the cylinder collapses (implodes) is the Thick Wall Cylinder CollapseStrength (σ
twc).
The TWC test was developed because it replicates of the geometry of perforationfailure (see figure 14)
End Cap
Sleeve
Hollow Cylinder Core Sample
End Cap
Pressure (P) Pressure (P)
Pressure (P)
σtwc
Increase load at0.5 - 5 MPa/min
Time
Pre
ssure
(P
)
Failure
Figure 13Hollow (thick wall) cylindercollapse
1
20
We
llb
ore
Perforation
Perforation
Laboratory measurements, confirmed by field experience, have shown that INITIALsand failure occurs when the near wellbore vertical effective stress at the perforationdepth equals the TWC strength.
The frictional strength characteristics of the rock will depend on grain size distribution,porosity and degree of roundness, which for a specific field, will be related to thedepositional environment. The frictional characteristics will have a major impact onthe TWC strength. In general, sands with a low degree of sorting, and sub-angulargrain geometrics, will exhibit higher strengths than sandstones composed of well-rounded grains, for similar degrees of cementation.
3.3.2.4 Triaxial Rock Strength MeasurementThis is the most sophisticated of the rock mechanical tests that will be discussed here.A cylinder of rock is mounted in a sleeve. The axial stress (σ
a) - imposed by the end
pieces - and the radial stress (σr) are controlled separately. The axial strain is measured
by strain gauges. Figure 15 shows that the maximum axial stress reached prior tofailure increases as the radial stress is increased i.e. the rock sample shows strongerbehaviour as the confinement stress increases. Much more information can be derivedabout the strength properties of the rock from this test compared to those describedearlier - this is illustrated in figure 16, where the various phases of rock failure underinfluence of the imposed stresses are illustrated.
Figure 14Crushing of thick walledcylinder simulatesperforation performance
2
Department of Petroleum Engineering, Heriot-Watt University 21
27Unstable Formations and Sand Control7
Axial strain
Traxial stressesimposed on core
Axialstress
A
r = 20 MPa
r = 10 MPa
r = 5.0 MPa
r = 2.5 MPa
r = 0
r
a
Axial strain
onset formation of small fractures
A B Elastic
B C Elastic-plastic hardening
C Elastic-plastic softening
well developedfracture system(sample fails)
Ax
ial
str
es
s
A
B
C
3.4 Prediction of Downhole Rock Failure(Sand) failure of rock in a producing borehole is a complex process. Characterisationof the downhole properties of the rock is a complex subject that was briefly touchedon in the paragraphs above. Similarly, (sand) failure of rock in a producing borehole
Figure 15Triaxial test strengthmeasurement
Figure 16Elasto-Plastic materialresponse involves theformation of fracturesduring sample testing tofailure
1
22
is complex - a number of possible failure modes are sketched in figure 17. Simple,analytical descriptions of some of the processes are available, but tend to be (highly)conservative. This is because they represent initial failure of the perforation tunnel butdo not capture the re-stabilisation processes such as cavity enlargement which givesreduced flow velocities and drawdowns, which delay the onset of unacceptable sandproduction. Development of complete descriptions of these processes require the useof complex numerical computer code. The actual results are dependant on:
Failuremode :
Shear orCompressive
Tensile Erosion
Cause ofFailure :
Far field stress+ drawdown
Drawdown Flow
FlowFlow
Flow Flow
Flow
(i) the rock failure model used.
(ii) the type of laboratory rock strength measurements made.
(iii) the accurate forecasting of the producing well conditions.
(iv) the availability of core material representative of the failed section of formation.
This latter point is often problematic - the weakest formation is likely to fail first.However, this is the least likely to be recovered by coring!
3.5 A Final WordThe prediction of sand failure from basic principles i.e. when field history data is notavailable is a difficult process. The incentive for doing this is considerable, due to thehigh cost of unnecessary installation of sand control (see next section). Managementof this process by use of a risk based approach to the timing of sand control installationas described is the preferred route - however, when estimating the (Net Present Value)economics of possible schemes, it should include the fact that field experience showsthat the installation of Remedial Sand Control in a well that has already undergonesand failure has a lower chance of success than the equivalent completion operationin a new well.
N.B. Proper estimation of the chance of success of both operations depends on localfield experience!
Figure 17Possible rock failure modesfor a producing borehole
2
Department of Petroleum Engineering, Heriot-Watt University 23
27Unstable Formations and Sand Control7
4. THE COST OF SAND CONTROL
High well productivity losses - typically 60% - are observed after well killing followedby gravel packing operations - see figure 18. This figure summarises the result of afield test in which the well’s productivity index was measured at various stages in thecompletion process. The final well productivity, despite a 6-month clean up period,was only 18% of that achieved prior to gravel packing. This high impairment has beennoted in many gravel packed wells completed in different fields.
31
10
4.42.6
0.7
Stage: perforate kill well gravelpack
produce45 days
produce180 days
% original
PI100 32 2 8 14
PI
= P
rod
uct
ivit
y In
dex
(b
bl/
psi
day
)
40
30
20
10
0
0
wellconstruction
starts
+ve
-ve
Sand failure
extra costinstallation
sand control
Time
Discountedcumulative
cash surplus
wellproduction
started
Sand controlinstalled during
initial completion
No sand failure
reduced (Present Value)due to cost of
remedial sand control installation
delay in remedial sandexclusion installation
Sand failure andremedial exclusion
Low well productivity
Figure 19
Manage risk of sand failure
by delaying instalation of
sand exclusion
Figure 18
Field measured changes in
well productivity
1
24
It implies that it can often be advantageous to delay installation of sand control i.e. ona project cumulative cash surplus basis the reduced cost of completion plus the higherinitial well productivity more than compensates for the later lost production andremedial costs to remedy sand failure (figure 19). The "time effect of money" reducesthe later cost of the installation of sand control measures on a discounted cash flowbasis.
There is a varying efficiency for the different types of sand control that can beinstalled. Typical well productivity and the implications for the various sand controloptions for a West African field study are:
Water Cut 20% 50% 80%
Internal Gravel 5 60 0 5 15 Lowest Pack External Gravel 15 55 0 0 5 Pack None or 30 30 0 0 5 HighestChemical Consolidation
Oil Recovery atField
Abandonment
Gaslift Requirements
(MMscf/d)
No. of WellsRequired For
FieldDevelopment
Average Well Productivity
Index (b/d/psi)
SandExclusion
Type
These different types are illustrated in figure 20. They will be explained in theremaining chapters of this module. However, as can be seen from the table above,installation of sand control:
(i) increases the number of development wells
(ii) increases the capital costs per well
(ii) requires earlier installation of gas lift
(iv) results in higher abandonment pressures and reduced oil recovery
5. SAND EXCLUSION
5.1 Introduction
There are essentially two types of sand exclusion:
(I) Mechanical techniques where “gravel” particles, a few times larger than theformation sand grains, are used to retain the formation in place by forming a filterthrough which the formation sand cannot pass. The gravel is itself held in place bya screen which has been sized so that it in turn can not pass through the gaps (figure20). In its simplest form, the gravel is omitted and the screen alone “holds back” theformation.
Table 5
Example: Impact of
completion efficiency on
Field Development
2
Department of Petroleum Engineering, Heriot-Watt University 25
27Unstable Formations and Sand Control7
(ii) Chemical Techniques where a chemical cement increases the strength of theformation while retaining a permeable pore structure.
The various types of sand control are illustrated in Figure 20. These will all bedescribed in detail in the next chapters. However, one can deduce from this figure thattheir installation is a complex process, which would only be entered into if thealternative, allowing the tubing or casing to fill with sand were not a practicalproposition. The following exercise shows that this is not a practical option; even ifthe sand fill in the tubing is only 1m deep:
Screen
Screen
Cement
Strengthened Zone
Producing Interval
Producing Interval
Perforation
Packer
Gravel
Gravel
Casing
SANDCONSOLIDATION
INTERNALGRAVEL
PACK
EXTERNALGRAVEL
PACK
"FRAC PACK"
Figure 20
Sand control variations
(open hole completions not
illustrated).
1
26
EXERCISE
Estimate the oil flow through a 100 cm long sand bridge in a 3.5 in OD tubing whenthere is a 100 bar pressure drop across it.
This exercise involves the application of Darcy’s Law for single phase, incompressibleflow in porous media. If the gravitational effects are neglected:
Q = k A ∆Pµ L (1)
where :Q = flow rate of fluid (cm3 / sec)A = Tubing Cross Sectional Area (cm2)µ = fluid viscosity (cp)L = Length of Sand bridge (cm)k = Permeability of sand in the bridge (cm2)∆P = Pressure Drop across the sand bridge (atm)
The permeability of most formations are less than 3 Darcies, so it is reasonable toassume the sand bridge permeability of 1 & 10 Darcies. The tubing ID is typically2.6 in or 6.6 cm and we will assume that the oil has the same viscosity as water (1cp). Substituting in Eqn 1 gives:
k / (Darcy) 1 10
Q / (cm3 / sec) 34 342
Q / (m3 / day) 3 30
Even though all the assumptions made have been optimistic in terms of maximisingthe flow through the sand bridge (high permeability and available pressure drop (100bar, low fluid viscosity); the resulting oil production is low. Thus as soon as eithera:
(i) sand bridge is formed or
(ii) a perforation is covered by sand fill
then production is essentially lost from below that point in the well. i.e. Fluid can flowradially into the well through formation sand; but pressure drops immediately becomeunacceptably high when linear flow through sand within the well takes place.
5.2 Different Types of Mechanical Sand exclusion
5.2.1 Open Hole CompletionsThe basic configurations for mechanical sand control are shown in Figure 20. Thesand exclusion system is either installed below the bottom of the lowermost casing orliner (see Figure 20 bottom for an external gravel pack). The advantages anddisadvantages of the various sand exclusion systems are summarised in table 6. Thesystems are described in detail as follows:
2
Department of Petroleum Engineering, Heriot-Watt University 27
27Unstable Formations and Sand Control7
5.2.1.1 Slotted PipeThis consists of steel pipe (e.g. tubing) where a series of parallel slots have been cutthrough the metal (figure 21). The width of these slots are normally made as small asmechanically practical so that they will retain as large a fraction of the formation sandas possible. The inflow area is low (2-3% of pipe surface area). Sand grains whichare small enough to pass through the slot can still form a stable arch around the slotin a similar manner to that described in section 2.1.3 where arch formation around aperforation was discussed. It is mainly used as a low price option to reinforce an openborehole and to retain a coarse grained formation; although narrower, laser cut slottedpipe is now becoming available (expensive).
ITEM SLOTTED LINER WIRE WRAPPED SCREEN PRE-PACKED SCREEN(Mild Steel) (Stainless Steel) Resin Coated Sand
Description rectilinear slots / wire wielded to longitudinal rods gravel sandwiched between two machined in pipe wire wrapped screens
Concept wellbore reinforcement; formation sand exclusion or gravel gravel provides sand exclusion sand bridges around slot retention
Material mild steel stainless steel on mild steel base stainless steel on mild steel base pipe pipe
Sand exclusion poor: 0.012” slot width better than Slotted Liner since excellent: as with gravel packminimum slot width 0.006” - 0.040”
Works with yes yes yes, but should not be necessarygravel pack
Flow restriction high low, ≈ 10 times flow area of high, as for wire wrapped screenslotted liner
Mechanical good poor to collapse / tension if base fair: base pipe reinforces structureresistance pipe omitted. Also susceptible to
erosion
Plugging low (too wide to retain moderate high: fines + mud cake. Also tendency formation sand) impairment while running in hole.
Cost cheapest 2-3 x slotted liner 2-3 x wire wrapped screen, butoften less than gravel pack
Application borehole reinforcement higher productivity wells medium retains sand grains of all sizes
coarse grained formation grained formation.allows fines production
5.2.1.2 A Wire Wrapped ScreenThis consists of a triangular shaped wire which is carefully wound so that there is aconstant gap between successive turns (figure 22). It is held in place by spot weldingthe wire to vertical formers placed at 1cm intervals around the internal diameter of thescreen. Wire wrapped screens have the advantage over a slotted liner that the gapbetween the wires can be made smaller and be held to the target value with a muchgreater accuracy; allowing the screen to retain finer grained formations than theslotted liner.
Figure 21
Slotted pipe used for Sand
Control
Table 6
Comparison of Liner and
Screen Characteristics
1
28
Vertical former (for strength)
Stainless steel wirein triangular shape
Gap kept constant
Key form allows anysand grain that passes narrowest point to be flushed from slot
Spot weld
Wire wrapped screens also have a much greater inflow area - making them moresuitable for higher productivity wells with a greater inflow rate per unit completionlength. Wire wrapped screens have to be handled carefully at the rig site - theirstrength is much lower than slotted pipe (figure 22 omitted the perforated base pipewhich is often used to increase the screen’s strength. This base pipe has been includedin the cross sectional view in figure 23).
They are ideally suited for use in higher productivity wells where it is required to retaina medium grain sized formation.
connection
perforatedbase pipe
triangularwire
Figure 23
Cross section of wire
wrapped screen showing
perforated base pipe inner
support
Figure 22
Wire wrapped screen
2
Department of Petroleum Engineering, Heriot-Watt University 29
27Unstable Formations and Sand Control7
5.2.1.3 Resin Coated Sand Pre-Packed ScreenPre-packed screens are constructed from two concentric screens with a layer of gravelplaced in between them (figure 24). The gravel had been coated with a layer ofthermosetting resin. The construction process is as follows:
(i) the dual concentric screens have been welded onto the base pipe
(ii) the gap between them is filled with the resin coated sand and the final weldsmade
(iii) the completed screen is placed in an oven where the thermosetting resin,hardens creating a strong ring of gravel
The pore throats of the consolidated gravel provide a series of narrow openings whichprovide the sand exclusion and retain the formation in place. The presence of thegravel with its narrow pore throat diameter provides a greater flow restriction than thewire wrapped screen alone; as well making the screen susceptible to plugging byformation fines etc. The greater complexity of the prepacked screen increases thecost.
������������������������
yyyyyyyyyyyyyyyyyyyyyyyy
Connection
Perforated BasePipe for Strength
Gravel Consolidated in Placeby Chemical Cement
(Large enough thatit cannot pass through the gaps in the screen)
OuterScreen
Inner Screen(wider gap than outer screen)
����������������
������������
���������
������������
��������
������������
������������
������������
��������
������������
������������
���������
������������������
������
������
���������
����������
������������
������������
������������
��������
����������������
������������
���������
������������
��������
������������
������������
������������
��������
������������
������������
���������
������������������
������
������������
��������
������������
������������
������������
��������
Formation Sand too smallto pass through pores in gravel)
5.2.1.4 Application of Open Hole CompletionsOpen hole completions have become much more popular since horizontal wellsbecame widespread. Their use avoids the high cost and technical difficulties incementing and perforating long (up to 2km) horizontal liners and casings. Some ofthe problems - and opportunities - presented to the completion design engineer whendesigning horizontal well completions are illustrated in figures 25 - 28.
Figure 24
Pre-packed screen
1
30
13.375"
9.625"
7"Pre-perforated liner
13.375"
9.625"
5.5"wire-wrapped screens
9.625"
ECP
7"Slotted liner
stage cementing collar
Figure 25 shows a “conventional” well design with the production casing cementedin place just above or just into producing formation. The 7” perforated liner providessupport to the open hole but does not provide sand control. Sand control is providedin the completion shown in figure 26 in which the 7” pre-perforated liner is replacedby a (smaller) 5.5” diameter wire wrapped screen. This change could increase the wellcost by typically 10 - 20%.
Formations consisting of coarse sand grains allow this well cost to be decreased bysubstituting a 7” slotted liner for the wire wrapped screen (figure 27). The well costcan be further decreased if the use of an external casing packer possibly backed up bycement placed via a stage cementing collar can give sufficient (zonal) isolation of theproducing interval from shallower formations. This low cost option gives a typicalwell cost of only 60% of that for above wire wrapped screen case.
Figure 25
Conventional completion,
no Sand Control
Figure 26
Conventional completion
with wire wrapped screen
for Sand Control
Figure 27
Low cost option
incorporating sand
exclusion
2
Department of Petroleum Engineering, Heriot-Watt University 31
27Unstable Formations and Sand Control7
Tubing retrievableSurface controlled subsurface safety valve
51/2" Tubing
41/2" Tubing
7" SlottedLiner
81/2" Horizontal hole
95/8" Packer (ca 60˚)
95/8" Shoe (90˚)
Gas lift valve in side pocket mandrel
Sliding side door (allows circulation)
Wirelinenipples
Polished bore andelastomer seals
Permanent guages with surface read out
ECP ECP
A full Horizontal Well Completion is pictured in figure 28.
It incorporates (from the top):
• tubing retrievable surface controlled, sub surface safety valve (SCSSSV) so as tobe able to close the well in an emergency
• 51/2" tubing reducing to 41/
2" tubing in the bottom section - where the accessories
(gas lift valves etc.) are installed. The use of 51/2" tubing maximises the flow
capacity while a 51/2" accessory would have too great a diameter for the casing.
Hence 41/2" accessories (and tubing just above the packer) are used.
The accessories consist of:
(i) gas lift valve installed in a side pocket mandrel
(ii) sliding side door which can be opened to allow fluid circulation between thetubing and the tubing / casing annulus
(iii) permanent gauges connected to surface for continuous monitoring of downholeproducing conditions
(iv) wireline nipples where plugs or other devices can be placed
(v) polished bore and elastomer seals. This allows tubing expansion or contractiondue to the temperature changes to be taken up. The tubing can also be (partially)retrieved to replace the SCSSSV as required without removing the packer.
The open hole section is completed with a 7" slotted liner in a 8 1/2" drilled hole.
External casing packers are included at regular intervals so that sections of thecompletion interval can be isolated at a later date to stop extraneous gas or waterproduction. Removal of the ECPs allow a gravel pack to be placed (Figure 28a).Gravel packing of long horizontal wells is a technology that has now matured.
Figure 28
Horizontal well completion
1
32
5.2.1.5 Enhanced Drilling Fluid Requirements for Open Hole Completion’sOpen hole completion’s with sand exclusion place special requirements on the drillingfluid properties. This is illustrated in figure 29a which is a schematic illustration ofthe results of large scale experiments in which a mud cake was deposited in a simulatedsection of a horizontal well. An acidisation treatment removed the mud cake coveringthe upper part of the hole but left extensive mud cake remnants on the lower sections.Simulated production from the reservoir into the well (figure 29b) resulted in morethan half the screen area being covered by mud cake, reducing fluid flow.
Remedial treatments to dissolve mud cakes are a poor option since:
ScreenMudcake fallen on
top of screen
Mudcake
ScreenFormation
(Partially dissolved)mudcake deposited at bottom
of horizontal hole after becoming detached from formation face following breaker treatment
(i) they are expensive
(ii) they tend to destabilise an already weakly / un-consolidated formation resultingin permeability impairment
(iii) It is difficult to ensure that the complete length of the open hole interval istreated with the dissolving fluid.
Remedial treatments can be avoided and higher productivity completions achieved byselecting the drilling fluid so that:
(i) the mud cake is easily and evenly lifted from the borehole when the well is
Figure 29a
Mud/mud cake deposits on
lower side of horizontal well
section observed after
displacement/breaker
treatment
Figure 29b
Screen appearance after
breaker treatment to remove
mud cake
2
Department of Petroleum Engineering, Heriot-Watt University 33
27Unstable Formations and Sand Control7
placed on production. Drawdown of a horizontal well is often low and decreasesfrom the heel of the well towards the toe. ONLY if the “filter-cake-lift-off” pressureis low will it have a chance of being removed from the complete length of the well.
(ii) the mud cake should break up and flow easily between the narrowest gaps inthe installed sand exclusion equipment.
This process can be accelerated by choosing (some of) the mud cake constituents tobe soluble in the produced fluids (water or oil). Specialist drilling fluids are nowavailable from the service companies that meet these requirements.
5.2.2 External Gravel PacksAll the options described above just used only the screen or liner as the basis of thecompletion. It was inserted into the open hole and the gap between the screen or linerand the borehole wall remain empty. Once the well was placed on production the mudcake should be produced through the screen as described above. The behaviour of theformation will depend on its strength:
(i) strong formations: borehole wall remains intact and gap between the liner andthe screen remains empty
(ii) weak formations: borehole collapses and the original liner/borehole gapbecomes filled with failed formation material.
An alternative is the under reamed, external gravel pack introduced earlier (bottomFigure 20 and Figure 30). This involves enlarging (by typically 10 -15 cm) the gapbetween the sandface and the screen using an under reamer. The under reamerreplaces the drill bit at the bottom of the drill string and consists of a series of arms withcutters at the end which expanded so as to enlarge the hole. When the drill string isrotated the hole enlargement is carried out with a non damaging fluid i.e. the filter cakewill decay or not prevent the flow of oil or gas from the enlarged borehole diameter . Theenlarged liner borehole gap is completely filled with gravel, a process known as gravelpacking. This is described in the remainder of this module.
1
34
Formation sandFormation sand
gravel
under reamedhole section
slotted liner orwire wrap screenretains the gravel
cement
packer
productiontubing
productioncasing
casing shoe
5.2.3 Internal Gravel PackThe screen or liner is placed inside a cased hole for an internal gravel pack with gravelbeing placed in the screen / casing annulus and in the perforations
An internal gravel pack is illustrated in middle of Figure 20 (right hand side) andenlarged as Figure 31
Scr
een
orlin
er
Gra
vel-f
illed
annu
lus
Larg
e D
iam
eter
Per
fora
tion
pack
ed
With
Gra
vel
Cas
ing
Cem
ent
Form
atio
n sa
nd
������������
���������������
������������������
������
������
���������
������
������������
������������
���������
������������������������
������������������
���������
������������
���������������
���������
���������
������
������
������
������
������
������
������
������
������
������
������
���������
���������
������������������������
������������������
������������
���������������
������
���������
���������
������������������������
���������
���������
������������
��������������������
���������
�����������������������
��
������
������
���������
���������
������������������������
������������������
���������������
������������
������������������������
��������
������������
���������
������
���������������������
���������������������
���������������
������������
������������
���������������
������������
���������
������
������
������
����
����
������������
������������
�����������������
���������
������������
���������������
����������������
���������
������������
���������
���������������������
���������
������������
������������
���������������
���������
���������
������������
����������������
���������
���������
����������������
���������������
������
������
������������
������
���������
���������
���������
������������
����������������
���������
���������
������������
���������������
���������
������������
���������
���������
������������
���������
���������������
���������
������������
���������������
���������
������������
������������
���������
������������
���������
���������������
������������
���������
���������������������
������������
���������������
���������������
������
���������
������
���������������
���������������������
������������������
������������
������
���������������
������
���������
���������������
������
������
������
������
������
���������
���������
���������
������������
���������
������������������������
���������
�������������������
���������
��������������
������������
���������
����������
��������������
���������
������������
���������������
������������������
������
���������
��������
������
���������������������
������������
������������
�����������������
������
���������
������
��������������
���������������
��������
������
��������
���������
��������������
�������������
���������
������
������������
���������
������������
���������
�����������
���������
������������
��������������
���������
������������
��������������������������������������������������������������������
������
���������
������
���������
yyyyyy
zzzzzzzzz
{{{{{{
|||||||||
Figure 30
Detail of an external
gravelpack
Figure 31
Detail of internal gravel
pack
2
Department of Petroleum Engineering, Heriot-Watt University 35
27Unstable Formations and Sand Control7
5.2.4 Special Gravel PacksThe combination of an internal gravel pack and a short, highly conductive, hydraulicfracture stimulation has become very popular in the Gulf of Mexico (USA). This isknown as “Frac and Pack” -an example is illustrated in the middle of Figure 20 (LeftHand Side).
A second specialist option is the use of “Resin Coated Gravel”. This product is similarto that used for making the pre-packed screens - The resin coated gravel sets into aconsolidated mass under the influence of the Well’s Bottom Hole Temperature.
5.3 Advantages / Disadvantages of Gravel PackingThese have been summarised as table 7 which is self explanatory when read inconjunction with the remainder of chapter 5.
Advantages Comment
Effective over long intervals >100 m operationally possible
Copes with varying rock properties Clay / Silt particles can pass through
gravel pack if large enough
Employs simple, non-toxic materials Gravel placed with water based fluids
Disadvantages
Mechanical restriction in wellbore e.g. Production logging not possible
Workover difficult Requires removal (fishing) of packers /
screens etc.
Expensive for multiple intervals Consider co-mingled production
Diminished workover options Difficult to identify source as well
as shut off undesirable water and gas
Sensitive to poor completion practices Many points where permeability damage
can be created
5.4 Gravel Pack Sand SelectionGravel pack sand is a well rounded, clean (minimum fines and acid soluble material)that has been sieved so that its size falls between carefully selected size ranges. If sizedcorrectly, it acts as a high permeability (i.e. not restricting flow of oil or gas) filterwhere the pore throats between the gravel grains are small enough to restrict thepassage of the formation sand grains. This is illustrated in figure 32.
Table 7
Gravel Packing
Considerations
1
36
gravel
formationsand
Too Large Gravel• Formation sand produced• Gravel pack failed
Too Small Gravel• Formation sand retained, but• low permeability of gravel creates unnecessary extra loss of well productivity
The formation sand is characterised by breaking down any consolidated, multi-grainparticles into the individual sand grains. These are then passed through a series of pre-weighed sieves of progressively smaller mesh size and the weight of sand trapped oneach sieve recorded. The cumulative distribution is then plotted in the manner shownin figure 33 using log-linear graph paper. The size corresponding to a cumulativeweight percentage of 10%, 40% and 90% is known as D
10, D
40, and D
90 respectively.
100
90
80
70
60
50
40
30
20
10
00.050.01 0.050.10.51.0
Grain Diameter (inches)Coarse Fine
Cu
mu
lati
ve W
eig
ht
Pe
rce
nta
ge
D90
D50
D40
D10
The sand grain size distribution is characterised by the Uniformity Co-efficient (C)which is defined as:
C = D40
/ D90
Figure 33
Sieve analysis of typical
formation sand
Figure 32
Effect of incorrect gravel
pack sand selection
2
Department of Petroleum Engineering, Heriot-Watt University 37
27Unstable Formations and Sand Control7
with formation sands being classified as:
C<3 well sorted, highly uniform sand3<C<5 uniform sand5<C<10 moderate/poorly sorted sandC>10 poorly sorted highly non-uniform sand
Examples of a well sorted (C≈2) and poorly sorted sand (C≈8) are compared in figure 34.
100
90
80
70
60
50
40
30
20
10
00.01
Sand Grain Diameter (cm)
We
igh
t (P
erc
en
t) 1
2
100
90
80
70
60
50
40
30
20
10
00.01
Sample 1 is wellsorted (C=2)
Sample 2 is poorlysorted (C=8)
0.1
Sand Grain Diameter (cm)
Cu
mu
lati
ve W
eig
ht
(Pe
rce
nt)
1
2
Figure 34
Example of well and poorly
sorted sand distribution
1
38
The gravel pack sand may be selected once the formation sand has been characterised.There are a number of criteria used - the most common being the Saucier criteria.
D50
{Gravel} = 6 * D50
{Formation Sand}
However this relationship makes no allowance for the sand uniformity. For poorlysorted sand, a second criteria attributed to Schwartz should also be examined:
D10
{Gravel} = 6 * D10
{Formation Sand} for C<5D
40 {Gravel} = 6 * D
40 {Formation Sand} for C<5
D70
{Gravel} = 6 * D70
{Formation Sand} for C<10
However only a limited range of gravel packsand sizes are available commercially (table 8)
N.B. Tight (cubic) packing of the gravel is required for the Schwartz or Saucierselection criteria to be effective (figure 35).
gravel
formationsand
d = 15% DCubic packing required
d = 42% DRectangular packing
undesirable
DD
d
d
In all cases the wire wrapped screen is selected so that the gap is equal to 0.5 times thesmallest gravel size. This criteria is often not practical for slotted liners - in which casethe smallest available possible slot size should be chosen.
100
90
80
70
60
50
40
30
20
10
0
Grain size
fine grainedlaminae(low permeability;non-productive)
range forproducing
sand formation
d50{gravel} = 6*d50{formation}
Coarse Fine
Cu
mu
lati
ve W
eig
ht
Pe
rce
nta
ge
range of total completion interval
Figure 35
Gravel selection criteria
only works if a tight pack
can be created
Figure 36
Typical gravel / formation
sand size disrtibution
2
Department of Petroleum Engineering, Heriot-Watt University 39
27Unstable Formations and Sand Control7
Frequently, several sieve analyses are available from formation sand samples takenfrom different depths in a particular completion interval. Engineering judgementneeds to be exercised - a typical scenario that may be encountered is sketched infigure 36. Possible changes in the permeability of the gravel / sand combination onceflow has commenced are shown in figure 37. Smaller gravel forms an effective filterfor the sand - but has a low permeability due to its small grain size. Invasion of thegravel by the formation sand begins to occur once the D
50 size is greater than 7. The
permeability of the sand / gravel mixture can actually become less than that of theformation sand if large scale mixing takes place.
1.0
0.8
0.6
0.4
0.2
00 2 4 6 8 10 12 14 16 18 20
well productivityreduced by usingtoo small gravel
sand producedcompletely
through gravel
gravel acts aseffective filter
Formation Sand invades gravel;50:50 mixture
has lower permeabilitythan formation sand alone
Pe
rme
ab
ility
(fi
na
l)P
erm
ea
bili
ty (
init
ial)
IDE
AL
D50 {gravel}D50 {sand}
Formation
Formation sand permeability
As the gravel increases in size (D50
> 17), the gravel pore throats become sufficientlylarge that the formation sand can pass through with minimal impact on the gravelpacks permeability.
The properties of the standard gravel pack sands are listed in table 8
Properties of Standard Gravel Pack Sands
US Mesh Range Median Gravel Typical PermeabilityGravel Size Diameter(µm) (µm) (D)
40 / 60 425 - 250 340 5520 / 40 850 - 425 640 17012 / 20 1700 - 850 1275 600
N.B. 16/30, 30/50 and 50/70 gravel can be made available to special order (higher cost)
5.4.1 Operational ConsiderationsThe choice of gravel size influences the points in the operation at which thepermeability of the pack can be damaged e.g. smaller gravel is more prone to
Table 8
Properties of standard
Gravel Pack sands
Figure 37
Schematic diagram of sand
/ gravel interaction
1
40
plugging by dirty completion fluids but is less prone to plugging by fines duringproduction.
The specification for the gravel used for gravel packing operations is laid down in a"Recommended Practice" by a working committee of the American PetroleumInstitute (API). It must not only be carefully sieved so that 98% out of the sample fallsbetween the maximum and minimum specified sieves, but also the source of the gravelis selected to meet minimum roundness, sphericity, grain strength criteria along witha maximum acid solubility level and percentage clay. Synthetic “gravel” is alsoavailable at a premium price - the grains are stronger (less permeability impairing finesproduced during pumping due to grain breakage) and are more spherical (higherpermeabilitys - typically 25% greater for the same nominal grain size).
800
700
600
500
400
300
200
100
00 25 50 10075
Pre
ssur
e dr
op a
cros
s pe
rfor
atio
n (p
si/p
erf)
Flows Rate (bbl/d/perf)
Perforation filledwith 0.5 Darcy formation sand
Perforation filledwith 180 Darcy gravel
Cross sectional areaof perforations
Diameter, in. Area, in.2
0.110
0.196
0.442
0.301
0.785
Perforation2 in long
fluid data µ = 1 cpsg = 0.8β = 2.3 x 104
3 "8
1 "2
3 "4
1 2
3 8
3 47 8
1
Gravel permeability is a key issue for internal gravel packs due to the high pressuredrops potentially associated with flow through the gravel filled perforation penetratingthe casing and cement. This is illustrated in figure 38 where the pressure dropcalculated by Darcy’s law is plotted as a function of flow rate for a cp oil flowingthrough different size perforations filled with 180 Darcy gravel. Minimal flow ratesare achieved through the perforation if formation sand replaces the gravel.
The completion engineer is in charge of specifying the perforation density (numberof perforations per metre) and the perforation diameter (design and weight ofexplosive for the perforating charges). The perforation programme should be chosenso that the gravel filled perforations do not limit the oil/gas inflow into the well.
5.5 Gravel Packing - Surface OperationsThe surface operations occuring during a gravel packing treatment are summarised inFigure 39
Figure 38
Importance of perforation-
tunnel size and filling
material
2
Department of Petroleum Engineering, Heriot-Watt University 41
27Unstable Formations and Sand Control7
A 1% wt potassium chloride (KCL) brine base fluid is used “as is” or is viscosifiedby the addition of a polymer e.g. hydroxyethyl cellulose. This base fluid is filtered toremove impairing contaminants and gravel added to the required concentration (lowconcentrations of gravel are added continuously (“on the fly”) while pumping; highconcentrations can be pre-mixed in the stirred vessel when viscous fluids are usedsince they have sufficient viscosity to retain the gravel in suspension). The slurry isthen pumped into the well using the high pressure pump.
Viscous or non-viscous fluid
base fluid (1% wt KCl brine)
Pump
Highpressure
pump
10µ 2µFilter unit
Gravel Filtered fluid
Pump to well
Polymer
Paddlemixer
5.6 Fluids for Gravel PackingThe properties of the gravel packing fluid can range from being similar to water tohighly viscous. In both cases the fluid performs functions described in table 9.
Function Comment
Control formation pressures Similar to drilling fluids
Transport gravel to completion BOTH at surface and downholeinterval Return (reverse) excess gravel As abovefrom completion interval to surface
“Clean” fluid essential Filter solid particles in water and (undissolved) polymer residue
Compatible with clay particles Add 1% wt KCl in formation
Viscosity of viscous fluids must Add chemical breaker (acid or oxidising agent)degrade to allow production to surface
Figure 39
Surface equipment for
batch gravel packing
operation
Table 9
Gravel Packing Fluid
Functions
1
42
5.6.1 Properties of Viscous Gravel Packing FluidsThe viscosity behaviour of a typical viscous gravel packing fluid {a dilute (80 lbspolymer / 1000 gal water) solution of hydroxy ethyl cellulose polymer in brine} isshown in Figure 40 where the viscosity is plotted as a function of shear rate. Two typesof behaviour are shown:
0.1 1 10 100 1000 100000.01
0.1
0.01
1
10
100
Vis
cosi
ty [P
a.s
]
{ { {{ {a settling particle drill pipecasingtubing
annulus formation
Newtonian behaviour at low shear rate
shear rate regime for:
Shear rate [s-1]
"Power" law regime
(i) A low shear rate “Newtonian” region where viscosity is independent of shear rate
i.e. µ = τ / δ (2)
where µ = coefficient of viscosity, τ = shear stress and δ = shear rate
(ii) A higher shear rate region where “power law” behaviour is shown i.e.
= k
= K -1µ δδ
δη
η∴ (3)
where = power index and k = consistency indexη
Newtonian fluids have a power index of 1, hence equations (2) and (3) becomeequivalent.
The slope of the log viscosity against log shear rate plot shown in figure 40 is the powerindex η. The typical shear rate region which is encountered in various parts of the wellare also indicated in the figure.
Figure 40
Hydro Ethyl Cellulose
(HEC) polymer viscosity as
a function of shear rate
2
Department of Petroleum Engineering, Heriot-Watt University 43
27Unstable Formations and Sand Control7
5.6.2 Other Base Brines1% wt KCl solution is the standard brine since it is compatible with the majority offormations {i.e. no permeability damage}. Geopressured formations require the useof higher density brines - see table 10
Completion Brines for Geopressured Formations
Chemical Formula Maximumspecific gravity
Ammonium Chloride NH4Cl 1.13
Potasium Chloride KCl 1.17
Sodium Chloride NaCl 1.2
Potasium Bromide KBr 1.31
Sodium Formate HC00Na 1.33
Calcium Chloride CaCl2 1.42
Sodium Bromide NaBr 1.54
Potasium Formate HC00K 1.6
Mixed Calcium Bromide/ Calcium Chloride CaBr/CaCl2 1.7
Calcium Bromide CaBr 1.87
Mixed Zinc Bromide/ Calcium Bromide ZnBr2/CaBr 2.4
(g.cm3)
The brine chosen will depend on the density required, cost, compatibility with theformation and well equipment, crystallisation temperature (for North Sea or otherlower temperature operations) as well as health and safety issues for the wellsite staff.N.B. A range of formate salts is being introduced to cover the same density range.Formate salts have an improved environmental profile compared to the higherdensity, bromide based brines.
5.7 Fluid Loss ControlGreat care is taken to prepare the well in an unimpaired state prior to the start of thegravel pack operation. This leads to the result that even a small overbalance of thecompletion fluid’s hydrostatic head compared to the reservoir pressure leads to largescale losses of completion fluid into the formation. This is undesirable since it oftenleads to extensive formation permeability impairment. Conventional lost circulationfluids, as employed during drilling, also lead to formation impairment. Four optionsare available:
(i) placing a small volume of a very high viscosity fluid across the perforated zone.This either contains a viscosity breaker (which will degrade the viscosity after aspecified time/temperature exposure period) or it is removed by pumping acidimmediately before (preferred) or after the gravel pack has been completed.
(ii) as above with filter cake forming solids (oil soluble resin or water solublesodium chloride crystals) which will dissolve once production commences.
(iii) mechanical solutions such as large bore flapper valves (figure 41).
Table 10
Completion Brines for
Geopressured Formations
1
44
The valve is run into the well in the open position - being held open by, for example,the wash pipe of the gravel pack assembly (see section 5.8). On withdrawal of thewashpipe, the spring holds the flapper valve closed against the seal; thus preventingfurther brine losses from the tubing to the formation. The valve is constructed fromglass or ceramic material so that it can be broken by pressure or mechanical impact.
Flapper
Top View
Side View
Spring
Spring
Seal
Connection
Tubing
(iv) Formation Inflow Valve (FIV) figure 42. This is a large bore adaptation of a drillstem test valve - it can be open and shut by either annulus or tubing pressure. The FIVis placed near the bottom of the tubing string. A seal assembly is placed at the bottomof the tubing string. The FIV is large enough for perforating guns mounted at the endof a work string to be run through it. The sequence of operations is as follows:
cemented line
packer(not yet set)
polished borereceptical attop of liner
formation perforatedthrough tubing
cement
cement
seals ontubing
formationinflowvalve
(a) Open FIV by pressurising annulus
(b) Perforate completion interval with tubing retrievable perforating guns.
(c) Recover perforating guns to above FIV.
(d) Close FIV by pressurising annulus to stop completion fluid losses.
Figure 41
Large - bore flapper valve
Figure 42
Horizontal well completion
schematic
2
Department of Petroleum Engineering, Heriot-Watt University 45
27Unstable Formations and Sand Control7
5.8 The Gravel Pack OperationFigure 43 schematically illustrates the main stages of a gravel pack operation. Thisfigure illustrates the process for an EGP; the process being the same for an IGP. Thewire wrapped screen or liner is placed across the completion interval with a length ofnormal tubing above it. A specialised gravel pack tool, called the “cross-over tool”,and a packer are mounted above the tubing. The “cross-over tool” allows variouscirculation paths from the tubing to the annulus to be selected. The operation is asfollows:
(a) Circulating gravel
gravelpack
washpipe
productioncasing
"cross-over"tool
"cross-over"tool
(b) Screen Out
gravelpack
(c) Reverse Out Excess Gravel
gravel slurry
�����������������������������������������������������������������������������������������yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy
screen
(i) Gravel slurry is pumped down the tubing and “crosses-over” into the liner /casing annulus. The gravel falls to the bottom of the hole where it builds upwards.The fluid flows through the liner, up the wash pipe and “crosses-over” so that it canreturn to the surface via the casing / tubing annulus (figure 43a).
(ii) The pressure rises rapidly once the (de-hydrated) gravel level has covered thetop of the liner since flow through the gravel leading to a much higher pressure drop.This is called a screen out (figure 43b).
(iii) The gravel pack tool is raised to allow circulation directly from the tubing to thetubing / casing annulus (the flow paths to the gravel pack itself are disconnected).Reverse circulation (DOWN casing/tubing annulus and UP the tubing) allows theexcess gravel slurry to be recovered at the surface. (figure 43c).
Figure 43
The gravel pack operation
using a crossover tool
1
46
The “cross-over” tool, packer and work string are recovered so that the final wellcompletion tubing can be run.
5.9 Gravel Placement with Low Viscosity FluidsThe manner in which the gravel pack is formed depends on the viscosity of the fluidand the deviation angle that the well is drilled through the completion interval. Thekey points concerning low viscosity fluids are summarised as follows:
(i) Gravel placement with low viscosity fluids is mainly applied to external gravelpacks or shallow formations - The low gravel concentrations (2lb gravel / gal fluid)and the low frictional pressure losses minimise the chance of fracturing these weakformations.
(ii) the time required to pump all the gravel and complete the gravel pack is longdue to this low gravel concentration, despite the high pump rate of 5-10 bbl/min.
(iii) This high pump rate can cause permeability impairment due to gravel/sandintermixing (see section 5.4).
(iv) the large volumes of fluid pumped during the operation through the screen/linerincreases the opportunity for liner or screen plugging.
Large scale laboratory tests have shown that the gravel pack is formed from the bottomupwards in vertical or low deviation (<45º) wells - see figure 44.
������������
yyyyyyyyyyyy Crossover
Packer
6
7
5
4
3
2
1
Wire-wrappedScreen
UpperTell Tale
Wash Pipe
Pac
king
Seq
uenc
e
The process that occurs in perforations - dune formation - is more complicated.(Figure 45.) The gravel is deposited at the mouth of the perforation since it cannot beheld in suspension by the slow moving fluid (leak off rate per perforation is low). Thedune builds up until the gap between it and the top of perforation is so narrow that thefluid velocity has increased sufficiently to carry the gravel further down the perforation.
Figure 44
The gravel packing process
with low viscosity fluids
2
Department of Petroleum Engineering, Heriot-Watt University 47
27Unstable Formations and Sand Control7
The dune can then develop down the perforation until the perforation tip is reachedas shown. The final packing stage is then backwards from the tip to the mouth of theperforation.
1 2 3 4 5 6 7 8 9
1110
Flow
Casing
Cement
5.10 Gravel Placement with High Viscosity FluidsInternal Gravel Packs are usually placed with high viscosity fluids. The high viscosity(up to 300 cp) carrier fluids allows the gravel to be suspended at high gravelconcentrations (up to 15lb gravel/gal carrier fluid or 50% vol). Also the resultingslurry can be pumped at low rates (0.5 - 1.5 bbl/min). This gives considerableadvantages which should lead to higher productivity completions:
(i) limited fluid volume → reduced screen plugging
(ii) low pump rate → no gravel/sand mixing
The gravel packing process is quite different with high viscosity fluids:
(i) gravel slurry dehydration is initiated at the perforation tip followed by formationof a node once the perforation is filled (figure 46). The process is driven by fluid leakoff into the formation.
1234567Flow
Casing
Gravel Node Formsat Perforation Mouth
Cement
Formation
(ii) simultaneously, a sheath of dehydrated gravel is formed around the screen - theprocess being driven by fluid flowing circulating through the screen and returningto the surface via the washpipe (figure 47).
Figure 45
The perforation packing
process with low viscosity
fluids
Figure 46
The perforation packing
process with viscous fluids
1
48
CrossoverPacker
8 9
6 7
54
3
21
Gravel Nodes at Perforation Mouth
Gravel Sheath Around Screen
Wire-wrappedScreen
Wash Pipe
(iii) The remainder of the gravel pack is then completed from the bottom upwards(figure 47). Once the gravel pack is completed as far as the top of the wire wrappedscreen, the pressure increases rapidly a screen out having been achieved. Productionmay be established once the gravel pack operation is completed.
(iv) A chemical breaker has to be mixed with the viscous fluid to degrade theviscosity so that production may be established once the gravel pack operation iscomplete.
(v) more difficulty is experienced in obtaining complete gravel packs when:
(a) the completion interval contains large permeability contrasts
(b) the completion zone is long (>15m)
(c) the deviation angles are high (>50º)
Use of special equipment changes - e.g. of a large diameter wash pipe - which helpdistribute the slurry across the complete completion interval can improve the gravelpack quality.
5.11 New Technology5.11.1 Gravel Pack EvaluationSeveral of the (nuclear) density logs can be used for gravel pack evaluation. Intervalswith an incomplete gravel pack give a lower shallow density reading than intervalswith a good pack. However, the readings are also influenced by changes in thecompletion equipment e.g. connections, packers, transition from wire wrapped screento tubing etc. These effects are schematically illustrated in figure 48.
Figure 47
The gravel packing process
with viscous fluids
2
Department of Petroleum Engineering, Heriot-Watt University 49
27Unstable Formations and Sand Control7
��yy��yy
������
yyyyyy5000
Top of Gravel
100% Complete Gravel Pack
Inomplete GravelPack
100% Complete Gravel Pack
ScreenPerforations
CasingCollarLog
Gravel
Density Log Reading
5.11.2 "Frac and Pack""Frac and Pack" was introduced in section 5.2.4. Field experience, particularly in theU.S. Gulf Coast, has shown that this combination of hydraulic fracturing and gravelpacking leads to completions with a lower “skin” and hence much higher wellproductivities (figure 49). The hydraulic fracture pierces the ring of formationpermeability impairment around the well by providing a high conductivity channel forthe flow of oil or gas figure 50. The (vertical) hydraulic fracture also ensures that thewell is connected to all thin, laminated sands, some of which might be missed by aconventional gravel pack.
• Gravel Pack Skins Range - 1-300+Average Gravel Pack Production Efficiency < 25%
• Frac - Pack Skins Range - 4 to 27Average Frac - Pack Production Efficiency < 95%
Gravel Packs
Frac - Packs
Ski
n V
alue
65
55
45
35
25
15
5
-50 10 20 30 40 50 60
Wells
Gulf Coast, USA, Frac - Pack Field Results Based on Well Build Up Tests
(36 Frac - Pack, 30 Gravel Pack)
Figure 49
"Frac Pack" results
Figure 48
Gravel pack evaluation with
a nuclear density log
1
50
Screen
Screen
Gravel Pack
DamagedZone
DamagedZone
100D
10mD
0.1mD
Propped Fracture
"Frac and Pack" fractures are short and fat, compared to the long thin fractures usedfor stimulating low permeability formations. This is because formations requiringsand control normally have a high permeability. The background to these statementswill be explained in the hydraulic fracturing module.
5.11.3 New Screen TechnologyIt was discussed earlier that high flow rates have often been observed to trigger sandinflow problems (figure 2). This has lead to a large effort has been made to developscreens with an increased area open to inflow from the formation i.e. screens that willminimise the inflow velocity. Figure 51 compares the inflow area for a perforatedcasing with that of various types of sand control screens and the ideal completion (anopen hole).
Figure 50
Conceptual view of frac
packing
2
Department of Petroleum Engineering, Heriot-Watt University 51
27Unstable Formations and Sand Control7
Inflo
w A
rea
as %
of O
penh
ole
Flo
w A
rea
6 SPF 12 SPF0.1
1
10
100
24 SPF
WrappedScreen
SlottedLiner
Casing with 0.75in. diameterperforations
(SPF = Shots per foot)Uniform
Membrane(High Cost)
Openhole
The inflow area of the wirewrapped screen is reduced to 3% once a gravel pack isplaced in front of it. High cost membranes with a size small enough to control sandparticle movement without need for a gravel pack have been developed with a veryhigh inflow area. The membrane is run in the hole during the completion process andassumes that the formation sand will collapse around it.
An alternative approach being developed is the use of an expandable sand screen. Herethe sand screens diameter is expanded by 33% to 50% by pumping, pushing, pullingor rotating an expansion tool through the screen once it has been placed across thecompletion interval (figure 52).
Expanded PipeUnexpanded Pipe
Expansion ConeFormation requiring support
by sand screen
Figure 51
Inflow area as a function of
completion type
Figure 52
The expandable screen
1
52
5.12 Chemical Sand ConsolidationChemical sand consolidation - the artificial strengthening of the oil/gas producingformation - is the least frequently used of the sand control. This is despite its twomajor advantages:
(i) unlike mechanical forms of sand exclusion, it leaves the wellbore completelyunobstructed without imposing any restrictions as far as future work-overs areconcerned (compared to a standard perforated completion)
(ii) the consolidation treatment can also be carried out through tubing i.e. a drillingor work-over rig does not have to be moved onto the well and the tubing etc. doesnot have to be pulled as the first stage of the well repair. Instead a pump truck anda series of chemical storage tanks with suitable manifolding has to be provided (seefigure 54). This can often be mobilised more quickly and at a lower cost than a rig.
The chemical glue must:
(i) wet and adhere to the sand grains
(ii) leave the pore spaces open so that the permeability is retained.
This is illustrated in figure 53 which shows an enlarged view of a number of sandgrains which have been covered with chemical cement. Organic resins (epoxy, furanor phenolic) are the most widely used chemical cements -though other materials havebeen used in the past -solder, water glass, alumina, nickel plating. The typical targetformation for chemical consolidation are thin ( < 3m), high permeability, clean (lowclay content) sands. This restriction on the formation type to be treated and theoperational complexity explains the lack of popularity of chemical consolidation as:
Layer of chemical cement
Openpore
space
Sand grain
Sand grain
Sand grainSand grain
Cement concentratedat intergraincontact point
(i) the consolidation treatment is operationally complex (figure 54). This is dueto the need to prepare the sand grain surface so that the chemical cement will adhereto it followed by the need to re-establish the formation permeability e.g. byoverflushing excess chemical cement away from the near wellbore area anddisplaceing it deeper into the formation. The complex series of fluids to be pumped
Figure 53
An enlarged view of a
chemically consolidated
formation
2
Department of Petroleum Engineering, Heriot-Watt University 53
27Unstable Formations and Sand Control7
could include some, or all, of: acid, neutraliser, preflush, spacer, resin, overflush,displacement fluid. Each fluid requires a separate tank which have to be manifoldedtogether so they can be pumped in the correct sequence in one smooth, continuousoperation:
Closed Tank For Returns
High Pressure
Pump
(truck)
Filtration Unit
To Mud Pits
Wellhead
Booster/Charge Pump
Storage Tanks
Diesel/Kerosene
BrinePre-flush Spacer
Mixer
A B C D
Small Tanks With Chemical Cement
Properties
(ii) many of the chemicals used have aggressive properties and due care andattention has to be given to operator safety and potential environmental impact.
(iii) the formation composition can be detrimental to the strength development ofthe chemical cement e.g. A chemical formulation capable of developing a strengthof 150 bar unconfined compressive strength in a formation containing clean sandgrains, will typically show a negligible strength development when the clay contentrises to 10% wt (see figure 55). The need to treat clean sand indicates that highpermeability formations are the main target for chemical consolidation.
Figure 54
Site layout for chemical
consolidation N.B. May
also require acid and
neutraliser tanks if pre-
consolidation matrix
stimulation treatment
required
1
54
150
100
00%
100
5%
95%
10%
90
Clay content
Sand content
Unc
onfin
ed c
ompr
essi
ve s
tren
gth
of c
onso
lidat
ed fo
rmat
ion
(bar
)
(iv) All the chemical flushes have to be applied to all the perforations in the correctorder since one or more incorrectly treated perforations could lead to future sandproduction. Field experience has shown that this requirement for correct placementmeans that only thin zones (maximum length 3m) should be treated at one time.Longer zones require a series of consecutive treatments.
Figure 55
Typical strength
development of a
consolidation system with
increasing clay content
2
Department of Petroleum Engineering, Heriot-Watt University 55
27Unstable Formations and Sand Control7
SAND CONTROL TUTORIAL
“Installing sand control equipment is sometimes a necessary evil”
Question 1.
Discuss this statement with respect to• Well productivity• Easy of well operation• Easy of future workovers
Rank the various types of control for their impact on well productivity
Answer 1.
Well productivityHigh well productivity losses (typically 60%) are observed after well killingfollowed by sand control equipment implementation such as gravel packing. Insome cases the final well productivity was only around 20% of that achieved priorto gravel packing. Such high impairment has been noted in many gravel packedwells.
Easy of well operationInstallation of gravel packing creates mechanical restriction in wellbore e.g.Production logging is normally no longer possible.
Easy of future workoversInstallation of gravel packing also makes future workover operations difficult.Requires removal (fishing) of packers / screens, etc. Workover options arediminished, making difficult to identify the source as well as making it moredifficult to shut-off undesirable water and gas.
The various types of sand control can be ranked by their impact on the aboveparameters as follows:
Sand control type Well Easy of well Easy of futureproduction operation workovers
Perforated completion 2 1 1(living with sand) (before sand failure) (before sand failure)
5 5(after sand failure) (after sand failure)
Screen/Slotted lines 1 1 3Internal gravel pack 5 4 4External gravel pack 4 4 5Consolidation 3 2 1
Question 2.
Pressure drops as a function of flow rate has been measured through a number ofsimulated perforations filled with gravel and gravel/sand mixtures. The results areshown in graphs a to d. These graphs illustrate many of the aspects of flow through
1
56
porous media. Identify those aspects which can be derived from the graphs andcomment on how these would influence your design of an internal gravel pack.
Answer 2.
Graph a
2000
1800
1600
1400
1200
1000
800
600
400
200
00 25 50 75 100 125 150 175
Flow rate per perforation (bbls/day)
Pre
ssur
e lo
ss p
er in
ch o
f per
fora
tion
tunn
el le
ngth
psi
Darcy lawcomputationK - 1 Darcy
EXPERIMENTA
L
Darcy law computation
Legend:3/8" perforation
0.017" - 0.033" gravelø = gravel porosity ø = 40%
ø = 35%
a) Turbulent flow through gravel packed perforations greatly increases the pressuredrop with increasing flow rate compared to the pressure drop value calculated fromDarcy’s Law (which is independent of flow rate).
b) Increased pack porosity gives increased permeability (reduced pressure drops)
Graph b
Flow rate (bbl/d per perforation)
PerforationDiameter
Pre
ssur
e dr
op (
psi)
200
150
100
50
00 5 10 15 20 25
A
B
D
D 3/8 inchC C 1/2 inch B 3/4 inch A 1 inch
a) Pressure drop across the gravel filled perforation decreases as the perforationdiameter increases
b) Larger perforations with lower flow velocities show reduced turbulent effect
2
Department of Petroleum Engineering, Heriot-Watt University 57
27Unstable Formations and Sand Control7
Graph c
Flow rate (bbl/d per perforation)
Pre
ssur
e dr
op (
psi)
200
150
100
50
00 5 10 15 20 25
A
B
1 DarcyA Formation sandB 40 -60 gravelC 20 - 40 gravelD 12 - 20 gravel
CD
1 Darcy formation sand
a) Even the smallest (40/60) gravel pack sand shows a greatly reduced pressure dropcompared to a good formation sand (1 Darcy) when packed in the formation tunnels.
b) “Extra” pressure drops due to turbulence (or non-Darcy flow) decreases as thegravel size increases.
Graph d
70
60
50
40
30
20
10
00 20 40 60 80 100
Per
mea
bilit
y (D
arcl
ea)
% sand mixture
a) Permeability of gravel/formation sand mixtures decreases as the concentration offormation sand increases.
b) Some mixtures show a permeability even lower than the formation sand alone.
1
58
6. FURTHER READING
(1) Golan M. & Whitson C.“Well Performance” 2nd editionpublished by Norwegian University of Science and Technology
(2) Allen T. & Roberts A.Volume 2 “Production Operations” 4th editionpublished by OGCI
(3) Economides M., Hill A. & Economides C.“Petroleum production Systems”published by Prentice Hall
(4) PENBERTHY, W.L. & SHAUGHNESSY, C.M.Sand Control. SPE Series on Special Topics. Volume 1.Richardson: SPE, 1992ISBN 1555630413
2
Department of Petroleum Engineering, Heriot-Watt University 59
27Unstable Formations and Sand Control7
Well Control 11
C O N T E N T S
1. INTRODUCTION2. FIELD DEVELOPMENT
2.1. Oil Fields2.2. Gas Fields
3. FIELD DEVELOPMENT EXAMPLES4. PRODUCTION PLATFORM FUNCTIONS5. PRODUCTION PHILOSOPHIES
5.1. Separation Objectives5.2. Oilfields5.3. Gas Fields
6. BASIC PROCESSING SCHEME6.1. Processing Conditions6.2. Process Plant Metallurgy6.3. Separators6.3.1. Horizontal Separators6.3.2. Vertical Separators6.4. Operational Production Problems with
Gravity Separators6.4.1. Foaming6.4.2. Solids6.4.3. Emulsion6.4.4. Surging Flow6.4.5. Production Chemicals6.4.6. Miscellaneous Processes and Comments
7. SEPARATOR SIZING BASICS7.1. Gas Capacity7.2. Liquid Capacity
8. TEST SEPARATOR9. COMPRESSORS10. OIL EXPORT11. GAS HANDLING
11.1. Increasing NGL Recovery11.2. NGL Stabilisation11.3. Gas Dehydration11.3.1. Dew Point Depression11.3.2. Hydrates11.4. Continuous Dehydration Process11.5. Batch Dehydration Process11.6. Acid Gas Treating
12. FURTHER READING
8Oil and Gas Processing8
1
2
LEARNING OUTCOMES:
Having worked through this chapter the student will be able to:
• Discuss the impact of the field location, well rate, produced fluid and secondaryprocessing on the design and operation of the production facilities
• Relate the required Production Facility services to the field’s oil recovery mechanism
• Draw a basic, outline production process scheme
• Describe the components and discuss the operation of a 3 phase separator
• Identify the advantages of horizontal and vertical separators
• Describe the operational problems associated with these separators
• Quantify the sizing (gas and liquid capacity) of a 3 phase separator
• Discuss fiscal measurement of produced cride oil
• Describe a pipeline “pigging” operation
• Describe the components of a gas handling facility viz NGL separation andstabilisation, gas dehydration and sweetenting.
Department of Petroleum Engineering, Heriot-Watt University 3
8Oil and Gas Processing8
1. INTRODUCTION
This section covers the treatment of the produced fluid from and during its passagefrom the well-head, through the facility to the point of sale. The nature, scope andgeographical spread of these facilities will vary greatly depending on the:
(i) Location: On or off-shore
(ii) Well and Field Production Rate
(iii) Oil, gas or condensate Field
(iv) Central gathering station or local (well-head) facilities
(v) Secondary processing requirements e.g. removal of contaminants such as H2S
However, in all cases the facility is designed to separate the (multi) well stream intothe three basic components (oil, gas and water) and to process the phases into:
(i) marketable products (i.e. to sales specification) - or
(ii) a form that they can be disposed of in an environmentally acceptable manner.
A simplified production scheme for an oil or gas field is given as Figure 1.
Manifold
Gas
Water
Reuse or disposal
Oil
Sales
Sales pipeline
Compressor
Lower water and hydrocarbon dewpoints.
Remove contaminents as required
Choke
Storage Tank
(Multi stage)Three phase
separators
Simplified processing oil facility scheme
The primary separation process takes place in gravity separators. The process isdriven by the density difference between the gas, oil and water phases and the highfluid pressure frequently available at the well head. Gas “flashes” from the well-headfluid as the pressure is reduced, while at the same time “free water” is separated fromthe oil.
Figure 1
Simplified processing oil
facility scheme
1
4
The gas must be compressed to the export pipeline pressure and treated to removewater vapour and heavier hydrocarbons. More complicated processes may berequired to “sweeten” the gas by removing other contaminants such as carbondioxide, hydrogen sulphide etc.
Typically, the oil and water will only show partial separation, and an “emulsion”, orintimate mixture of oil or water droplets in the other phase, will require more intensiveprocessing to achieve a suitable level of separation. Any produced solid material isalso normally found in this layer.
2. FIELD DEVELOPMENT
One of the fundamental decisions to be made during field development is whether todevelop the discovered field as an oil or gas field. This has a large impact on:
(i) the sales contract
(ii) the field development philosophy and any measures taken for improving thereservoir recovery efficiency and
(iii) the production facilities.
The design and operation of the production facilities influences the (relative) recoveryof oil and gas. In virtually all cases, the production facilities will be designed tomaximise the recovery of (hydrocarbon) liquids since their sale is normally moreprofitable than gas. However, more complex facilities are usually required to recovera higher proportion of liquids; resulting in a trade off between increased capitalexpenditure, (possibly) reduced operating cost and increased revenue.
2.1 Oil FieldsThe production from oil fields build up rapidly as more wells are drilled and broughtonto production. The oil field is then produced at a maximum (plateau) for the nextfew years. This rate is determined by the capacity of the production facility. The oilproduction then gradually decreases until the income from the oil production no longerpays for the field operating expenses (Figure 2). The field is then ready for abandonment.
The plateau oil production rate will be determined by the individual well rates, and thenumbers of wells to be drilled; not to mention the geology, reservoir fluid propertiesetc. The length of the production plateau and the decline will also be a function of thereservoir size and the recovery mechanism.
Oil fields will typically produce a much larger volume of water than oil. Hence, oil/water separation typically occurs near to the production well to minimise unnecessaryexpense in the pumping of large volumes of water over large distances. In particular,offshore separation will be employed in the case of offshore fields; while wellheadseparation (at the sea floor, and disposal into an injection well) is now being developedfor subsea wells. Offshore operations are thus becoming more similar to landoperations, where piping and pipeline costs are minimised by:
Department of Petroleum Engineering, Heriot-Watt University 5
8Oil and Gas Processing8
(i) local separation and disposal of the bulk of the produced water
(ii) oil (with a relatively low water content) being sent to a central gathering stationfor final water separation treatment.
ProductionDecline
Time
Oil
Pro
duct
ion
Rat
e
Plateau
Production
Pro
duct
ion
Bui
ld-U
pField Abandonment
Operating Expenses (Oil Equivalent)
First OilD
isco
very
Large volumes of sales quality crude oil have to be piped to the coastal export terminalvia a pipeline where it can become part of the world crude oil trade. This is particularlytrue for onshore oil fields where the only economic means of transporting significantvolumes is by pipeline, though smaller volumes are sometimes transported by train/barge or road. Offshore, the choice between a pipeline and local storage near theplatform together with a shuttle tanker to transport the crude oil directly to the refineryor to the export terminal will be determined by the economics of the two scenarios.If a dedicated pipeline to the coast can not be justified, then the presence of existinginfrastructure - i.e. a nearby pipeline - with sufficient available capacity to transportthat projected volumes is the key factor. The increasing density of the pipelinetransport network in the North Sea is illustrated in Figure 3.
Gas export is normally only possible when the volumes are sufficient to make buildingof a dedicated pipeline economic. Thus, the available pipeline infrastructure is evenmore important in the case of produced gas since there are no storage alternatives - ithas to be exported, used or otherwise disposed of at the same time as it is produced.Many oil developments will use (part of) the gas as a fuel to power the platform andfacilities. However, the changes in the volumes of produced fluids from the wells(table 3), imply that there may be excess gas early in the life of the project availablefor export, while the platform becomes gas-short later on in the project lifetime. Theinstallation of a gas pipeline allows the field to export gas in the early years whileimporting gas to power platform equipment in late project life. This is particularlyattractive if the operators own a nearby field with surplus gas. All these factors havea large impact on the design of the platform facilities.
Figure 2
Oil field production
1
6
Tern
Eider
NorthCormorant
SouthCormorant "A"
Magnus"A"
Stratfjord
North Alwyn
BRENTA
B
C
D
Hutton
N.W.Hutton
CentralCormorant
UMC
Heather "A"
Central
Ninian Nth
Ninian Sth
Don
Murchison
Thistle "A"
Ospery
Any remaining gas has to be disposed of by flaring, venting (releasing to theatmosphere without combustion) or underground disposal. Reduction in the emitted(flared and vented) gas has become a governmental objective with the imposition ofa carbon tax in some countries. Further, company management has encouraged thedevelopment of many innovative facility designs/field development options toachieve this.
As mentioned earlier, large volumes of water are “co-produced” with the oil.Frequently large volumes of water are injected into the reservoir to increase the oilrecovery by “sweeping” the oil from the injection to the production well, and bymaintaining the down-hole reservoir pressure.
This requires facilities to:
(i) extract large volumes of water from a surface (sea or river) or under groundsource (dedicated water production wells),
(ii) remove solids (filtration) and (corrosive) oxygen to the specified levels and
(iii) inject the water into a dedicated injection well.
2.2 Gas FieldsThe development of a hydrocarbon resource such as a gas field implies that, unlikeoilfields, the produced fluid will be mainly gas accompanied by small volumes ofwater and condensate {sometimes called “natural gasoline” or natural gas liquids
Figure 3
Northern North Sea
Pipeline Systems
Department of Petroleum Engineering, Heriot-Watt University 7
8Oil and Gas Processing8
(NGL)}. Offshore separation is normally employed with the (small volumes of) waterbeing disposed of to the sea and the gas/condensate being exported to the coast viamultiphase pipelines. Advances in control engineering (automation, data transmis-sion, infra red gas detectors etc.) mean that (small and mature) fields can now be runon a “not-normally-manned” basis with the consequent reduction in operatingexpense and manpower employed.
A new type of field that is being developed in the North Sea is the High Temperature,High Pressure Gas Condensate field. These fields produce (relatively) much largervolumes of condensate compared to gas fields. They also exhibit unusually largechanges in wellhead pressures and fluid compositions as the reservoir depletes.
3. FIELD DEVELOPMENT EXAMPLES
The majority of offshore production platforms are supported by steel jackets whichare held in place on the sea bottom by steel piles (see Figure 4 and 5, a large, integratedsteel jacket platform). These welded pipe structures provide a support for variousprefabricated modules, e.g. accommodation, power facilities etc. {see section 5 -‘Platform Functions’ for a more detailed discussion}.
Mono Pod Platform(Usually Unmanned)
Tens ion Leg Platform
Steel Jacket Platform
Jack - Up Produ ction PlatformWith Subsea Stora ge
Concrete Gravity Platform
Met
ers
0
50
100
150
200
250
300
Float ing Produ ction PlatformWith Subsea Colle ction Man ifold
Figure 4
Types of offshore platforms
Figure 5
A large steel jacket
platform
1
8
Concrete gravity platforms depend on their large weight to hold them in place on thesea bed. The Brent D platform weighs over 200,000 tons, but is also capable of storing1,000,000 barrels of oil; so as well as avoiding the need for piling, it provides localcrude storage capacity for transfer at regular intervals to the shuttle tankers.
An offshore storage buoy or storage tanker would be required if this integral storagewas not available within the platform construction. Offshore loading is illustrated inFigure 6.
Fixed platforms become uneconomic as the water depth increases and floatingplatforms are attached - either tethered (anchored) over subsea production wells ormore rigidly held in place by tensioned cables. The most recent developments toexploit the smaller hydrocarbon reservoirs currently being developed, can be foundin the combination of subsea wells drilled from one or more drilling centres, with aFPSO (Floating Production Storage and Offtake) vessel. The FPSO hosts all thenormal platform functions such as production separation and water injection equip-ment. As discussed earlier, gas export requires a pipeline while oil export can be bypipeline or shuttle tanker. Figure 7 illustrates a typical example of this type ofdevelopment. An FPSO is not the only option for this type of development - dependingon the reserves, location etc., a fixed platform can play host to the production facilities(Figure 8) for the subsea wells
4. PRODUCTION PLATFORM FUNCTIONS
Production platforms perform a multitude of functions as listed below. All functionsmay be carried out in one large platform - as in the case of integrated platformsillustrated in Figure 5. Alternatively they may be split into a series of separateplatforms - each with one or more of the following functions - connected by a bridgeso that personnel and equipment can move easily between them. Figure 9 illustratesa typical organisation of the following modules.
Figure 6
Offshore loading
Department of Petroleum Engineering, Heriot-Watt University 9
8Oil and Gas Processing8
FPSO
P3
P2
P1
P4
Gas export
Oil Line
Water Injection Line
H
Field A
Field B
Field C
ProductionManifold
Water InjectionManifold
Producers
Injector
Injection
Producer
Field 1
Field 4
Field 3
Field 2
ToField 4
Drill Centre A
Drill Centre B
Drill Centre D3-5km to Host Platform
Drill Centre C
To Field 1
Diverter
DiverterCrossover Subsea
Isolation Valve
2 x 6" Gas Pipelines
16" OilPipeline
Existing 20" Gas Pipeline
Riser Platform> 100km
20" GasPipeline
Figure 7
A subsea FPSO
development
Figure 8
A host platform connected
to several subsea fields
1
10
1. Drilling derrick with wellheads situated directly underneath2. Drilling support module such as drilling mud preparation3. Process equipment where oil, gas and water are separated and treated to sales
or end-user specifications4. Export and (sometimes import of) sales quality crude oil and gas4a. Compression module where gas is compressed to the required pressure for
export or own use (gas lift, gas injection or power generation). Excess gas isflared via the flare boom
4b. Oil export module which house the pumps for exporting the crude to pipeline,local (floating) storage or shuttle tanker
4c. Water Injection module consisting of filtration, deoxygenation, chlorination,and high pressure injection pumps
5. Platform utilities such as power generation (frequently gas turbines but stand-by diesel power in the case that the gas supply is unavailable)
6. Accommodation / life support for the personnel manning the platform. Forsafety reasons, this is normally situated as far as possible from the process/compression/wellhead areas.
7. Control room, the hub of the platform’s operation. Safety and loss controlsystems are monitored here, as is the fiscal metering that accurately measuresthe volumes of crude oil exported.
8. Maintenance workshops, communications, cranes to lift supplies and equipment onto the platform from supply vessels, transport (helicopters) etc.
The chosen Reservoir Oil Recovery Mechanism and associated artificial lift used toincrease the energy available to lift fluids from the downhole to the surface will havea significant impact on the required (platform) services (Table 1).
Crude Oil,Gas and Water from Reservoir
Sea WaterInjection
Flareboom
Drilling Derrick &Substructure
Helideck
DrillingModule
WellheadModule
ProcessModule
CellarDecks
Sea Level
UtilityModule
PowerGeneration
ControlRoom Accomodation
Gas
Lift
Oil
GasGas Turbines
Gas ForFuel
Oil
Exp
ort
Gas
Exp
ort
Gas to Fuel
Figure 9
An integrated steel jacket
platform
Department of Petroleum Engineering, Heriot-Watt University 11
8Oil and Gas Processing8
Recovery Mechanism Process Impact
Primary Recovery Natural Flow - No extra platform services
Artificial Lift Gas Lift Extra Gas Compression
ESP* Extra Electrical Power Generation
Secondary Recovery Water Flood Water Injection Facilities
Pressure Water injection Water Injection Facilities
Maintenance Gas injection (High Pressure) Gas Compression
Tertiary Recovery Steam,
Carbon Dioxide,
Miscible Gas,
Chemical etc.
* ESP - Electric Submersible Pump
Dedicated, Specialised Facilities
5. PRODUCTION PHILOSOPHIES
The process plant capacity is set by the number of production wells and the wellproduction rates (determined by reservoir permeability, fluid properties, and pressurethickness of the pay zone, well design, individual well reserves etc.) and the fieldproduction philosophy. These differ between oil and gas fields:
5.1 Separation ObjectivesThe complete separation process system is designed to produce on specificationexport fluids. The actual values will depend on the specific (crude oil or gas)properties and the transport route (Table 2).
5.2 OilfieldsOilfields are produced at maximum plateau rates dictated by the facility design for aslong as possible. The (oil) production rate will then decline usually accompanied byincreasing water production until the minimum economic rate is reached. This is thepoint at which operating expenses are no longer paid for by the oil income (Figure 2and Table 3).
Table 1
Scale and type of platform
services depend on recovery
mechanism.
1
12
PRODUCT PROPERTY SPECIFICATION
CRUDE OIL Vapour pressure < 16 psi at 25º C *
Water content < 0.5% wt
Salt content < 70 g/m3
Temperature < 40º C
Pressure ~ atmospheric for tanker export
or pipeline operating pressure
* Higher values allowed pipeline export
GAS Calorific value ~ specified limits
Liquids - none
Hydrocarbon dewpoint < -3º C
Water dewpoint < -8º C in NW Europe
Carbon dioxide < 3% wt
Hydrogen sulphide < 4 ppm
Temperature < 40º C
Pressure ~ pipeline operating value
* varies with ambient conditions e.g. 0º C in sub-tropical areas, -20º C in Canada
WATER Dispersed oil < 40 ppm for marine discharge in N.W. Europe,
< 32 ppm in Gulf of Mexico
Phase Fluid Early Mid-Life Mature Abandonment
Oil (bopd) 100k 100k 30k 10k
Gas (MM sft3/d) 100 100 30 10
Water (bwpd) <500 30k 100k >100k
Water cut <0.5% 30% 77% >90%
The gas associated with the oil production must be used at the time of production -whether for export, fuel, lift gas, re-injection or flared. If the opportunity for pipelineexport does not exist, then projects frequently have excess gas early in their early/midlifes, while they are short of gas in the later, decline phase.
5.3 Gas FieldsThe pattern of gas usage tends to be seasonal i.e. with periods of high and low demand:
(i) Winter peak in N.W. Europe, Northern USA and Canada, here space heatinguses a large proportion of the gas.
Table 2
Typical sales/disposal
specifications
Table 3
Typical changes in
production rates during the
lifetime of a 100,000 bopd
oil field
Department of Petroleum Engineering, Heriot-Watt University 13
8Oil and Gas Processing8
(ii) Summer peak in Southern U.S.A. where air conditioning is the major user.
(iii) Constant demand when industry is the main user (e.g. Aluminium Smelting) orwhen supplying a Liquefied Natural Gas Plant.
A typical nomination contract in N.W. Europe will specify both a total yearlyproduction and a minimum production rate at the end of the winter period.
The composition of any produced water will change significantly during the field lifetime. Initially it will be (fresh) condensed water vapour that was originally present inthe gas phase in the reservoir. Later on, (liquid) formation water will start to beproduced from the reservoir - resulting in a gradual increase in salinity until theformation water value is reached if large scale water production occurs. Frequently,the easiest export route for any liquid hydrocarbons (condensate) recovered during theseparation process is to spike (inject) them back into the gas export pipeline for laterrecovery at a central, onshore facility.
A typical production profile is illustrated in Figure 10 and Table 4.
Years
Pro
duct
ion
1 2 3 4 5 6 7 8
Plateau
Decline
First Gas
Dis
cove
ry
Operating Expenses (Gas Equivalent)
Abandonment
Annual Peak e.g. Winter in NW Europe
Phase Fluid Early Mid-Life Declining Abandonment
Gas (MM sft3/d) 500 500 50 20
Condensate (bpd) 5000 5000 500 200
Water (bwpd) 250 500 500 200
Figure 10
A typical gas field
production profile
Table 4
Typical changes in
production rates during the
lifetime of a typical
southern North Sea gas
field
1
14
6.BASIC PROCESSING SCHEME
The building blocks of a simplified processing scheme are shown in Figure 11. Thereservoir fluids are produced to surface at the wellhead followed by separation into theoil, gas and water phases. Each phase is then treated individually to reach the salesspecification.
(i) The oil content of the produced water is reduced to a level suitable for disposal.Any recovered oil is passed to the final stages of oil treatment.
(ii) The separated gas is treated to sales specification by reducing its water andliquid hydrocarbon content. The recovered liquid hydrocarbons are “spiked” intothe crude oil stream if it is being transported by pipeline. Alternatively, it may beadded to the sales gas pipeline, followed by onshore separation. The gas may be sold,used for power generation, gas lift, injection back into the reservoir with any excessbeing flared.
(iii) The crude oil is transported by pipeline or tanker (sea, road, or rail) after beingtreated to the appropriate specifications. The latter option requires larger, localstorage than when transport is by pipeline.
Gas InjectionGas Lift
Gas Treatment
Gas FuelGas
Possible Own Use
Recovered Oil
Recovered Liquid
Hydrocarbons
Disposal
Flare
Gas Pipeline if Present
Oil Pipeline
or
Tanker
Manifold Primary Oil/Gas/WaterSeparation
Final OilTreatment
WaterTreatment
LocalStorage
6.1 Processing ConditionsThe process equipment design and the materials of construction are determined by:
Figure 11
A basic process scheme
Department of Petroleum Engineering, Heriot-Watt University 15
8Oil and Gas Processing8
Fluid QualityWellhead pressure and temperature of the produced reservoir fluid. These changedramatically over the lifetime of the field - see table 3 and module 10, section 10.2.
Fluid PropertiesThe properties of the crude oil are analysed in the laboratory, and phase equilibria(“PVT”) properties are determined. This aspect is discussed in detail in the ReservoirEngineering part of this course. Hydrocarbon mixtures can be classified into five maintypes based on these phase equilibria properties as summarised in Table 5.
Hydrocarbon type Low High RetrogradeShrinkage Oil Shrinkage Oil Condensate Wet Gas Dry Gas
Field Development type Oil reservoir Oil reservoir Oil reservoir Gas reservoir Gas reservoir
Stabilised crude density < 30°API 30° < API < 50° API < 60° API > 50° API > 50°
Gas Oil Ratio(sft3/bbl) < 500 500 < GOR > < 70,000 < 100,000 > 100,000
8,000Comments Broad phase
envelope with high proportion of heavy hydrocarbons
Narrow phaseenvelope with few heavy hydrocarbons
Even more lighter and fewer heavier hydrocarbons
The hydrocarbon fluid properties along with the wellhead pressures determine thenumber of separation stages and the process conditions under which the oil/gas/waterseparation is carried out. Many of the stabilised crude oils produced by the differentNorth Sea fields have similar properties. These are summarised in Table 6.
Property Value
Specific Gravity 0.84 g/cm
36° API in oilfield units
Viscosity 4 cSt at 50° C
Sulphur content < 1.0 % wt
Gas Oil Ratio 1000 scf/stb
Crude Oil Composition 50% wt paraffins
30% wt napthenes
20% wt aromatics
Wax variable, but often low (<5%)
Asphaltenes variable, but usually low (<2%)
Gas Composition > 80% methane
Carbon Dioxide and Hydrogen Sulphide in the gas phase are also normally low.
Table 5
Types of Hydrocarbons
produced
Table 6
Typical North Sea crude oil
properties
1
16
The similarity of the crudes developed from the many fields means that they can bepiped to the coastal terminal in a “common carrier” pipeline. The lower the specificgravity (the higher the API gravity), the greater the market value of the crude due tothe increased yield in high value, refinery products. Economics thus dictate thatdenser, more viscous, lower API gravity crudes - such as that produced by the Albafield - are transferred to shore via a shuttle tanker, even when a “common carrier”pipeline with sufficient free capacity is available. This is to protect the value of theother crude streams using the pipeline.
6.2 Process Plant MetallurgyMaximum producing wellhead temperature and pressures along with the presence ofcorrosive components (e.g. carbon dioxide, hydrogen sulphide, saline formationwaters) determine the specification (metallurgy) of the process equipment. Processconditions can be modified so that a more economical construction metallurgy can bechosen viz:
(i) Inlet temperatures can be reduced by cooling in a heat exchanger (seawater isavailable in large quantities as a cheap, cooling medium for offshore operations)
(ii) Pressures can be reduced by use of a choke
(iii) Corrosivity can be reduced by the injection of corrosion inhibitors. These area mixture of chemicals injected into the process stream at a low dosage level whichcoat the metal surfaces, reducing their susceptibility to corrosive attack.
6.3 SeparatorsSeparators form the heart of the production process. There are two basic types:
(i) Gravity separators which depend on the density difference between the phasesto be separated
(ii) Centrifugal separators in which the effect of gravity is enhanced by spinning thefluids at a high velocity
Gravity separators are essentially large cylindrical pressure vessels up to 5m indiameter and 20m long. They are used in either 2-phase (liquid/gas) separation, or 3-phase (water/oil/gas separation). They are normally mounted in a series of 2, 3 or even4 separators (figure 12) with inlet and outlet under pressure control. They be mountedeither vertically or horizontally.
Department of Petroleum Engineering, Heriot-Watt University 17
8Oil and Gas Processing8
ChokeHigh Pressure
SeparatorIntermediate Pressure
SeparatorLow Pressure
SeparatorLC LCLC
PC PC PC
Pressure Control at 100 bar
Level Control Valve
Level Control Valve
Pressure Control at 20 bar
Pressure Control at 4 barHigh
PressureGas
From WellManifold
IntermediatePressure
Gas
LowPressure
Gas
Vent Gas(Pressure =1.5 bar)
Storage Tank
Crude Oil
Gravity separators (see figure 14) consist of an:
(i) Inlet section with momentum breaker/inlet deflector to rapidly change the inletliquid velocity; hence helping disengage free gas.
(ii) Gravity settling section, typically sized so that:
(a) 2-phase separators: sufficient gas phase residence time such that liquiddroplets of 100 µm will separate from the gas.
(b) 3-phase separators: as for 2-phase separators plus sufficient oil phaseresidence time such that 500 µm water droplets will settle into the waterphase. The separated water phase will typically contain 500 ppmdispersed oil, hence requiring further treatment before disposal.
The above criteria are based on settling theory (see section 8.7); alternatively,experiment may be carried out to measure the rate of oil/water separation. Typicalresults for such a settling experiment are shown in Figure 13 - the initially producedintimate oil and water mixture (emulsion) separates into a lower, "clean" (or low oilcontent) water layer and an upper, "dry" (or low water content) oil layer with a (morepersistent) emulsion layer in-between.
(iii) Gas outlet with mist extractor. One design of which is a wire pad of finelywoven stainless steel wire wrapped in a cylinder. It is designed to remove liquiddroplets between 10mm and 100µm. These droplets impinge on the wire, coalesceand flow down in to the liquid phase. Efficient operation depends on operating withthe correct gas velocity:
(a) too high gas velocity - liquid drops “eroded” from wire and is re-entrained.
(b) too low liquid velocity - liquid drops drift past mesh without impinging.
A gas scrubber is employed if larger droplets (up to 500µm) have to be removed sincea mist eliminator would flood with this level of liquid loading.
(iv) Liquid outlet under level control to evacuate liquid or separate oil and waterphase (2/3-phase separation operation, respectively). The outlet is usually equippedwith a vortex breaker to prevent re-entrainment of gas.
Figure 12
Simplified 3-stage, two
phase (gas / liquid)
separation
1
18
"DRY" OIL
EMULSION EMULSION
"CLEAN" WATER
h
h oh e
h w
h w/h
"DRY" OIL
"CLEAN" WATER
TimeAppearance Oil/Water Emulsion Sample Behaviour of Oil/Water Emulsion With Time
6.3.1 Horizontal Separators (Figure 14)These are most suited to separation of large volumes of gas from liquid. Theiradvantages/disadvantages compared to vertical separators are:
(i) Larger interface area gives better foam/emulsion handling characterisation
(ii) Can be modularised, but require larger surface area
(iii) Solids removal is less efficient and requires a more complex jet wash system
(iv) Lower surge capacity i.e. reduced ability to deal with uneven, inlet flow
Gas
Mist Eliminator
Oil Outlet andLevel Control
Water Outlet andLevel Control To Oil Export
To ProducedWater Treatment
InletFrom ProductionManifold
Inlet Deflector/ Momentum Breaker
Pressure Control Valve
OIL and EMULSION
WATER
GAS
PC
Weir
OIL
6.3.2 Vertical Separators (Figure 15)These are most suited for separation of gas from large volumes of liquid. Comparedto horizontal separators they are:
(i) Good for uneven, surging inlet flow due to their greater height
(ii) The oil/gas and oil/water interface level control is less critical for the same reason
(iii) However, they do tend to be larger than their horizontal equivalent for the sameseparation capacity
Figure 13
Separation of oil / water /
emulsion liquid with time
Figure 14
Three phase horizontal
separator
Department of Petroleum Engineering, Heriot-Watt University 19
8Oil and Gas Processing8
Gas
Mist Eliminator
Oil Outlet andLevel Control
Water Outlet andLevel Control
OIL
WATER
To Oil Export
To Water Disposal
Inlet
Inlet Deflector/ Momentum Breaker
Pressure Control Valve
PC
6.4 Operational Production Problems with Gravity SeparatorsThe process equipment used in platform operations is designed to achieve the requiredproduct specifications using standard chemical engineering design procedures. How-ever, a number of operational problems referred to above can be encountered ifallowance is not made for the fact that we are not dealing with “pure” fluids. Thecomplex mixture that makes up the produced hydrocarbon contains many minorcomponents that cause:
6.4.1 Foaming:Presence of semi stable gas bubbles at the oil/gas surface that prevent the gasdisengaging quickly and cleanly from the liquid surface. A de-foaming chemical(surfactant) is injected at a low concentration to overcome this problem; coupled withthe use of a mist eliminator which will remove liquid droplets from in the size range10-100µm.
6.4.2 SolidsLow concentrations of solids are frequently produced to surface with the well fluids.They may be present in a low concentration but represent a large absolute solidvolumes given the large volumes of produced fluid These solids range from “fines”- (micron) sized clay particles which can flow through the formation pore throatstructure - to individual sand grains / larger “lumps” of failed formation which canbe produced from weak or unconsolidated formations. These solids collect in the base
Figure 15
Vertical three phase
separator
1
20
of the separators where they reduce the separator performance (shorter residence time)and can lead to corrosion since the fluid is stagnent. They have to be regularlyremoved. Water jets are mounted at the bottom of the tank - the jetting action re-suspends the solids followed by transport to a holding tank for ultimate disposal.Alternatively, manual removal of the solid deposit is required.
Several other solid phases can form in the gravity separators and the other componentsof the production system. These include:
(i) ScaleBy the mixing of incompatible waters {e.g. barium sulphate from mixing sea water(sulphate source) and produced water (source of barium), or due to pressure/temperature changes (e.g. to calcium carbonate scale)}. The point of deposition canbe controlled by the addition of scale ‘inhibitors’.
(ii) WaxCooling of the crude oil can result in its paraffin content precipitating as a solid wax.This is avoided by preventing the crude cooling below the wax cloudpoint tempera-ture; while the temperature at which deposition occurs can be controlled to someextent by the use of wax inhibitors.
(iii) AsphaltenesPressure reduction below the bubble point and the consequent loss of the morevolatile compenents can lead to the precipetation of asphaltenes from some crude oils
6.4.3 EmulsionMomentum breakers and corrugated plate settling packs are installed internally in theseparator to promote oil droplet coalescence and separation. However, emulsionseparation may not be sufficiently complete within the residence time available in theprimary separator. This depends on the chemical properties of the crude oil/watersystem (see section 6.3) and the time/rate of shear that the fluids have been subjectedto during the production process. Frequently used, practical solutions include theaddition of heat (viscosity reduction) or emulsion breaking chemicals together withacceptance of reduced oil and water quality being passed to the subsequent separationstages. Reverse emulsion breakers can be employed for treating water-in-oil emul-sions.
6.4.4 Surging FlowThe primary separator provides pressure control on all three (gas/oil/water) outlets. Itcontrols residence time via the oil/water level controls. Production wells frequentlydo not produce fluid at a constant even flow rate. This is due to the length and topologyof the production tubing and flow lines. These problems are accentuated byincorrectly set gas lift (well heading) as well as the long flow lines associated withsubsea wells. The separation system must be capable of dealing with the resultinghigh, instantaneous feed rates which can cause the levels to increase above theirnormal, operating values. The control system needs to be suitably adjusted to dealwith this.
Department of Petroleum Engineering, Heriot-Watt University 21
8Oil and Gas Processing8
6.4.5 Production ChemicalsIn addition to those already discussed, other production chemicals used includebiocides, corrosion inhibitors. Fig 16 is a typical example of where they might be usedin a oil/gas/water separation system.
NGLRecovery
HPSep
MPSep
LPSep
FlashDrum
SlopOil
GasSweetening
NGLMetering
Crude OilMetering
Gas Dehydration
Glycol Regeneration
NGL
Water
Fuel Gas
GasCooler
Cooler Cooler
ExportPumpBooster
Pump
Knock OutDrumCompressor
Compressor
Oil Pipeline
Compressor Knock OutDrum
Sour Gas
GasInjection
CorrosionInhibitor
CorrosionInhibitor
CorrosionInhibitor
FromOpen Drains
FromClosed Drains
To LPSeparator
Produced WaterDisposalCaisson
CoolerOily Water Treatment
GasDeoilingChemical
Defoamer
Demulsifier
Oil Cooler
ScaleInhibitor
WaxInhibitor
CorrosionInhibitor
WaxInhibitor
ScaleInhibitor
Well FluidsOil (& NGL)GasWaterGlycol
*: Deoiling Chemicals
6.4.6 Miscellaneous Processes and CommentsA fresh water wash may be required to reduce the salt content in the crude oil to thespecification value if the formation water has a high salt content. This involves mixingthe crude with the water followed by subsequent separation.
The main separation train may be twinned (i.e. two separation trains of equal capacity)to increase the reliability (uptime) of the separation facility. In addition a smaller testseparator will almost always be installed to allow individual well rates to be measuredas part of the field monitoring programme.
7. HORIZONTAL SEPARATOR SIZING BASICS
The liquid droplets will settle at a velocity determined by equating the gravity forceon the drop with the drag force caused by its relative motion to the fluid continuousphase. Typical tragectories of the settling liquid droplets are shown in Figure 17. Thisassumes that the internals of the separator have been designed so that turbulenceeffects are minimised e.g. by use of inlet deflectors/momentum breakers, vortexbreakers on the outlet.
Figure 16
Typical Oil / Gas / Water
process scheme
1
22
GAS
OIL
WATER
GAS
OIL
WATER
Vf
Vw
Vw = Water droplet settling velocity in oilVo = Oil droplet rising velocity in waterVf = Liquid droplet settling velocity in gas
Vo
Plan View Axial View
Vf
Vw
Vo
7.1 Gas CapacitySeparators are typically designed to allow liquid droplets larger than 100µ to settlefrom the gas phase to the liquid interface. The maximum allowable gas velocity (V
g)
which achieves this separation may be calculated by the Souder-Brown equation:
Vg = k p - p
pl g
g
where pl
= liquid densityp
g= gas density
k = constant
The constant k is related to the diameter of the droplet to be separated, the gas viscosityas well as the densities of the liquid and gas phases. Calculation of the constant k iscomplicated by the fact that the settling flow regime is not laminar, and qualitativecalculations are beyond the scope of this text. However, suitable correlation’s can befound in the Chemical Engineering literature (e.g. Arnold and Stewart, in section 8.11).
The minimum area required for gas flow (Ag) is then:
Ag = Q
g / V
g
where Qg is the specified maximum gas flow.
7.2 Liquid Capacity
The separation velocity (V) of one liquid from another is described by Stokes Law:
Figure 17
Setting droplet trajectories
Department of Petroleum Engineering, Heriot-Watt University 23
8Oil and Gas Processing8
Kd2 (ρd -ρc)µc
v =
whered = minimum specified droplet size to be separatedρ
d= density discontinuous phase
ρc
= density continuous phaseµ
c= viscosity of continuous phase
K - a constant
Also Required droplet separation time = vertical height of continuous phaseseparation velocity
The separation of water droplets from oil is normally more difficult than oil from watersince oil is normally more viscous than water. If no other information is available,field experience indicates that if 500µm water droplets are removed from the oil theresulting oil-in-water content will typically be 2000 ppm or less. Similar argumentsapply to the calculation of the constant K as was discussed for gas capacity. Thesecalculations, along with the specified liquid flow rates, allow the area required forliquid flow to be specified.
An alternative approach is to base the separation requirements on emulsion separationtests which specify a required separation time (Section 6.3 and Figure 13). Separatorsare typically run 50-75% liquid full, so knowledge of the design liquid flow rates andthe required residence times allow the separator length and diameter to be calculated.An increased separator diameter is required to compensate for surging flow as well asallowing sufficient volume to give an adequate response time between high/low levelalarms before the high/low level trip leads to a shut down. An additional 30-60seconds hold-up volume is typically specified. The resulting performance of a typicalthree stage separation process is summarised in Table 7
Separator Water Oil Target Water Target OilStage Residence Time Residence Time in Oil in Water
(min) (min) (%vol) (ppm) 1st 4 2 2 2000
2nd >6 4 1 400
3rd >8 6 0.25 100
Table 7
Performance targets for
typical 3-stage separation
system
1
24
8. TEST SEPARATOR
The production from each well has to be regularly measured accurately as part of thefield's well surveillence programme. Accurate three phase flow meters have onlyrecently become available, so traditionally a small three phase separator (the testseparator) has been used for this purpose. Here, the well production is split into itsconstituent phases (oil/water/gas); the flow rates of each measured separately withconventional orifice meters. The oil and water flows recombined for further processing.
Gas Out
Gas Measurement
Water Flow Meters
Oil Flow Meters
BSW
InterfaceControllers
RecombinedFlow
Inletfrom Well
Orifice Box
Gas Oil Water
Multi phase flow meters are now becoming reliable. This is reducing the need to installa test separator (platform space and weight saving) as well as the need to install asecond (test) pipeline in subsea field developments between the manifold and the hostfacility (module 10, figure 7).
9. COMPRESSORS
The gas liberated from the produced fluids must be delivered to the export pipeline atthe specified pressure. It is flashed from the produced liquid at a variety of pressuresduring the production process - so economically achieving this export pressurerequires a number of gas compressors rather than one large one. Minimising thecompression power requirements while maximising the liquid recovery (determinedby the thermodynamics of the produced fluids), will dictate that the gas from each oilseparation stage is compressed to the operating pressure of the previous stage. Thisis schematically illustrated in Figure 19.
Figure 18
A test separator
Department of Petroleum Engineering, Heriot-Watt University 25
8Oil and Gas Processing8
1st Stage Compressor 2nd Stage Compressor 3rd Stage Compressor
LKO LKO LKO
(Pressure= 20 bar)
Out
(Pressure= 1.5 bar)
(Pressure= 4 bar)Gas
Gas from high pressure separator
Gas from intermediatepressure separator
Gas from lowpressure separator
Vent gas from storage tanks etc.
Liquid out
Liquid Knock Out vessel Cooler Compressor LKO
KEY
(Pressure= 100 bar)
The action of the compressor performing work on the gas being compressed raises itstemperature considerably (the “bicycle pump effect”). Minimisation of the compres-sor power requirements together with maximising of the liquid recovery (whichcondenses on cooling the gas) dictates the use of interstage coolers, as shown. Off-shore, sea water is used for this cooling duty.
An example of the combined result of the above effects is illustrated in Table 8.
Separator Pressure(bar) Oil Production(m3) Compression Power *(kw)
80, 5 1300 640
80, 30, 5 1315 370
80, 30, 12, 5 1321 295
*Omission of compressor interstage coolers increases this by 50%
The prime mover or engine chosen to power the compressor depends on thecompressor’s location and the power required. Gas turbines, diesel engines andelectric motors are all frequently employed.
Centrifugal and positive displacement reciprocating compressors are both commonlyused in oil field applications. Both compressor types are susceptible to damage byliquid droplets, hence the presence of the liquid knockout vessels prior to eachcompressor.
Table 8
Effect of multiple separator/
compressor stages
Figure 19
Schematic 3-stage
compression scheme
1
26
Centrifugal compressors are large, complex machines that contain internals that rotateat high velocities. They are much more difficult to install and maintain than crude oilor water pumps. One of the critical areas in compressor design is to ensure that a re-cycle valve is included which opens to prevent compressor “surge”. This occurs whenthe feed rate is insufficient to allow the compressor to reach its design dischargepressure.
The subsequent treatment dehydration etc. of the gas to achieve export specificationsonce it has been compressed is similar to that described in gas field operations.(Section8.10).
10. OIL EXPORT
The volume of oil being exported has to be measured to the highest accuracy - sincenot only does it define the project product and cashflow, but also involves both fiscal(tax/royalty etc.) and intercompany transfer (e.g. to a “common carrier” pipeline)aspects.
Turbine flow meters - which involve the measurement of the number of times that theflowing oil revolves a paddle or turbine placed in the oil flow path - have a high,intrinsic accuracy and are normally used for this purpose. However, they have to becalibrated - or “proved” at regular intervals. A bank - say 5 to 10 - of smaller capacityturbine flow meters are used rather than a single, large meter. The calibration processis illustrated in Figure 20. The flow rate as measured by the turbine flow meter iscompared with that calculated from the time for a sphere to be displaced between twodetectors and the volume between these detectors.
Department of Petroleum Engineering, Heriot-Watt University 27
8Oil and Gas Processing8
Sphere
SphereDeflectors
Open Open
Open Shut
Shut Open
TurbineFlow Meter Tu
rbin
eFl
ow M
eter
in N
omal
Ope
ratio
n
Turb
ine
Flow
Met
er B
eing
Cal
ibra
ted
Inlet Header
Export Header
Reversing Valve
Oil export depends on efficient pipeline operation. The pipeline requires regularcleaning by a “pig”. This removes settled sand, stagnant water collected at low points(for corrosion prevention), wax deposits etc. The “pig” may be in the form of a sphereto displace fluids or a cylinder with brushes to scrape the inside surface of the line.Alternatively it may have “intelligence” in that it can inspect the pipeline condition
Figure 20
Meter calibration system for
oil export pipe line
1
28
and record the results for later replay. The “intelligent” pig uses similar techniquesas employed for well (tubing) condition monitoring {mechanical arms to “feel” forcorrosion pits and grooves, acoustic wall thickness measurement devices etc.}. As anaid to recovery, they often carry a transmitter so that their position can be pinpointedif they become stuck in the pipeline.
The pig is “launched” into the pipeline using a pig launcher (see Figure 21). This isan oversize barrel with end closure, pressure gauge and venting/purging system toallow any hydrocarbons present in the “launcher” to be depressurised/disposed of ina safe manner. Opening of any pressurised lines which could potentially containhydrocarbons is a hazardous operation which requires proper attention to safety andenvironmental aspects.
A pig “signaller” is built into the launcher.This confirms the successful launch of thepig into the pipeline. A duplicate pig “signaller” and “launcher” is installed at thereceiving end of the pipeline to record the arrival of the pig and to allow its saferemoval from the pipeline.
Drain
Purge
Pipeline
Pig Passage Signaler
End Closure
Pipeline
Pig
Vent
Pig Launcher
Pressure Guage
A B
C
Valve AStatus
ClosedNormal Pipeline Operation
Displace Pig in Launcher Open
Open
Closed
Closed
Open
Valve B Valve C
11. GAS HANDLING
Similar process units are used in oil field gas handling as are used in gas fields. In bothcases the objective is to maximise the recovery of liquid Natural Gas Liquids or NGL's(mainly hydrocarbons with a chain of four or five carbon atoms). This is achieved inoil fields by multi-stage separation and (gas) compression - as discussed in theprevious chapter. Single stage separation, with optional gas compression andprocessing as dictated by the producing wellhead pressure and gas compositionrespectively, is the norm for gas fields. Figure 22 illustrates typical gas field facilities.
11.1 Increasing NGL RecoveryThe increased NGL yield, and reduction in the water content, is achieved by coolingthe gas. The equipment used is:
Figure 21
Schematic of pig launch
system
Department of Petroleum Engineering, Heriot-Watt University 29
8Oil and Gas Processing8
(i) Refrigeration to between -30ºC and -40ºC using a conventional refrigerationplant. (Freon and Propane as refrigeration agents, respectively).
(ii) Joule-Thomson expansion - the gas will be cooled when its pressure isreduced by flowing through a throttle or choke (isenthalpic conditions). This processis favoured when there is a large pressure drop between wellhead and pipelineoperating pressures. The gas may require compression back to the export pipelineoperating pressure subsequent to the separation of the condensed liquids.
(iii) Turbo Expansion - this involves extracting energy from the gas by getting itto do work e.g. powering a compressor or generating electricity while it’s pressureis being reduced. This achives a lower temperature than for Joule Thomsonexpansion e.g. Shell’s St. Fergus gas plant cools the gas to -99ºC using two-stageturbo expansion.
(iv) Processing of Chilled LiquidsGlycol is added prior to cooling the gas so that any condensed water does not form(solid) ice or hydrates (see Section 10.3.2). The cool water/glycol/liquid hydrocarbonmixture is lead to a three-phase separator where the water/glycol mixture is removedfrom the bottom while the hydrocarbon liquid (middle layer) is pumped to adistillation column {see Section 10.2 on NGL Stabilisation for further details}.Adjustment of the distillation conditions (number of trays, bottom temperature,pressure etc.) allows ethane, propane and butane to be recovered separately e.g. foruse as a chemical feed stock. The (denser) NGL’s are recovered from the columnbottom. Cryogenic distillation conditions (<-50ºC) are necessary if recovery ofsubstantial volumes of ethane is desired.
LiquidSeparation
Dehydration
CompressorHigh PressureSeparation
Wells
Cooling
GasProcessing
GasSales
Water
Gas
Gas
Produced
Fluids
Water
Water
NGL
NGL
NGLStabilisation
Column
NGLSales
Oily WaterDisposal
Figure 22
Schematic gas field facility
system
1
30
11.2 NGL StabilisationThe separated NGL has a high vapour pressure due to dissolved, volatile gasses(methane, ethane etc.). It is not suitable for storage in a tank or onward transport tocustomers. The vapour pressure is reduced by heating the NGL to progressivelyhigher temperature and allowing the gas to “flash-off” at constant pressure (Figure23) {See section on PVT properties in Reservoir Engineering module} The sameprocess can be carried out more efficiently in a distillation column. The columnconsists of a number of trays containing bubble caps (Figure 24) which forces the gasrising upwards into intimate contact with the liquid on the tray, ensuring they cometo equilibrium. The higher trays are operated at a progressively lower temperaturethan the lower trays. Hence the liquid undergoes a series of multiple flashes atincreasing temperature and constant pressure as it trickles down the column from thetop tray to the bottom tray. Any water that is present will tend to concentrate at thetray operating at just below its boiling point under the column operating pressure. Aseparator may be installed on this tray and the water drawn off for disposal.
PC
LowTemperature
Separator
IntermediateTemperature
Separator
HighTemperature
Separator
Heater
Heater
Cooler
Stabilised NGL Storage
Unstabilised NGL
Gas Gas
Pressure Control
Packing material is frequently used as an alternative to distillation trays. These ensurethat the liquid and gas come into equilibrium by providing a large surface area withwhich the upcoming gas and the downward flowing liquid can brought into equilib-rium. There are many proprietary types of packing - some examples are illustrated inFigure 25..11.3 Gas DehydrationProduced Natural gas is saturated with water vapour under the prevailing downhole,reservoir conditions. Production to the surface and subsequent processing normallyinvolves a progressive reduction in the gas temperature (and pressure). The equilibriumwater content of the gas phase decreases as the temperature drops or the pressureincreases, resulting in the separation of the liquid water. (Figure 26). Water presentsas vapour in the gas phase does not present a problem to the gas facility operation -unlike liquid or solid water (ice):
Figure 23
Process scheme for NGL
stabilisation via multiple
flashes at constant pressure
and increasing temperature
Department of Petroleum Engineering, Heriot-Watt University 31
8Oil and Gas Processing8
Separator
Water to OilyWater Treatment
Water to OilyWater Treatment
Gas
Produced Fluid
From Wells
NGL
Heat
Reboiler
Cold Water
DistillationColumn
Cooler
NGL Storage
Gas
Gas
Gas
Gas
Gas Vent
Customers
Bubble Caps
Top Tray
SeparatorTray
Liquid water accumulates at pipeline low points, reducing well capacity and acceler-ating corrosion. It also forms solid hydrates (see Section 10.3.2) which, similar to ice,can plug lines, valves etc.
Figure 24
NGL separation and
stabilization
Figure 25
Examples of column
packing
1
32
The saturated water content of natural gas under various conditions of temperature andpressure can be derived from Figure 26 The Dew Point is the temperature at a givenpressure, at which the gas is saturated with water vapour, and any further reductionsin temperature will cause the water to begin to condense. Thus, if the water contentof a gas stream is known, its Dew Point can be derived from Figure 26.
Water Content of Hydrocarbon Gas
Temperature ºF
Wat
er C
onte
nt o
f Nat
ural
Gas
(lb
wat
er/m
illio
n cu
ft. o
f wet
gas
at 6
0º F
and
14.
7 ps
ia)
80000
60000
40000
20000
10000
8000
6000
4000
2000
1000
800
600
400
200
100
80
60
40
20
10
8
6
4
2
1
80000
60000
40000
20000
10000
8000
6000
4000
2000
1000
800
600
400
200
100
80
60
40
20
10
8
6
4
2
1
0.6
20 25 30 35 40 45 50
50 ºF100 ºF
150 ºF
200 ºF
250 ºF
300 ºF
0.8 1.0
1.0
0.9
0.8
0.71.2 1.4 1.6 1.8
Molecular Weight
Correction Factor for Gas Gravity
Gas Relative Density
CG
1.00
0.98
0.96
0.94
0.92
0.900 1 2 3 4
Total Solids in Brine. %
Cs
=
H20
Fro
m B
rine
H20
Fro
m W
ater
Correction Factor for Salinity
-60 -40 -20 0 20 40 60 80 100 120 140 160 180 200 240 280
HY
DR
AT
E F
OR
MA
TIO
N L
INE
14.7
psi
a25
50
100
200
300
400
500
600
1000
1500
2000
4000
5000
6000
1000
0
800030
00
800
14.7
psi
a25
50
100
200
300
400
500
600
1000
1500 20
00
800
Position of this lineis a function of gas composition
Water contents of natural gasses with corrections for salinity and relative density.After McKetta and Wehe, Hydrocarbon Processing. August 1958
Figure 26
Dew point of natural gas.
(adapted from Gas
Processors Suppliers
Association, Engineering
Data Book, 10th. Edition)
Department of Petroleum Engineering, Heriot-Watt University 33
8Oil and Gas Processing8
11.3.1 Dew Point DepressionDew Point Depression is the number of degrees that the dew point (or watercondensation temperature) has to be lowered by the dehydration process e.g. a 80ºFDew Point Depression for a gas at 100ºF and 1000 psi would require the removal{60 - 4.5 =} 55.5 lbs water/MM ft3 gas and implies that liquid water condensationwould not occur above 20ºF (at 1000 psi) or even -10ºF if the pressure was reducedto 200 psi.
11.3.2 HydratesHydrates are solid materials which form from light hydrocarbons and liquid water.The light, hydrocarbon molecule is embedded in the solid water crystalline lattice,which makes up the majority of the compound’s composition. Viz:
Methane CH4
• 7H
2O
Ethane C2H
6• 8H
2O
Figure 26 shows that they are formed above the normal freezing point of water andindicates the maximum temperature at which they are encountered.
One solution to preventing operational problems from hydrate formation is to add an(expensive) inhibitor e.g. methanol to the gas stream or to reduce the water contentof the gas such that liquid water cannot form under the full range of operatingconditions. Water vapour removal is achieved by:
(I) Refrigeration or cooling (see section 10.1)
(ii) Continuous absorption in a liquid desiccant (e.g. glycol)
(iii) Batch absorption by a solid desiccant (e.g. silica gel)
11.4 Continuous Dehydration ProcessThe process scheme for counter current absorption in a (packed) column using a liquiddesiccant such as (triethylene) glycol is shown in Figure 27. The wet gas enters thecontactor column at the bottom while the “lean” (low water content) glycol is pumpedin at the top. The glycol trickling down the column is brought into intimate contactwith the gas bubbling upwards. The water vapour is absorbed by the glycol since thepartial pressure of water in the gas phase is greater than its partial pressure in the liquid(glycol) phase. The lower the water content of the “lean” glycol, the lower theresulting dew point temperature and water content of the gas phase.
1
34
Glycol / GasColumn
Glycol StrippingColumn
Glycol Boiler
Gas Dehydration
"Lean"Glycol Storage
"Rich" Glycol / Water Mixture
Stripping Gas(Optional)
Reflux Condenser
Column
Glycol Regeneration
Dry Gas Outlet
Water Vapour
Low Pressure
Wet Gas Inlet
The “rich” glycol/ water mixture is drawn off from the bottom of the contactor tower.The absorbed water vapour dramatically lowers the mixtures boiling point, so that theglycol may be regenerated by boiling in a distillation column. It works in a similarmanner to those described earlier in section 10.2; except that a condenser (cooler) isplaced at the top of the columns. The temperature of the condenser is adjusted so thatthe glycol condenses and is returned as column reflux, while the water vapour passesthrough it as a gas. The “lean” glycol is drawn from the bottom of the still and cannow be used to treat further more wet gas. The water content of the “lean” glycol canbe further reduced by gas stripping. This involves bubbling a gas through the leanglycol - either in the regeneration column itself or, if higher purity glycol is required,in a separate small column prior to the gas being passed to the main, regenerationcolumn. Low pressure (say 50 psi at 60ºF) gas is used since increasing its temperatureto column operating temperatures will increase its water carrying capacity from, say,250 to 100,000 lbs/MMscf.
Triethylene glycol can typically achieve a dew point depression of 55ºC by operatingthe glycol reboiler at 205ºC. This can be increased to 85ºC by use of gas stripping.Further reductions in the dewpoint e.g. to -100ºC for gas destined for a cryogenicliquefied natural gas plant, require use of a solid desiccant.
11.5 Batch Dehydration ProcessVery low dew points, such as those required for cryogenic LNG plants where a watervapour concentration of less than 1ppm is required, can only be achieved with solidbed dehydration systems such as illustrated in Figure 28. It consists of:
(i) two or more desiccant filled contactor vessels,
(ii) a heater to supply hot, regeneration gas,
(ii) a cooler, to condense water from the used regeneration gas, together with a separator.
The solid desiccant (e.g. silica gel, alumina gel, molecular sieve etc.) has a very largespecific surface area. This allows it to absorb water from the (wet) gas passing
Figure 27
Continuous gas
dehydration
Department of Petroleum Engineering, Heriot-Watt University 35
8Oil and Gas Processing8
downwards through the contactor at low near ambient temperatures (30º - 40ºC).Once the entire bed has become saturated with water, the inlet (wet) gas is switchedto a second tower containing fresh absorbent. The saturated (spent) tower is thenregenerated by passing heated (260º - 350ºC) gas through it. The water absorbed bythe solid dessicant is vaporised into the hot gas stream. Cooling of this hot, wet gasreduces the water saturation level and liquid water is recovered from the separator. Ahot tower will not work efficiently. The dessicant bed needs to be cooled - by passingcool inlet gas - to its operating temperature (30º - 40ºC) before it will work efficiently.The “fresh” tower is then available for operation when the tower currently in usebecomes saturated with water vapour. The towers are typically sized so that they workon an 8-hour adsorption cycle followed by regeneration through six hours heating andtwo hours cooling.
Inlet, Wet Gas Cooler
Heater
Regeneration Gas
DehydratedSales Gas
Cooling Gas
Tower BeingRegenerated
AbsorbingTower
in Operation
Separator
Water
Regenerationand Cooling Gas to Compressor
11.6 Acid Gas TreatingNatural gas frequently contains other contaminants than those discussed to date.These include carbon dioxide (CO
2), hydrogen sulphide (H
2S) and other sulphur
compounds such as mercaptans. Since these materials form acidic solutions whendissolved in water, they are known as acid gasses. These compounds are undesirablesince they:
(i) Cause corrosione.g. CO
2 is corrosive in the presence of liquid water at a partial pressure (=total
pressure * mole% of CO2) of 30 psi. H
2S can lead to sulphide stress cracking and
hydrogen embrittlement of many metals at partial pressures as low as 0.05 psi}
Figure 28
Schematic flow diagram for
solid bed dehydration
1
36
(ii) reduce the heating value of the gas and,
(iii) in the case of H2S, can be poisonous in quite low concentrations
(H2S has a distinct odour at 0.15 ppm, exposure to 100 ppm H
2S leads to drowsiness
after 15 minutes while 500 ppm causes sufficiently severe breathing problems after5 minutes that prompt the requirement of artificial respiration. Unfortunately, H
2S
cannot be smelt at these lethal concentrations.
The maximum allowable CO2 and H
2S concentrations are normally specified in the
gas sales contract - typical values are 3% and 4 ppm respectively. Low concentrationsof H
2S may be removed by solid absorbents (e.g. iron oxide, zinc oxide) which are
replaced when the bed is spent. Higher concentrations of H2S as well as CO
2 are
removed in solvent extraction process similar to a glycol dehydration unit (section10.4 and figure 26).
The extraction unit consists of two parts - an adsorption column in which the acidicgas to be treated is fed in at the bottom and the liquid solvent is added at the top. Thecolumn internals - trays or packing - ensure that there is intimate contact between thetwo phases and the process conditions adjusted so that the CO
2 / H
2S concentrations
are reduced to specification levels.
One class of solvents used for treating acid gasses are alkaline liquids that reversiblyreact with the acidic gases H
2S and CO
2 e.g. mono- ethanolamine.
2 RNH2 + H2S (RNH3)2 S
(RNH3)2 S + H2S 2(RNH3)HS
2 RNH2 + CO2 RNHCOONH3 R
heat
heat
heat
An alternative is a physical solvent e.g. Sulfinol in which one can dissolve largequantities of CO
2 and H
2S at low temperature; while releasing them again at high
temperatures. Thus, in both cases, a solvent “rich” in absorbed CO2 and H
2S is
recovered from the bottom of the adsorption column and transferred to the regenera-tion column, where heat liberates the absorbed CO
2 or H
2S. The regenerated solvent
is used again. The gasses are vented or flared {H2S being converted to sulphur dioxide
(SO2)}. However, there are normally strict environmental constraints on the quanti-
ties of H2S and SO
2 that can be released into the environment; while the emission
of CO2 attracts a tax penalty in some countries. Alternatives are:
(i) Injection into an underground disposal reservoir
(ii) Conversion of the H2S into solid sulphur which can be sold to the chemical
industry. The Claus process is frequently used - this involves the oxidation of partof the H
2S to SO
2; followed by conversion to sulphur by reaction with further H
2S.
2 H2 S + 3 O
2 → 2 SO
2 + 2H
2 O
SO2 + 2H
2S → 3 S + 2H
2O
Department of Petroleum Engineering, Heriot-Watt University 37
8Oil and Gas Processing8
12. FURTHER READING
Arnold K and Stewart M. "Surface Production Operations"Volume 1 (2nd Edition, 1998) and Volume 2 (1st Edition, 1989)Published by Gulf publishing CompanyISBN 0-88415-821-7 and 0-87201-175-5
Kennedy J "Oil and Gas Pipeline Fundamentals"Published by Penwell Books, 1993ISBN 0-87814-390-4
.
Well Control 11
C O N T E N T S
1. CHEMICAL PROPERTIES OF FORMATION/ PRODUCED WATER.1.1. Mineral Scale1.2. Scale Inhibiton1.3. Corrosion1.3.1. Carbon Dioxide Corrosion1.3.2. Oxygen Corrosion1.3.3. Hydrogen Sulphide
2. PRODUCED WATER TREATMENT2.1. Introduction2.2. (Corrugated) Plate Interceptors2.3. Flocculation / Coagulation2.4. Flotation2.5. Hydrocyclones2.6. Coalescer Units2.7. Centrifuges
3. OTHER SOURCES OF (PLATFORM)WATER REQUIRING TREATMENT
4. DISPOSAL OF PRODUCED WATER4.1. Marine Discharge4.1.1. Production Chemicals4.2. Reducing Environmental Impact of
Produced Water Discharge5. WATER INJECTION
5.1. Water Sources5.2. Water Quality5.3. Injection Water Treatment5.3.1. Example Process Schemes Employing
Coarse and Fine Filtration5.3.2. Suspended Solids Removal5.3.3. Filter Systems5.3.4. De-Oxygenation5.3.5. Hydrogen Sulphide Removal5.3.6. Chemical Treatments
6. FURTHER READING
9Water Handling9
1
2
LEARNING OBJECTIVES:
Having worked through this chapter the student will be able to:
• Discuss the (highly) variable chemical composition of formation water
• Describe operational problems associated with production water (scale, corrosion, etc)
• Identify equipment used to treat oily water separated by the primary gravityseparation equipment (corrugated plate interceptor, flocculation, coagulation,(gas) flotation and hydrocyclones and coalascers)
• Discuss the disposal options for produced water
• Identify the sources of water to be used for water injection
• Outline the processes required to “clean up” water for water injection (filtration,deoxygenation, chlorination, desulphation etc)
Department of Petroleum Engineering, Heriot-Watt University 3
9Water Handling9
1. CHEMICAL PROPERTIES OF FORMATION / PRODUCED WATER
Formation water, and hence produced water, show a wide range of chemicalcompositions. The formation water and their formation rocks are in equilibrium withone another under the reservoir conditions of temperature and chemical pressure,where they have normally been in contact with one another for a long time. The rangeof formation water chemistry encountered in the North sea is illustrated in Table 1.The first three fields can be regarded as normal with the relatively low salinity {TotalDissolved Solids (TDS) 25,000 - 60,000 ppm}. The latter three fields illustrate themore extreme values encountered in high salinity brines with high cation levels ingeneral, in barium cation (Ba++) concentration and in sulphate anion (SO
4)
levels
respectively.
1.1 Mineral ScaleThe production process involves bringing formation fluids from their original hot,high pressure reservoir conditions to the surface. This temperature and pressurereduction can result in the solubility of one or more minerals in the water phasebecoming less than the amount present in the water. The mineral precipitates as a solidphase (scale) which separates from the water.
A second source of scale is mixing of:
(i) Incompatible formation waters (if more than one reservoir zone is beingproduced at the same time), or
(ii) (Incompatible) formation water and injection water (frequently sea water) arebeing produced from different perforations producing from the same formation
Frequently, the amounts of scale that can be produced are apparently small - in theorder of mg solid per litre of water. However, a large number of litres are producedfrom a well in a short period of time. Thus the potential rate of deposition {or volumeof solid scale produced in unit time} is large. These solid precipitates may eitherremain as a suspension in water or form a layer adhering to the pipe wall. Such layersrestricts fluid flow in the tubing, as well as plugging restrictions such as (narrow)orifices etc.
(Calcium) Carbonate (CaCO3) scales are normally caused by decreases in the calcium
bicarbonate containing fluids temperature and pressure.
Ca (HCO3)2 CaCO3 (dissolved) + H2O + CO2 (dissolved)
→ Ca CO3 (solid) + H2O + ↑ CO2 (gas)
The partial pressure of Carbon Dioxide (CO2) above the liquid phase, and hence the
concentration of CO2 that will dissolve in the liquid, will decrease when the pressure
is reduced. CO2 will be liberated from the liquid to the gas phase; driving the above
equilibrium to the right and precipitating part of the CaCO3. The presence of other
salts e.g. a sodium chloride concentration to 200,000 ppm will significantly increasethe solubility of CaCO
3 {up to 250% for this example}.
1
4
Exa
mp
les
of
the
ran
ge
of
No
rth
Sea
Oil
Fie
ld F
orm
atio
n W
ater
Ch
emis
try
Io
n (p
pm)
Fie
ld
Na+
K
+
Mg+
+
Ca+
+
Sr+
+
Ba+
+
Si
Fe
C
l'
SO
" 4 H
CO
' 3 p
H
TD
S*
C
omm
ents
Bea
tric
e 19
,000
12
5 25
0 1,
900
165
12
- 0.
1
35,5
30
- 46
5 7.
8 57
,440
Trol
l 18
,469
43
8 40
8 1,
856
324
14
- -
33
,288
-
- 6.
4 5
4,98
9
Bre
nt
9,00
0 20
0 50
25
0 25
60
-
-
14,3
00
10
1,05
0 -
24,9
45
Rav
ensp
urn
69,2
00
1,50
0 3,
700
25,5
00
1,10
0 15
55
16
0
142,
200
260
3.
7 26
5,77
0 H
igh
TD
S
and
Cat
ions
Mill
er
28,8
00
1,82
0 11
5 1,
060
110
1,03
0-
10
47,6
80
7 2,
070
6.7
82,6
90
Ext
rem
e B
ariu
m
Pip
er23
,800
34
1 66
7 3,
480
156
1 -
-
41,8
90
1,84
5 69
1 -
71,8
27
Hig
h S
ulph
ate
(Car
boni
fero
us)
Sea
wat
er10
,890
462
1366
428
80
--
1969
929
6212
36.
5-
Not
e su
lpha
te
leve
l
* T
DS
- T
otal
Dis
solv
ed S
olid
s (p
pm o
r m
g / l
)
extr
acte
d fr
om :
‘Nor
th S
ea F
orm
atio
n W
ater
Atla
s’ e
dite
d by
E.A
. War
ren
and
P.C
. Sm
alle
yG
eolo
gica
l Soc
iety
Mem
oir
No.
15
Pub
lishe
d by
The
Geo
logi
cal S
ocie
ty, 1
994
Table 1
Examples of the range of
North Sea oil field
formation water chemistry
Department of Petroleum Engineering, Heriot-Watt University 5
9Water Handling9
By contrast, chemical incompatibility between injected sea water and formation wateris the cause of the deposition of sulphate mineral scales. This occurs because sea watercontains reasonable concentrations of Sulphate anions (up to 2,800 ppm) but is lowin divalent cations {420 ppm Calcium (Ca++), trace Ba++ and Strontium (Sr++)}. Bycontrast, many formation waters contain significant concentrations of barium - fromtens of parts per million in the Brent field formation waters to thousands in the Millerfield. The solubility of Barium Sulphate (BaSO
4or Barite) is very low, being only 4%
of that of calcium carbonate. Barite is precipitated by the reaction:
Ba++ + SO"4 → Ba SO
4 ↓
BaSO4 is one of the most insoluble of the scaling minerals. A similar problem is
encountered with Strontium Sulphate (celestite or SrSO4) in some fields.
A less frequently encountered scale is Calcium Sulphate (CaSO4). This is due to the
unusual solubility behaviour of Gypsum (CaSO4. 2H
2O), the most commonly encoun-
tered form of calcium sulphate. Gypsum has a solubility maximum at 40°C (i.e. itshows reduced solubility at both higher and lower temperatures). The issue iscomplicated by the fact that the equilibrium form above 40ºC is Anhydrite (Ca SO
4);
which is even less soluble. A further complication is that this transition temperatureis itself dependent on the salinity. Further information on this complicated system canbe found in chemical textbooks.
The key facts summarising formation of the above and other scales, are summarisedin Table 2.
Typical Causes and Types of Inorganic Scales Encountered in Oilfield Operations
Scale Formula Principal Cause of Scale Formation
Calcium Carbonate (Calcite) Ca CO3 Reduced partial pressure of CO2 due to pressure reduction
Barium Sulphate (barite) Ba SO4 Mixing of formation waterand Sea Water
Strontium Sulphate (celestite) Sr SO4
Calcium Sulphate (gypsum) Ca SO4.2H2O Temperature change - maximum solubility at 40ºC
Ferrous Sulphide FeS Corrosion followed by contact with Hydrogen Sulphide
Salt NaCl Temperature and pressure reduction {some (liquid) water also vaporises into gas phase}. Gas wells only
Sulphur S Temperature and pressure reduction. Sour gas wells only
Table 2
Typical causes and types of
scales encountered in
oilfield operations
1
6
1.2 Scale InhibitonThe exact location at which scale will actually form is a function of many factors -temperature and pressure changes, mixing patterns, water chemistry, precipitationkinetics etc. Scale formation is observed in the reservoir formation, the perforations,producing well tubing, pumps and topside facilities. Formation damage (permeabilityreduction) in the near wellbore area of production wells has also been observed.
An effective scale management programme consists of:
(i) Predicting whether scale formation is likely to occur in the lifetime of theproducing asset based on the Formation and Injection Water Chemistry andpredicted producing conditions in the reservoir, well and the facilities
(ii) Identification of when significant changes in producing conditions haveoccurred that could initiate scale formation e.g. sea water breakthrough into aproducing well
(iii) Instituting a scale inhibitor injection programme.A scale inhibitor is a chemical e.g. a polyphosphonate, which has the ability toprevent the (tiny) barium sulphate seed crystals from growing large enough thatthey can form bulk precipitates; i.e. scale inhibitors reduce the speed (kinetics)of solid scale formation, but do not alter the final equilibrium (thermodynamics).They achieve this by absorbing on the active sites of the growing bariumsulphate crystal. This (temporarily) inhibits the accretion process which leadsto crystal enlargement. The (minute) seed crystals are thus unable to sticktogether and are transported from the well with the produced fluids.
The scale inhibitor can be injected:
(a) Downhole: the inhibitor is injected (squeezed) into the formation where it isabsorbed onto the formation rock. It is then re-dissolved in the produced fluidat low concentration over a period of many months. Scale inhibitors show a“threshold” activity level i.e. if they are present above this level in the aqueousphase they will inhibit scale formation. Thus a monitoring programme needsto be instituted to measure the inhibitor level in the produced water. Providingthe inhibitor concentration remains above a threshold value, it will protect thenear wellbore formation and the production tubing from scale formation. Afurther inhibitor treatment is required when the concentration in the producedwater becomes too low (approaches the threshold (or minimum) inhibitoractivity level.
(b) Inhibitor may be continuously injected into the tubing at a low rate throughan injection valve situated near the bottom of the well. The inhibitor is carriedto the injection valve via the casing / tubing annulus. This, obviously, does notprovide protection to the (near wellbore) formation, perforations or to thecasing / tubing below the injection point.
(c) At the wellhead to provide further protection to the facilities.
Department of Petroleum Engineering, Heriot-Watt University 7
9Water Handling9
1.3 Corrosion
Corrosion of production facilities costs operators many millions of pounds every year.Corroded downhole equipment - tubulars or valves etc. are difficult and expensive toreplace while corrosion of the well’s casing can threaten the integrity of the well itself.Corrosion of surface facilities is in principle easier to repair - since access is simpler- but can (potentially) cause the loss of production (income) from all wells tied intothat facility; as well as resulting in significant environmental damage if the operatorwas not aware that he had a significant corrosion problem.
Iron bacteria deposit a sheath of iron hydroxide around them as they grow. This isobtained from soluble iron ions in water. Once established, an iron bacteria colonycan precipitate large quantities of plugging iron hydroxide as well as provide a hostenvironment for other bacteria, e.g. sulphate reducers, that lead to localised pittingcorrosion.
1.3.1 Carbon Dioxide CorrosionCarbon dioxide (CO
2) dissolved in water can form a sufficiently strong acidic solution
to enable steel to dissolve in water if the CO2 partial pressure is greater than 30 psi.
CO2 + H
2O → H
2CO
3 (carbonic acid)
H2 CO
3 + Fe → FeCO
3 + H
2
CO2 corrosion is characterised by a uniform attack which gradually reduces the metal
thickness.
1.3.2 Oxygen CorrosionOxygen (O
2) dissolved in water, as may occur in water injection projects or in process
equipment with operating pressures below one atmosphere, is highly corrosive:
4 Fe + 6H2O + 30
2 → 4 Fe(OH)
3
Oxygen corrosion is more aggressive than CO2 corrosion since it is not uniform - it
leads to pits in the metal surface which can rapidly grow to form holes which piercethe metal itself.
1.3.3 Hydrogen SulphideHydrogen Sulphide (H
2S) partial pressures greater than 0.5 psi, can lead to both metal
loss corrosion and to hydrogen embrittlement:
H2S + Fe → FeS + 2H•
2H• → H2
Hydrogen embrittlement involves the steel loosing its ductability. Soft steels, whichare capable of relieving the internal stresses caused by atomic hydrogen penetratingthe (crystal) structure and subsequently forming molecular hydrogen, are less andsusceptible to this form of attack.
Any form of dissolved iron present in the water phase due to corrosion processes willbe precipitated by contact with hydrogen sulphide:
1
8
Fe++ + H2S → 2H+ + FeS ↓
2Fe+++ + 3H2S → 6H+ + Fe
2 S
3 ↓
The iron sulphides have a very low solubility in water. These solids can form copiousplugging precipitates which create extensive formation damage.
2. PRODUCED WATER TREATMENT
The large changes in the volumes of produced water which are typically encounteredduring the life of a production project are summarised in table 3. The facilities haveto be designed to operate efficiently over these wide range of conditions.
Phase Fluid Early Mid-Life Mature Abandonment
Oil (bopd) 100k 100k 30k 10k
Gas (MM sft3 / d) 100 100 30 10
Water (bwpd) <500 30k 100k >100k
Water cut <0.5% 30% 77% >90%
2.1 IntroductionGravity separators - the normal method for separating water from the co-produced oiland gas phases - was discussed in section 6 of chapter 8 describing the "Fundamentalsof Oil and Gas Processing" The basic mechanism of the separator operation wasdescribed. We also discussed that this primary separation may be enhanced by:
(i) Heating of the crude oil to enhance water separation via viscosity reduction,
(ii) Addition of demulsification chemicals which alter the interfacial tensionbetween the oil droplets and the water; so allowing the emulsion to break and thewater droplets to settle. These emulsions can have a much greater (up to 100 times)viscosity than that of either the water or oil phases. This partly accounts for theirstability.
(iii) Electrostatic separation may be used to further reduce the water content ofrelatively dry oil. The water droplets suspended in the oil carry a small electricalcharge. Their rate of settling can be increased by imposing the appropriate electricalfield across (part) of the settling region inside the separator. The gravity separatorhas now become an electrostatic separator - it is not widely used but is occasionallyemployed in conjunction with the more difficult to separate, typically denser, crudeoils. Electrical short circuiting occurs when the water content becomes too high, sayabove 10% volume.
Even with the above enhancements; the first stage gravity separator leaves typically500 - 2000 ppm oil still present in the water. A further treatment / processing step(s)are required before the water is fit for disposal e.g. the oil content has to be reducedto the 40 ppm average that is required by legislation in N.W. Europe. The following6 different process schemes (sections 9.2.2 to 9.2.7) that have been developed toreduce this oil content of this oily-water.
Table 3
Typical changes in
production rates during the
lifetime of a 100,000 bopd
oil field
Department of Petroleum Engineering, Heriot-Watt University 9
9Water Handling9
2.2 (Corrugated) Plate InterceptorsPlate interceptors work by reducing the distance required for a droplet to migratebefore it comes into contact with other oil droplets and coalesces. The interceptor packconsists of a series of closely spaced, parallel plates mounted at an angle of 45º (Figure 1).These plates have a corrugated profile (similar to the well known roofing material -figure 2). This profile allows the oil to collect at the high point, where the individualdroplets coalesce and the resulting larger droplet moves upwards, under the influenceof gravity and against the direction of water flow.
Oil Outlet
Oil Weir
Inlet WeirWaterOutletWeir
WaterOutletPipe
Oily WaterInlet Pipe
Oil Layer
Oil Droplet
(Corrugated) Plate Interceptor
Small oil droplets rising under gravity
Large oil droplet formed by coalescence of many small droplets on top surface of corrugated plate
The corrugated plate interceptor works off because of the short separation distancebetween the plates. The ability of the oil droplets to move upwards through the waterphase increases rapidly (with the square of the oil particle diameter) as the droplet sizeincreases. Hence the overall rate of oil-water separation will be increased if the oildroplets can be encouraged to grow larger in a shorter time. The oil layer collectedat the surface, is skimmed off by an oil outlet weir.
The configuration of the water inlet and outlet pipes is so arranged that any turbulencethat might lead to re-mixing of the small oil droplets is minimised. Removal of oildroplets as small as 60 µm is frequently achieved. These plate separation packs mayalso be installed inside gravity separators so as to enhance their performance.
Figure 1
(Corrugated) Plate
Interceptor
Figure 2
Oil droplets coalescing in a
corrugated plate
1
10
2.3 Flocculation / CoagulationThe separation of suspended liquids (oil) and solid material can be enhanced byartificially increasing their size and their ability to coalesce. This is done by chemicalflocculation. This process uses a chemical (such as Ferrous Sulphate) which forms avoluminous precipitate in contact with water. This precipitate has the ability tocoagulate into large flocs and, in the process, destabilise any suspended particles at thesame time. A typical flocculation unit is shown in Figure 3 - it consists of a mixingchamber and a floc growth chamber. This growth chamber gives the microflocs,formed in the mixing chamber, the opportunity to collide, grow and form large flocs.This process can be speeded up by the addition of floc growth catalyst. These areproprietary chemicals available from service companies.
Floc GrowthFloc Formation
Floc GrowthCatalyst
Paddle Mixer Paddle MixerOutlet
WaterInlet
Flocculant(e.g. Fe SO4
The flocs are subsequently removed from the water stream by settling / sedimentation(gravity separation) or by flotation (see section 9.2.4). Organic polymers withpolyectrolyte properties work efficiently (only 1-10 ppm addition required) in salinewaters. These polymers can be cationic, anionic or amphoteric - the optimum choicedepending on the water salinity and the nature of the solid particles. This type of unitis mainly used for cleaning water prior to re-injection.
2.4 FlotationGas flotation uses accelerated separation via the buoyancy (density reduction) effectobtained from rising gas bubbles attaching themselves to the oil droplets. Once theyhave been carried to the surface they can coalesce, followed by separation of the oilusing a weir. The gas can be added to the oily water in two manners:
(i) Injected into the water and dispersed by a rapidly rotating impeller (dispersedgas flotation - see Figure 4). It is normal practice to arrange a number (e.g. four) ofthese units in series. A typical design criteria is that they should be capable ofremoving oil droplets larger than 15 µm.
(ii) Dissolved in the water under high pressure. Oily water is mixed with the gas orair in the pump suction. It travels via a pressure vessel where a limited residence time
Figure 3
Flocculation unit
Department of Petroleum Engineering, Heriot-Watt University 11
9Water Handling9
allows the gas to dissolve. When the pressure is rapidly reduced - by passage of thewater through a throttling valve - the gas comes out of solution in the form of manysmall bubbles (the champagne bottle effect) - see Figure 5.
Oil Froth Gas Injection
Perforated Hood
Rapid Rotation of Paddle Produces Small Gas Bubble
CleanWater OutletOily Water Inlet
Oil Outlet
Gas / Oil Bubbles
In both cases the oil froth is removed at the top and denser sediment / solids from thebottom. The water is re-cycled a number of times to improve the unit’s effectiveness,with cleaned water being drawn from the bottom of the flotation unit.
Water Recycle
Oily Water
Clean Water
Gas or Air
Gas or AirPUMPPRESSURE VESSEL
(provides residence timefor gas adsorption)
PressureReducing
Valve
Gas Vent
Oil
Gas Bubbles
SedimentSolids Removal
Oil Froth
2.5 HydrocyclonesHydrocyclones have become the standard device for cleaning oily water. Developedin the early 1990s, they work by using centrifugal force to increase the effect ofgravity. They have proven to be compact, efficient, and - with the increasing use ofFloating Production and Storage Vessels to develop marginal oil fields - operateindependently of platform motion.
Figure 4
Dispered gas flotation
Figure 5
Dissolved gas flotation unit
1
12
A schematic diagram of a hydrocyclone is included as Figure 6. This long, thin devicehas an upper inlet, where oil-water feed flows tangentially into the upper, swirlsection. The oil- water swirls around in a circular path, and is then accelerated as theflow enters the taper section with it’s reducing diameter. This swirling action leadsto an increase in the effective gravitational component, in the basic separationequation. The lighter hydrocarbon phase migrates to the central core which flowsupwards and exits the device via the upper outlet. The (cleaned) water phase passesinto the tail section equipped with a back pressure valve to control the pressure dropacross the device, and then exits the hydrocyclone.
Hydrocarbon (Lightphase) Outlet
Water Outlet (Underflow)
Oily-Water Inlet
SWIRL SECTION(with circular flow)
TAPER SECTION
TAIL SECTION
Reverse FlowingCentral Core
Lighter PhaseMigrates to Central Core
Back PressureDevice Figure 6
Schematic of a
hydrocyclone
Department of Petroleum Engineering, Heriot-Watt University 13
9Water Handling9
Hydrocyclone performance is sensitive to:
(i) Feed rate - excessively high flow rates reduces performance by causing turbulence.
(ii) Pressure drop - this is controlled by the back pressure device mounted in thetail section. Proper adjustment of the pressure drop maintains the efficiency of thedevice’s performance. This results in a large turn down ratio i.e. the hydrocyclonewill separate oil from water efficiently over a wide range of flow rates.
(iii) High oil content in the oily-water feed - this reduces the separation efficiency.
In practice, a number of cyclones are mounted in parallel in a pressure vessel withcommon inlet and outlets that are suitable for modular construction.
A typical modern flow scheme for produced water is shown in Figure 7 - the oily-waterphases recovered from the first and second stage separators are treated in separatebanks of de-oiling Hydrocyclones. The addition of a de-oiling chemical to thehydrocyclone feed is optional and based on the actual producing circumstances. Theoil recovered from the Hydrocyclones is passed to the low pressure separator whilecleaned water is ready for disposal after a further de-gassing stage.
It has become the “equipment-of-choice” replacing gas flotation for cleaningproduced water. This is due to its compact size, efficient performance and it’soperational robustness in the presence of solids.
(Optional)DemulsifierChemical
(Optional)De-oiling Chemical
(Optional)De-oiling Chemical
Disposal Water
De-oiling Hydrocyclones
De-oiling Hydrocyclones
Gas
Gas
Oil Oil
Gas
Oil plus some Water
Oil Recovery
1st Stage Gravity Separator
Produced Water De-gasser
2nd Stage Gravity Separator
Oily Water
Disposal Water
Oily Water
Export Crude OilFrom Wells
2.6 Coalescer UnitsThere are a number of designs of coalescer units. The basic concept is to provide a(usually oleophilic) surface on which the small droplets of oil can collect, grow andeventually break free and be removed for subsequent separation. They are capable ofproducing the lowest oil concentrations (5 ppm oil in water has been achieved in ideal
Figure 7
Modern scheme for clean
Produced Water
1
14
circumstances). However, they are very susceptible to plugging by organic (wax,asphaltenes etc.) or inorganic solids (scale, formation material, corrosion productsetc.). They can be (partially) regenerated by backwashing - though disposal of thebackwash water may present a problem.
One design is to pack the coalescer vessel with granular, fibrous material. The (small)input oil droplets coalesce on the granular fibres. The resulting, larger oil droplets areseparated downstream of the coalescer. A second type of coalescer is where the oily- water feed is passed down a filter tube. The properties of the semi-porous materialare adjusted so that water can pass through it, concentrating the remaining oil /emulsion phases which can then be more efficiently processed by other means, suchas gravity separators.
2.7 CentrifugesHydrocyclones are typically capable of removing 50% of the 10 µm oil droplets risingto 100% of the 20 µm ones. The principle of enhanced gravitational force employedby Hydrocyclones can be further extended by use of centrifuges where an externalelectric motor is used to spin the fluid at high velocity together with a suitably designedinternals to promote oil/water separation. Experience has showed that centrifugeshave the following advantages:
(i) more effective than Hydrocyclones (complete removal of all oil droplets largerthan 5µm is claimed),
(ii) their operation is not dependent on the available fluid pressure drop,
(iii) the above has to be weighed against their increased space, weight, power andmaintenance requirements.
3. OTHER SOURCES OF (PLATFORM) WATER REQUIRINGTREATMENT
Produced water makes up the bulk of the water that has to be processed for disposal.Particularly in offshore operations, there are other sources of water that requiretreatment before disposal. These include:
(i) Water used for washing / cleaning of equipment,
(ii) Sea spray and rain water,
(iii) Utility water previously used for heating and cooling duty,
(iv) Displacement water from crude oil storage systems and shuttle tankers. (Thestorage cells are kept full of water which is displaced by the produced oil. Theprocess is reversed when the oil is pumped out for onward transport.)
Care needs to be taken with respect to compatibility problems when mixing waters ofdifferent chemical compositions (see section 1.2 on scaling). Scale inhibitors may berequired if severe problems are encountered.
Department of Petroleum Engineering, Heriot-Watt University 15
9Water Handling9
4. DISPOSAL OF PRODUCED WATER
4.1 Marine DischargeHistorically, the vast majority of offshore produced water and platform water hasbeen disposed of to the marine environment after reducing the oil content to at leastthat specified by the national legislation e.g. (40 ppm oil in water in N.W. Europe witha voluntary company limit of 30 ppm agreed for the UK Continental Shelf for 1999).Onshore, disposal has been by underground injection, either returned to the producingzone as part of a water flood / pressure maintainance scheme or injected into adedicated disposal well or interval. Although the contribution of offshore operationsto the amount of oil reaching the North Sea is low (table 4); the amount contributedby produced water is rising since the maturity (increasing number of years inproduction and average water cut) is continually increasing. E.g. In 1985, producedwater contributed about 10% of the oil entering the North Sea due to Drilling andProduction Operations. This had increased to more than 50% by 1995. (see chapter10.4 - Environmental Impact of production Operations).
Source Input
(million tonnes per annum)
Marine Transportation 1.13 - 2.13
Offshore Production 0.08 - 0.2
Coastal Oil Refineries 0.2 - 0.3
Industrial Waste 0.3 - 1.98
Municipal Waste 0.3 - 0.45
River Runoff 0.3
Natural Seeps 1.6
Atmospheric Rainout 0.6 - 9.0
Total Pollutants 2.06 - 4.91
This rise in the importance of produced water as a source of oil emission to theenvironment has lead to the desire to implement produced water re-injection inoffshore operations. This is an active topic of research and development at the present.The water will mainly be as a (partial) replacement for the large volumes of sea waterinjected for the purpose of maintenance and improved reservoir sweep (see section 5.3).
4.1.1 Production ChemicalsMany production chemicals that are injected into the produced fluids, both downholeand at the surface into the production facilities. Most of these chemicals aresurfactants and are partially or fully soluble in the water phase. This implies that they(or their residues) may also require final disposal with the produced water. There arestrict rules and regulations with respect to environmental acceptability testing of the(fresh) production chemicals - but there are currently no standards with respect to thetesting of the produced water. In fact, while figures for the tonnage of chemicals usedare available (table 5), there are no generally accepted accounting standards tomeasure the fraction of this figure that is disposed of to the environment.
Table 4
Oil in the marine
environment
1
16
4.2 Reducing Environmental Impact of Produced Water DischargeThe environmental impact of produced water can be minimised by:
(i) Replacing the use of fresh injection water (see section 5) by produced water
(ii) Reduce the oil content of produced water. The relatively recent developmentof the hydrocyclone has contributed to increasing the ability of the operators tomeet the discharge regulations. Further reductions of this figure (20 ppm oil-in-water monthly average with a maximum of 100 ppm) are currently underdiscussion
(iii) Reduce the volume of produced Water which has to be disposed of
Production Chemical Usage (in Tonnes)
Norway 1988 UK in 1987
Biocides 1372 2316
Corrosion Inhibitors 1062 1100
Oxygen Scavengers 997 4348
Scale Inhibitors 766 4025
Emulsion Breakers 174 2132
Oil Removers 8 293
Defoamers 128 34
Flocculating Agents 166 108
Surfactants 0 53
Cleaning Agents 13 552
Gas Treatments - Glycol, Methanol etc. 2234 9237
C.M. Hudgins “Chemical Use in North Sea Oil and Gas E & P” J. Petroleum Technology, SPE, Jan 1994
The major International Oil companies are currently producing, treating and disposingof a similar volume of produced water as their oil production. More maturecompanies, with mainly USA production, may have average water cuts of 75% ormore. Total produced water costs are typically US $ 0.20 - 0.75 / bbl for lifting,treating and disposal. This is a very significant economic incentive for reducing itsvolume. This is being achieved in a number of different ways:
(i) Water shut off in the well (cement squeezes, scab liners, opening / closingdownhole flow control valves etc).
(ii) Water shut off in the near wellbore formation or at greater depth from the wellby injecting chemical treatments (e.g. total blocking of the pores with animpermeable polymer gel).
Both these methods require identification of the completion zone that is producingexcessive amounts of water. Further technologies that can be employed are:
Table 5
Chemicals in the
environment
Department of Petroleum Engineering, Heriot-Watt University 17
9Water Handling9
(iii) Re-design of the wells so that they produce at a greater oil / water ratio e.g. useof horizontal wells which can be placed at the maximum possible distanceabove the oil / water contact while the resulting reduced drawdown achievedby the long exposure to the reservoir minimises water coning effects.
(iv) Downhole water separation of the bulk of the “free” water using hydrocyclonetechnology followed by injection in a separate lateral (Figure 8). The disposalzone may be above or below the production zone. One or two electricsubmersible pumps may be used too depending on the depths of the productionand disposal zones, the formation pressure, and permeability’s as well as theflow rates. Over 100 installations of this type had been installed by 1999.
Casing
Tubing
Oil Rich Stream(75% Water)
Separated "Free" Water
Produced Oil (4%) andWater(96%)
Producing Zone
Disposal Zone(Open Hole)
Downhole Hydrocyclone Separator and One or Two Submersible Pumps Run from a Single Motor
5. WATER INJECTION
Disposal of produced water was discussed in the previous section. However, largevolumes of water are also injected into the sub-surface for pressure maintenance andimproved oil recovery. North Sea light oil reservoirs, which might recover 20 - 30%of the original oil in place under primary (pressure depletion) can have their recoverydoubled to 40 - 50% by use of water injection. The processes involved are:
(i) Sweep of mobile oil from injection well to production well.
(ii) No loss of oil by maintaining formation pressure above the bubble point.
(iii) Improved cash flow / profitability by increasing early oil production andreducing the project lifetime.
(iv) Reduced / no requirement for artificial lift.
Figure 8
Downhole Separation /
Injection
1
18
The reservoir engineering study carried out as part of the field development plan willdictate where, how much, and at what rate the water should be injected. The task ofthe Production Technologist is to write the injection water specifications so that it willnot be:
(i) Corrosive to the injection well facilities and
(ii) Compatible with the injection formation so that the injectivity is maintainedover the injection project lifetime. Large volumes (30,000 bwpd) may be injectedinto each well while the total field requirements can be up to 1,000,000 bwpd in thecase of the giant, Middle East oil fields.
5.1 Water SourcesThere are often several potential water supply sources that are frequently used. Thewater from the different sources will have different properties and will imposedifferent treatment costs to make it suitable for injection. When a source is chosen,studies are required so that there is sufficient water available for the project lifetime.Sources include:
(i) Sea water
(ii) Fresh surface water
(iii) Aquifer water (not oil producing zone)
(iv) Produced water (from oil producing zone)
Surface waters will normally show seasonal variations in their quality. Henceanalyses need to be made over a complete 12-month cycle. Careful positioning of thewater inlet will help minimise water treatment costs, e.g. a seawater intake is bestsighted upstream of platform discharge points, e.g. of drill cuttings and drilling mud,and at below the depth to which summer algal blooms occur; while being sufficientlyabove the sea bed to be unaffected by bottom sediment stirred up by currents, storms etc.
5.2 Water QualityThe principal factors which define the water quality are:
(i) solids - dissolved or suspended
(ii) dispersed oil
(iii) dissolved gasses
(iv) bacteria
Their likely effect and solution of some quality issues are is summarised in table 6. Inaddition, some receiving formations contain clay particles which swell and dispersein the presence of fresh water. This leads to blockage of pore throats and reducedpermeability. Hence, prior to the implementation of a water injection project, the
Department of Petroleum Engineering, Heriot-Watt University 19
9Water Handling9
compatibility of the injection water and formation should be tested by pumpinginjection water through a formation core sample and measuring any permeability changes.
Issue Effect Treatment
Suspended solids Plugging of Injection formation Filtration
Suspended oil Plugging of Injection formation Hydrocyclones /
(particularly in presence of solids) Flotation / Filtration
Dissolved Gases Corrosion of well and facilities. Degasification
{O2 / CO2 / H2 S} Plugging of formation by corrosion products Corrosion inhibitor
Injection
Formation of Solids Equipment and formation plugging Scale inhibitor
{CaCO3 / Ba SO4 / CaSO4 / FeS} by scale Injection
Bacteria Formation plugging by bacterial Biocides
{Aerobic / Anaerobic (sulphate reducing)} residues or corrosion products
Water incompatible with formation Loss of permeability of injection formation - Pretreat formation
(clay stabilisers)
- Alter injection water chemistry
5.3 Injection Water TreatmentDump flooding is the only field proven case in which the injection water does notrequire treatment. This is because the water is never produced to the surface (Figure 9).Specific geological conditions required for dump flooding - a prolific, high pressureaquifer has to under-or-over-lay the (lower pressure) oil producing zone. Injectedwater volumes are unknown, unless a downhole flow meter is positioned between thetwo intervals. The water injection rate can be increased by positioning an electricsubmersible pump between water production and injection internals (powered dumpflooding).
Aquifer Water Supply
Prolific (High kh), High Pressure Aquifer
Casing
Seat for Plug to Isolate Oil Zone or for Electric Submersible Pump (for Powered Dump Flood)
Low PressureOil Producing Zone
Injection Interval
Table 6
Quality issues in Water
Injection
Figure 9
Dump water flooding
1
20
Research projects are being conducted on “Raw (sea) water injection”. These areexamining the conditions in which untreated (sea) water could be injected. This wouldbe a valuable technology for the development of marginal oil fields using minimumfacilities and subsea wells.
5.3.1 Example Process Schemes Employing Coarse and Fine FiltrationFigure 10 is an example of a typical process scheme - it employs coarse and finefiltration prior to de-oxygenation together with the addition of several inhibitors. Theprocess scheme operates as follows:
Chlorine Injection
Foam Inhibitor
Corrosion Inhibitor
To Water Injection
Well
Injection Pump
(Sea) WaterLift Pump
Stripping Gas InletIntake With Fish Barrier
Stripping Gas Out
Biocide
De-Oxygenating Tower
Oxygen Scavenger
Back Flush Solids
Polyelectrolyte
Scale Inhibitor
Coarse Filter Fine Filter
The (sea) water lift pumps are protected from large solids (e.g. fish) by barrier filters.Chlorine is injected into feed water in the suction of the lift pumps. It is then pumped,after the addition of polyelectrolyte coagulant and scale inhibitor, to the coarse andthen to the fine filters. The filters retain the solids while the chlorine kills the bacteriawhich could have lead to microbial growth on the filters. The solids removed fromthe water will eventually block filters - they are removed by back-flushing (reversingthe flow of process water) so that their solids removal capacity is regenerated. Thefiltered water, after the addition of a defoaming chemical if necessary, is pumped tothe top of the de-oxygenating tower, stripping gas is fed in at the bottom of the tower.A packed or trayed tower (section 5.3.4 and figure 11) is used to bring the water andgas into intimate contact to reduce the oxygen to the required level. The de-oxygenated water is drawn off at the bottom of the tower. An alternative to gasstripping is to reduce the pressure to a low value and run the tower as a vacuum dearator(figure 11). The treatments do not completely reduce the oxygen to the very low levelsrequired - this is achieved by the addition of an oxygen scavenger.
The water is now ready for pumping into the injection wells after the addition of furtherscale inhibitor and biocide to ensure the injection water sterility.
Figure 10
Injection process scheme
with fine filtration
Department of Petroleum Engineering, Heriot-Watt University 21
9Water Handling9
Gas
Gas
Gas
Bubble Caps
Vent Gas
Stripping Gas
Raw Water Inlet
Raw Water Inlet
Deoxygenated Water Deoxygenated Water
High Vacuum
Packing
5.3.2 Suspended Solids Removal(i) Sedimentation - sometimes in land operations the water source may contain ahigh concentration of coarse particles. These can be removed by allowing the waterto stay stagnant for a period of time and allowing the solids to settle. Stokes Lawdescribes the settling process; it is a function of the particle diameter and density -hence in a continuous process the required residence time to clarify the liquid at anydepth can be calculated from the time required for particles originally at the top tosettle to this depth.
Hydrocyclones can be used to speed up the removal of these solid particles. TheHydrocyclones principle of operation was described in section 9.2.5. For this duty,the centrifugal force ensures that the denser solid particles are concentrated in theunderflow while the (cleaner) water phase exits at the top.
(ii) Coagulation / Flocculation - This was described in section 9.2.3.
5.3.3 Filter Systems(i) Single and Dual Media FiltersThese filters consist of a pressure vessel filled with a filter medium. The flowdirection may be either up or down. Single media filters are filled with homogenuouslymixed sand particles or with layers made up of differently sized particles - the sandparticles graded from coarser to finer size. This arrangement by which the coarserparticles are contacted prior to the finer particles is called “deep bed” filtration -since the whole bed takes part in the filtration operation. Dual media filters, e.g.garnet / anthracite, are built up of layers of the different media arranged so that theleast dense medium also has the largest particle size.
These filters are cleaned by backwashing - removing the solids by flowing in thereverse direction at a high velocity. They can be fouled by oil since it will also be
Figure 11
Gas stripping to
deoxygenate water (left)
Vacuum dearator to
deoxygenate water (right)
1
22
retained in the filter media. Washing with detergent may help with oil removal. Fewproblems are encountered with the operation of these filters.
(ii) Diatomaceous Earth FiltersThe advantage of Diatomaceous Earth (DE) filters is that they are lighter than singleor mixed media filters. DE is nearly pure silica material originally formed by a small,single cell, marine organism. It is slurried with water and fed to the filter where itforms a porous bed on a filter support (the “precoat”). The action of this effectivefilter bed is enhanced by continually adding a small concentration of DE to the waterbeing filtered (“body feed”). This helps maintain the porosity and permeability ofthe filter cake, resulting in a longer life between removal of the filter cake bybackwashing.
The size and weight of DE filters is counter-balanced by their tendency to be operatorand maintenance intensive, their susceptibility to plugging by oil (>10 ppm) and high(>20 ppm) suspended solids concentration. Also there is a possibility of rupture ofthe filter support; resulting in injection of DE (an excellent formation pluggingmaterial) into the injection well. This is prevented by employing a “guard” cartridgefilter.
(iii) Cartridge FiltersCartridge filter units consist of one or more cartridges (similar to a sock) mountedover a perforated pipe support. The flow direction is from the outside (greatestsurface area, where the filtering action takes place) to the inside. A wide range ofparticle sizes can be removed - the cartridge construction material can be chosen sothat it can remove particles from 100 to 2 µm. The cartridges have a relatively limitedsolids removal capacity - they are most frequently used as a final (“polishing”)filtration stage, where they act as a safety check. In this duty, they would onlynormally deal with low solids concentrations e.g. intercepting solids introduced asa result of filter backwashing operations.
These units are relatively light in weight compared to the alternative filter typesdiscussed earlier. They are often used for pilot test applications and for filteringcompletion fluids during work-over operations.
5.3.4 De-OxygenationThe presence of Oxygen in concentrations greater than 5 x 10-3 g/m3 (5 ppb) in waterflood operations can cause severe corrosion and plugging of the formation bycorrosion products.
(i) Gas StrippingRemoval of oxygen by gas stripping is based on lowering of the solubility of oxygenin water by reducing the oxygen partial vapour pressure. Henry’s Law states thatgas’s solubility is proportional to the vapour pressure of the gas over water. Oxygenfrom the water may be stripped by passing a (low oxygen content) stripping gasthrough the water in co-current or counter-current flow.
Gas Stripping is normally performed in towers containing packing or perforatedtrays. The water runs into the top of the tower and the stripping gas is fed in at the
Department of Petroleum Engineering, Heriot-Watt University 23
9Water Handling9
bottom. It bubbles up through the water; the trays or packing provide good contactbetween the water and the gas (Figure 11).
The primary requirement of the gas is that it be oxygen and hydrogen sulphide free.Nitrogen natural gas or the exhaust gas from engines are commonly used. Theprinciple of removal is to reduce the concentration of oxygen in the gas coming inwith the water by dilution with the stripping gas. This reduces the partial pressureof oxygen in the gas mixture and lowers the concentration of oxygen dissolved in thewater. N.B. Where oxygenated water must be artificially lifted to charge the watersupply pumps, the use of gas lift has been found to remove a large portion of theoxygen.
The lowest residual oxygen values in the injection water require the use of oxygenfree stripping gas. Cryogenic separation of air can provide nitrogen with a very lowoxygen content - however this is often not practical offshore, particularly in floatingproduction facilities. This can be overcome by, for example, the Minox processwhich the oxygen content of the air is removed in a catalytic converter where theoxygen reacts with methanol. The nitrogen stripping gas is re-circulated. Once thestripping plant has been charged with oxygen free nitrogen, the methanol basedcatalytic oxidiser is only used to remove the oxygen stripped out from the treatedinjection water.
(ii) Vacuum De-aerationThe principle of vacuum de-aeration is to reduce the partial pressure of oxygen byboiling the water. At a temperature of 15ºC, water boils at a pressure of about 0.017atm and the residual water oxygen content is reduced to 150 ppb.
(iii) Chemical Treatment With Oxygen ScavengersOxygen removal to the required 5 ppb level is rarely possible. Oxygen scavengersare used to achieve this very low value. Oxygen scavengers remove oxygen fromwater by chemical reaction. A large number of chemical compounds can be used forthis purpose. Selection of appropriate compounds should be based on cost,compatibility of these compounds or their reaction products with other additivesused (bactericides, corrosion inhibitors, etc.) and ease of handling.
Sodium sulphite is the scavenger most frequently used in water flooding Itsscavenging effect is explained by the following chemical reaction:
2Na2SO
3 + O
2 → Na
2SO
4
Theoretically, an addition of 7.9 ppm sodium sulphite is required per 1 ppm oxygento be removed. Usually an additional 10-15 ppm is added to ensure that there is nopossibility of un-reacted oxygen being carried into a well. As the reaction proceedsslowly at room temperature a catalyst, usually 1 ppm cobalt ion, is added. Otherbivalent metals such as copper, nickel, iron and manganese ions will also catalyse thereaction. The sulphite reaction is also sensitive to pH - a minimum pH value of 6.0being required.
1
24
5.3.5 Hydrogen Sulphide RemovalAs discussed in section 1.3.3, Hydrogen sulphide is highly corrosive as well asrepresenting a safety hazard. it can be removed in a similar manner to oxygen:
(i) Mechanical removal by vacuum or counter current gas stripping,
(ii) Chemical removal by the addition of oxidising agents e.g. chlorine, potassiumpermanganate etc.
5.3.6 Chemical Treatments(i) Chlorination
Chlorination is a widely used, inexpensive, effective biocide. Chlorine hydrolyses toform hydrochloric and hypochlorous acid with water:
Cl2 + H2O = H+ + Cl' + HOCl → 2H+ + Cl' + OCl'
The degree of ionisation is dependent on the pH and the higher the pH, the lesseffective is a given quantity of chlorine (Table 7):
pH 6-8 8-9 9-10
Chlorine residual 0.2 0.4 0.8 required for 10 minute bacterial "kill” (g/m3)
The chlorine residual or ‘free’ chlorine is the total sum of Cl2, HOCI and OCI'.
Chlorine, reacts with many materials; being a strong oxidising agent. The quantityrequired is termed the ‘chlorine demand’. This includes the chlorine used up by:
(a) oxidising ferrous to ferric ion(b) reacting with H
2S to form H
2SO
4
(c) reacting with some organic corrosion inhibitors and scale inhibitors(d) reacting with sulphite oxygen scavengers.
A residual chlorine of 0.4 g/m3 is recommended to ensure bacterial control (table 7).
(ii) PolyelectrolytesPolyelectrolytes are large water soluble organic molecules made up of small buildingblocks called monomers, repeated in a long chain. They usually incorporate in theirstructures ion exchange sites which give the molecule an ionic charge. Those havinga positive charge are cationic, and those with a negative charge are anionic. Thesemolecules react with colloidal solids in the water by neutralising any electricalcharge or by bridging (tying together) individual particles to form a visible, insolubleprecipitate or floc. These chemicals therefore greatly enhance the efficiency offiltration systems.
Ferric chloride is also used as a coagulant prior to the water filtration treatment step.It's action is, however, very sensitive to pH.
Table 7
Minimum residual chlorine
content required for 10
minute bacterial "kill"
Department of Petroleum Engineering, Heriot-Watt University 25
9Water Handling9
(iii) Oxygen ScavengersOxygen scavengers were discussed in section 5.3.4.
(iv) Corrosion Inhibitors
Corrosion will occur if oxygen is present in a treatment system. Treatment options are:
(i) Remove the oxygen mechanically or chemically,(ii) Prevent oxygen ingress into the system,(iii) Coat or line the system or use corrosion resistant materials,(iv) Use an organic corrosion inhibitor which will form a protective film at themetal/water interface, thus reducing the corrosion rate.
There are a large number of chemicals available. Inhibitor selection must be tailoredto each specific system.
(v) Scale InhibitorsMost scale inhibitors in use are organic. Examples are:
(i) Organic phosphate esters. Not effective above 100ºC(ii) Organic phosphonates. Effective up to about 175ºC(iii) Organic polymers. High temperature application.
Their use was discussed in section 9.1.2
(vi) BiocidesAdditional bacterial protection can be provided by injection of Biocides. Microbi-ology is a very complex subject. The introduction of sulphate reducing bacteria intoa reservoir may cause souring (generation of H
2S) of the reservoir. The implications
of this can obviously be severe in terms of potential corrosive attack of the producingwells and facilities.
(vii) Antifoam AgentThis is occasionally used to break the foam on deaerator tower trays.
(viii) Desulphation of SeawaterEffective inhibition of barium sulphate scale may not be practical for the very highbarium content formation waters (table 6) when using sea water as the injection watersource. This can be overcome by (partial) sea water desulphation by employing thereverse osmosis process. In this process, (figure12) (Sea) Water is pumped througha semi-permeable membrane under a high pressure. The membrane is chosen so thatit restricts the flow of sulphate ions - they are retained on the high pressure side ata higher concentration than is found in sea water - while the water passed to the lowpressure side has a much reduced sulphate concentration. The water with theenhanced sulphate concentration is returned to the sea as waste.
The costs of this desulphation process has to be balanced against the costs ofdeveloping an alternative water source e.g. an aquifer.
1
26
Figure 12
Desulphation of seawater
by reverse osmosis
High Pressure
Low Pressure
Semi-Permeable Membrane
Sea WaterFeed
High Sulphate Fluid to Waste
Low Sulphate Concentration Fluid for Water Injection
Department of Petroleum Engineering, Heriot-Watt University 27
9Water Handling9
6. FURTHER READING
Arnold K. andStewart M."Surface Production Operations" Volume (2nd. Edition)Published By Gulf Publishing companyISBN 0-88415-821-7
2
Well Control 11
C O N T E N T S
1. INTRODUCTION2. MANAGEMENT OF UNCERTAINTY IN
PRODUCTION OPERATIONS2.1. Dissolved Gas Drive2.2. Gas Cap Drive2.3. Water Drive2.4. The “Real World”2.5. Overall Impact2.6. Resulting Production Constraints
3. MANAGEMENT AND OPTIMISATION OFPRODUCTION OPERATIONS3.1. Introduction3.2. Production Target Setting3.3. Data Gathering3.4. Use of Model3.5. Production Planning / Scheduling and
Forecasting4. THE ENVIRONMENTAL IMPACT OF
PRODUCTION OPERATIONS4.1. Overview of Discharges From The Oil
and Gas Industry4.2. Potential Environmental Impacts of
Production Operations4.2.1. Produced Water4.2.2. Other Aqueous Discharges4.2.3. Solids Disposal4.2.4. Atmospheric Emissions4.2.5. Accidental Spills and Flaring4.2.6. Other Forms of Environmental Impact4.3. A Waste Management Strategy4.4. Produced Water4.4.1. Regulatory Requirements4.4.2. Minimisation of Produced Water
Volumes4.4.3. Alternatives to Surface Disposal4.5. Atmospheric Emissions
2
10Oil and Gas Field Operations10
1
2
LEARNING OBJECTIVES:
Having worked through this chapter the student will be able to:
• Discuss the management of Uncertainty in Production Operations
• Relate this uncertainty management to the various drive mechanisms and the extracomplexity of the “real world”
• Identify the activities required to manage and optimise production operations
• Discuss the need for a realistic well gathering system model complemented by up-to-date, realistic field data
• Describe the use of the model for preparation of a production forecast and theidentification of profitable workover opportunities
• Identify the main elements of the environmental impact of production operations
• Discuss the elements of a waste management strategy
• Propose options to reduce the marine discharge of produced water
• Discuss the importance of atmospheric emissions
2
Department of Petroleum Engineering, Heriot-Watt University 3
210Oil and Gas Field Operations10
1. INTRODUCTION
This chapter deals with three aspects of oil and gas field operations:
(i) Management of the uncertainty involved in facility design and operations,
(ii) Management and optimisation of production,
(iii) The environmental impact of production operations.
2. MANAGEMENT OF UNCERTAINTY IN PRODUCTION OPERATIONS
It must be recognised that designing and operating an oil or gas field is very differentfrom designing and operating a refinery or chemical plant. Refineries and chemicalplants are much more complex, but the composition of the inlet flows are well definedat the design stage. Also the quantities and qualities of products remain unchangedduring the lifetime of the project. In contrast, oil and gas production involves simplerprocessing; but has to deal with the large uncertainty represented by our lack of a fulldescription of the initial state of the reservoir as well as our inability to predict howit will behave once reservoir fluids are being produced. Alteration of the productionfacilities is frequently very expensive due to their remote locations e.g. offshore. Infact, addition of extra equipment may be impractical due to space or weightlimitations. Cost effective management of this factor is becoming more important inthe current, stringent economic climate of low oil prices. “Slim” platform construc-tion techniques mounting minimum facilities are being inevitably favoured in the raceto profitably develop marginal prospects.
Typical examples of uncertainty that need to be taken into account within the FieldDevelopment Plan and the facility design include:
(i) Wellhead conditions are subject to uncertainty until the formation is actuallydrilled e.g. if the fluid in an isolated fault block was found to be gas rather than oilbearing, the required wellhead completion could change from the planned 5000 psioil producer to a 10,000 psi gas producer. This has a large impact on the specificationfor wellhead equipment, surface flow lines etc.
(ii) Field (potential) plateau production and reserves may be (much) larger or a(small) fraction of the facility design values. It is normally a very profitable projectto de-bottleneck the oil production facilities to take advantage of any greater-than-expected well / reservoir production potential.
(iii) The oil and gas production rates will decrease and the water production rateincrease as the field matures.
The rate at which these fluids change, and the long term production performance willdepend on the reservoir production mechanism e.g. the drive mechanism which forcesoil from a distant location in the reservoir into the well, and subsequently to thesurface. Typical examples are discussed below. Note that a homogeneous reservoiris assumed.
1
4
2.1. Dissolved Gas Drive
Time
Gas
- O
il R
atio
Time
Res
ervo
ir P
ress
ure
Pressure depletes to bubble point and gascap formed N.B. Assumes Constant
Production Rate
Depletion Bubble Point
Gas Expansion
Time
Tubi
ng H
ead
Pre
ssur
e
Bubble Point
Well Dies
Constant GORReducing BHP
Rate of reduction in BHP slowsGOR reduces
OIL
Gas Bubbles Will
Eventually Form Gas Cap
During the initial, depletion stage of reservoir production, the producing gas-oil ratiowill remain constant so long as the field produces above the bubble point; thereservoir pressure decreasing as fluid is removed from the reservoir. The Gas-OilRatio (GOR) will then decrease once the bubble point pressure has been passed - thegas will have started to separate from the oil prior to it reaching the production well.It will eventually form a gas cap whose expansion will “cushion” or reduce the rateof reservoir pressure decrease due to further fluid production. The producing TubingHead Pressure (THP) will reduce steadily until the bubble pressure is reached in thereservoir. The THP will then reduce more rapidly as the GOR decreases since thiswill increase the hydrostatic head of the fluid column within the well itself.
Figure 1
Production mechanism and
development of GOR and
Reservoir Pressure with
time for a dissolved gas
drive mechanism
2
Department of Petroleum Engineering, Heriot-Watt University 5
210Oil and Gas Field Operations10
10.2.2. Gas Cap Drive
Time
Gas
- O
il R
atio
(G
OR
)
Time
Res
ervo
ir P
ress
ure
GORdecreases furtheras gas comes outof soloution
Gas cap reachestop perforation
Gas Expansion "cushions"pressure depletion
N.B. Assumes Constant Production Rate
Time
Tubi
ng H
ead
Pre
ssur
e
Gas Breakthrough
Well Dies
Reducing GORLimited changeto BHP
Increased tubing frictionfrom extra gas overridingthe reduction in hydrostatichead
GOR increasesReduced averagefluid density
OIL
Gas Cap
This is similar to the latter stages of the dissolved gas drive. The decrease in reservoirpressure due to fluid production from the reservoir is “cushioned” by expansion ofthe gas cap. The gas-oil ratio continuously reduces as the reservoir pressuredecreases due to well production. This extra, liberated gas will further decrease therate in reservoir pressure reduction, but the (surface) gas-oil ratio of the producedfluid will decrease, resulting in greater pressure drops across the production tubing.Eventually, the gas will expand to such an extent that the gas-oil contact will reachthe top perforation, and a dramatic increase in gas-oil ratio will occur.
The THP will show a continuous decrease, driven by the reducing GOR and reservoirpressure. It will initially increase on gas cap breakthrough, due to the increased GORreducing the average hydrostatic head. However, it will then decrease, as the everincreasing gas rate leads to a more rapidly increasing frictional pressure drop in thetubing than the corresponding reduction in the fluid column’s hydrostatic head.
Figure 2
Production mechanism and
development of GOR and
Reservoir Pressure with
time for a gas cap drive
mechanism
1
6
2.3. Water Drive
time
Res
ervo
ir P
ress
ure
N.B. Assumes Constant Production Rate
Aquifer response begins
strong waterdrive
water breakthrough atbottom perforationweak waterdrive
time
Wat
er C
ut
Strong Waterdrive
Weak Waterdrive
Time
Tubi
ng H
ead
Pre
ssur
e
Well Dies
Water Breakthrough
OIL
With this production mechanism, the rate of the reservoir pressure reduction due towell production is reduced due to expansion of a (large) aquifer that is in goodpressure communication with the oil reservoir. This expansion can only begin tohappen once the reservoir pressure has reduced, e.g. there is a lag between initiationof production and the pressure support from the aquifer becoming apparent. Thisaquifer support manifests itself as a reduction in the rate of pressure drop withcontinuing well production. Strong aquifer support (large aquifer with goodconnections to the oil reservoir) will lead to limited further pressure decline, whilea weak aquifer will have proportionally less effect. The rising oil/water contact willeventually reach the bottom perforation and the water cut will rapidly increase. Thetubing head pressure shows a similar story - an initial drop prior to initiation ofaquifer support, followed by a period when the rate of pressure reduction depends onthe strength of the water drive, i.e. the THP is controlled by the changes in thereservoir pressure. Once water break-through occurs, the fluid column density (andhydrostatic head) increases rapidly, and the well ceases production shortly thereafter.
These (simplified) examples illustrate how the reservoir drive mechanism affects thereservoir wellhead and pressures, and the composition of the gas/oil mixture whichis to be processed in the surface facilities. It further defines the need for artificial lift,e.g. if gas lift is to be employed, then gas compression facilities will be requiredunless a separate source of high pressure gas is available. The use of electricsubmersible pumps will increase electrical generation requirements.
Figure 3
Production mechanism and
development of Water Cut
and Reservior Pressure
with time for a water drive
mechanism
2
Department of Petroleum Engineering, Heriot-Watt University 7
210Oil and Gas Field Operations10
2.4. The “Real World”This is much more complex than the simplified examples discussed above.
(i) The reservoir will be heterogeneous - leading to rapid reserve from highpermeability streaks, followed by early water or gas break-through.
(ii) The reservoir may be compartmentalised by sealing faults which reduce thereserves connected to the well. This would lead to an (unexpected) rapid drop inbottom hole pressure. Possible early installation of artificial lift or the need to side-track the well to a new location will then be required.
(iii) A strong natural water drive may have been assumed. Provision has to be madee.g. space for water injection plant and spare conductors in the case of an offshoreplatform so that water injection can be implemented before the project economics areput in jeopardy.
These factors all point to the need to develop a systematic approach to the collectionand analysis of production data, so that changes in the pattern of well production canbe recognised, and remedial measures initiated. This is discussed in greater detail insection 3. The carrying out of special tests, e.g. well interference and build up tests,cased hole production logs etc. in addition to the routine work, will add to theunderstanding of the reservoir performance.
2.5. Overall ImpactThe combination of the above - and many more - factors result in large changes in therate at which oil, gas and water are produced during the project’s lifetime. A typicalexample of these changes during the early, mid-life and mature phases of an oil andgas field’s production history is summarised in tables 1 & 2 respectively:
Fluid Phase
Early Mid-Life Mature Abandonment
Oil (bopd) 100k 100k 30k 10k
Gas (MM sft3/d) 100 100 30 10
Water (bwpd) <500 30k 100K >100k
Water cut <0.5% 30% 77% >90%
Typical changes in Production Rates during the Lifetime of a 100,000 bopd oil field
Table 1
Typical changes in
Production Rates during
the Lifetime of a
100,000 bopd oil field
1
8
Fluid Phase
Early Mid-Life Mature Abandonment
Gas (MM sft3/d) 500 500 50 20
Condensate (bpd) 5000 5000 500 200
Water (bwpd) 250 5000 500 200
Typical Changes in Produced Gas Volumes From a ‘Typical’ Southern North Sea Gas Field
2.6. Resulting Production ConstraintsThe production phases are described in table 4. The production philosophy for oilfields is often to build up to plateau production as quickly as possible after "first oil"and to produce at this maximum (plateau) rate for as long as possible, giving theproduction and operational cost profile illustrated in figure 4. Wells are often pre-drilled prior to installation of the platform to maximise the early production. Workoverof existing wells sidetracking of wells from depleted zones and infill drilling are allused to ensure that the decline rate is minimised. In the final phase of field life theoperating costs are lowered as much as possible to delay field abandonment. Thisoccurs when the net revenue is less than the operating expenses. Thus operational costreduction is the main driver in late field life.
Oil
Pro
duct
ion
Rat
eC
ost
Plateau
Platform&Facilities
Bui
ld-U
p
Decline
First Oil
DiscoveryWell
AppraisalWell
Abandonment
Economic Limit
Pipeline
Drilling
Operations
Time
Figure 4
An oil production scenario
Table 2
Typical Changes in
Produced Gas Volumes
From a ‘Typical’ Southern
North Sea Gas Field
2
Department of Petroleum Engineering, Heriot-Watt University 9
210Oil and Gas Field Operations10
Discovery/ • Exploration successful. Minimise time to "First Oil"
Appraisal • Drill appraisal wells to provide further information as required
First Oil • Time between discovery and first oil should be minimised.
• Fast track development reduces (discounted) exploration costs.
Build-Up • Objective is to reach plateau production as quickly as possible.
• Well production potential limits (number of wells) production
• Build up phase can be reduced by:
- Pre-drilling wells,
- Using more than one drilling rig,
- Drill most prolific well first.
Plateau • Production reaches maximum and remains constant
• One factor in the well/facility/pipeline system will be constraining production
• Excess well deliverability often available in early plateau period
• Consider de-bottlenecking facility or increasing pipeline capacity e.g. by use
of drag reducer to raise plateau production rate
• Drilling extra wells extends plateau period
• Well draw down increased to maintain oil production as water cut increases in
later plateau phase
• Eventually oil production will be constrained by, for example, the facility water
handling capacity e.g. Maintain plateau by shutting in high water cut production wells
Decline • Maximise oil production within production constraints
• Identify candidate wells for work-over/re-completion/stimulation/artificial lift to
produce remaining reserves
• Examine operating cost structure e.g. alliances, outsourcing etc. to maintain
income above expenses for as long as possible. This maximises reserves.
Abandonment • Plug wells and remove platform / facilities / pipelines etc.
Production Constraint Phase
3. MANAGEMENT AND OPTIMISATION OF PRODUCTIONOPERATIONS
3.1. IntroductionProduction Management is based on controlling all aspects (technical and organisa-tional) of the production process with each well in a field being produced to plan. Thisplan’s objective is to maximise return on investment by maximising oil productionwithin constraints. Constraints can occur at the reservoir, well, gathering network,process and marketing levels. Incentive schemes will frequently reward the fieldmanager based on his ability to achieve the economic optimum production thatmaximises the field profitability within this constraint framework.
Table 4
The Production Phases
1
10
Reservoir and well conditions change throughout the life of a field. Productionmanagement is a continuous process of surveillance and reaction to these changes.Systematic comparison of forecast and actual production will allow predictions to beimproved, and production maximised. Production Optimisation aims to maximisereturn of investment by:
(i) Maximising Economic Production Rates - the impact of temporary operationaldifficulties can be minimised by re-routing wells, adjustment of producing condi-tions etc.
(ii) Minimising Production Downtime - this requires an effective surveillanceprogram - to rapidly identify well problems and programme a repair by providing anoperating baseline against which actual measured rates can be compared.
(iii) Identifying Production Restrictions - locate those factors limiting production.e.g. compressor capacity, pipeline restrictions, process limits, well productivity etc.The economic benefit resulting from capital expenditure to remove the restrictioncan then be assessed.
(iv) Maintenance Planning - Field lift gas allocation, production rate setting etc.need to be reset in order to maximise production when equipment has to be taken outof service for maintenance.
(v) Planning for the future - It is common practice to operate the platform and/orpipeline as an independent profit centre. Production forecasts can be made bysimulating expected future production conditions. The optimum time for theinstallation of the need for artificial lift, additional compression capacity, drilling ofnew wells or development of near field potential etc. can then be identified. Theavailability of a realistic, up-to-date, production optimisation model allows prepa-ration of ‘what-if’ contingency scenarios with minimal delay.
3.2. Production Target SettingThe conventional practice is to base future production forecasts on an extrapolationof past history. This produces an achievable forecast based on the current operatingpractices. It does not necessarily take into account the field’s (untapped) productionpotential. This can be done with a (calibrated) production system mode. Thisestimates the field’s production with optimum operating practices (lift gas allocation,well choke settings, separator pressures etc.). The (current) operational efficiency canthen be estimated by comparing how close the actual “oil in the tank” is to thisproduction potential.
Operation at maximum well potential may not be practical for a number of reasons e.g.process limitations, sand production, contract or quota restrictions, gathering systemor pipeline limitations etc. Proper production management calls for the effect of eachrestriction on the overall production rates to be quantified. One can then check whetherretention of this restriction is (still) economically justified - as well as to identify thecost (capital expenditure) of its removal.
The production management system consists of two parts:
2
Department of Petroleum Engineering, Heriot-Watt University 11
210Oil and Gas Field Operations10
(i) TechnicalThis consists of four component parts;
(a) a reservoir fluid description (matched PVT model)
(b) a systems analysis models and software used to match the production performance of each well
(c) a gathering system model (pipeline geometry, length, elevation, diameter,roughness) used to perform production optimisation calculations
(d) a set of system constraints at a well, gathering system and process level. Typicalwell constraints are water coning, sand production, reservoir depletion. Gatheringsystem constraints will include maximum velocities (corrosion/erosioneffects), maximum allowable export rate and pressure etc
(ii) Field DataAccurate and up-to-date data from the field are required to provide realistic
predictions from the technical model. Well performance is continuouslychanging as the reserves are recovered, reservoir pressure depletes and thewell's produced fluid ratio changes. This is shown in figure 5
Wellbore Flowing Pressure
Absolute Open Flow
P RRese
rvoir pressu
re
IPR-1IPR-2IPR-3
IPR-4
IPR
-5
IPR
-6
ProducingRate
Time
IPR @ time-0
3.3. Data GatheringThe implementation of modern measurement and control systems at the wellheadallow the well's production to be controlled, while the well production performanceand that of the facilities are continuously measured and recorded. The situation isillustrated in figure 6, which illustrates the process for a remote oil field operated withflowing wells in addition to many types of artificial lift (Beam pumping / gas lift /electric submersible pump). The production performance of the well and theefficiency of the artificial lift process (if any) are measured and transmitted; alongwith data from the gathering station / well test facility, to the (central) Field OperationsOffice (FOO). Monitoring of the (automated) Distributed Control System (DCS)which controls the field’s operation also takes place at the FOO. Two-way commu-nication has been installed, so personnel at the field operations office can intervenewhen required and adjust producing conditions. Digital microwave communicationsbetween the FOO and the headquarters office ensures that the data can be shared withall the petroleum engineers monitoring the field.
Figure 5
Deterioration of IPR with
reservoir recovery
1
12
Remote Surveillance Data Transfer
Field OverviewWell Test
Individual Well Optimisation
Key: RTU - Remote Transmitter UnitDSC - Distributed Control System
Subsciber Radio
Main office
Field Operations Office
DCS
UHF Radio
Solar Power RTU
RTU
RPC
Rod PumpESP Well
Gas Lift well
Digital Microwave
RTU
The problem presented by a remote land location is similar to that found offshore. Thisis illustrated by the development where a minimum facilities platform acts as the FOOfor a number of platform wells shown in figure 7 (A field) as well as associated subseadevelopments (B, C and D Fields). Aberdeen, a distance of some 125km, is theheadquarters office.
A Field
Key SSV - Subsea Safety Valve
B Field
C Field
D Field
Minimum Process FacilitiesMinimum Manning
Platform
TenderAssistedDrilling
SSSVDiverter
Satellite FieldsProducing to
Mother Platform
Early productionby Pre-drillingSatellite Wells
Underwater Completionsfor Satellite Fields
Pre-installedFlowline Bundles
PipelinesBundlesUmbilicals
KEY
To Shore Existing oil export line
Spur to tie into existing
gas pipeline to shore
Figure 6
An integrated well
management system for a
land field
Figure 7
A subsea development
2
Department of Petroleum Engineering, Heriot-Watt University 13
210Oil and Gas Field Operations10
For both the land and offshore cases, a production systems model is required to turnthe data recorded by the various measuring instruments into “information” which canbe used to control and optimise the field’s operation. The procedures used to acquirethis data, carry out quality control, and display the way it is presented into well testand down-hole pressure survey data. This is used to maintain the model by carryinga systematic review of modelled and measured performances. Success depends on theavailability of good quality data and trained personnel to analyse it.
The key to the process is the provision of quality data. Operational personnel are mostmotivated to obtain this once they can see a use for the data - this is achieved in theset up described above since the system model is made available to them to help themwith their duties in the FOO.
3.4. Use of ModelOnce the field model has been set up and calibrated, the next step is to run anoptimisation to determine whether additional production can be obtained by re-allocating lift gas or adjusting well chokes. Constraints must be carefully set tocorrectly model system limitations. Obvious constraints are compressor and separa-tion capacity. Well performance constraints may also be required to deal with welland pipeline stability issues. For example, minimum lift gas allocation rates can beused to ensure that the operator does not attempt to run wells at an unstable rate.
Reservoir level constraints also need to be included. e.g. if gas flaring is a productionconstraint, then high GOR producers should be beaned back or even shut in. Ifreservoir considerations require that this well continue producing at a certain mini-mum rate, it can be kept on production. However, this will incur a production penalty,as lower GOR producers must be then choked back to enable the higher priority wellsto produce.
These calculations can be done automatically by setting up and calibrating the modelto the well test rates. An automatic, optimisation algorithm can be used by setting theseparator pressures and lift gas availability and starting the calculations. Results mustbe carefully checked to determine if the resulting recommended well rates areacceptable for the prevailing reservoir and production policy.
As mentioned above, each active constraint in the system will result in lessproduction. By comparing optimisation run results, the cost to production (and cashflow, profitability etc.) of reservoir policies can be quantified.
Optimisation of a large field is a complex task. All aspects of building, calibrating andmaintaining the model must be properly organised if accurate results are to beobtained and manpower requirements minimised. A first step is to systematicallyorganise the PVT and well completion data. This ensures that the current wellconfiguration information can be readily incorporated into the well models. Aneffective well file system ensures accuracy and minimises the time required to builda well model.
A large volume of routine production test data is generated during field operation.Each individual well test carried out with the test seporator provides valuableinformation on reservoir and well performance for use in history matching purposes.
1
14
Such well tests are used in production allocation (allocation of the produced oil to thevariuous wells being produced); hence they directly affect reservoir material balancecalculations.
Each test must be carefully screened before being used for production allocation. Suchsystems analysis models are a convenient way to identify poor tests and identifychanges in reservoir or well conditions by comparing the well model and the actualtest rates.
Flowing bottom-hole pressure surveys enable well models to be calibrated. Thequality of the field data can be evaluated by comparison with the well model results.The reason for any significant difference should be investigated and the well modelupdated, if appropriate. This ensures that the well model is always current, so wellproblems can be quickly identified.
Jan Feb Mar Apr May
platform shutdown
sidetrack
new well
tie in satelite field
compression upgrade
Platform Activities
Indi
vidu
al W
ell P
erfo
rman
ce
Time
Water Cut
Production
GOR
Process and Pipeline Model
Production Forecast
Individual Well Performance
Flow Constraints +Equipment Uptime
Field Production (or Cash) Forecast
Pro
duct
ion
Time
Figure 8
Development of a
production forecast
2
Department of Petroleum Engineering, Heriot-Watt University 15
210Oil and Gas Field Operations10
3.5. Production Planning / Scheduling and ForecastingA Field production forecast is normally prepared every month. The productionconditions can be re-optimised as part of the process or earlier if significant changesin well or facility operation occur. It is essential that the field optimisation model hasbeen updated to represent the current well status. This will ensure that the results arevalid and represent current, actual field performance. The interplay of the variousparameters is shown in figure 8. A field-wide production allocation should be doneto check for discrepancies in measurement and allocation. Differences should beidentified and eliminated as necessary.
With the model properly calibrated, an optimisation is run to determine the productiontargets, operating pressures and lift gas allocation for the next month. These resultsform the basis of the field operating plan. Additional optimisation runs can be madeto provide contingency operating conditions for planned maintenance or equipmentbreakdowns (e.g. compressor trips). The production plan can also include a prioritylist of wells to adjust in the case of e.g. lift gas shortage, water disposal limitations,gas export limitations etc. The production plan also includes reservoir pressure andinjection targets as appropriate.
The final part of the planning process is a summary of well repair and work-overopportunities. Problem wells are identified as part of the model calibration process.Likely repair costs and post-work-over production rates need to be prepared in orderto provide data for allocating capital and scheduling work-overs. Analysis of thecompeting production improvement opportunities is the basis of the medium termwell operations workover plan.
Pipelines and platforms are frequently treated as independent profit centres. Sparecapacity can be used to exploit company owned near field potential {extended reachwells, satellite, subsea completion’s employing natural flow, down-hole artificial liftor subsea (surface) pressure boosting}. Alternatively, this capacity may be used forthird party business - some platforms in the North Sea meet more than 50% of theiroperating costs by processing / pumping production from other companies. Similary,the Shell/BP central North Sea ETAP development was only economically possiblebecause both companies developed a cluster of fields as a single project. Here, a BPplatform based development (Marnoch) provides host to a series of Shell subsea fields(Heron, Skua etc). Neither company could have developed their own fields as a "standalone" project.
4. THE ENVIRONMENTAL IMPACT OF PRODUCTIONOPERATIONS
Production Operations take place within a regulatory framework set by internationalagreements or by the government of the nation where the oil or gas field is situated.In addition to meeting these legal requirements, the production operation has to beconducted in such a manner that they meet the expectations of the public at large, asinfluenced by publicity appearing in the media and by overview by non-governmentalenvironmental organisations such as “Green Peace”, “Friends of the Earth” etc. Thiscombination of legal and social issues is summed up by the need for the operator toconduct his operations in such a manner that he earns a “licence to operate”.
1
16
This chapter summarises the various routine discharges to the environment that occurduring routine operations and unplanned emissions (oil spills etc.).
4.1. Overview of Discharges From The Oil and Gas IndustryThe North Sea is relatively enclosed and bordered by highly populated and industr-ialised nations. In addition to effluent discharged or dumped directly into the sea;several major rivers such as the Rhine, Thames, Elbe etc. flow into the sea carryinga high contaminant load - making it one of the worst polluted seas.
Relatively speaking, the pollution from the North Sea oil and gas installations aremodest. Hydrocarbon discharges provide the greatest input, accounting by weight forover 97% of material discharged into the marine environment (i.e. excluding atmos-pheric emissions). Hydrocarbons from the North Sea oil and gas installations accountfor about 18% of the total, or 24% if inputs from shore based installations are included(figure 9).
Inshore Oil (6*109.kg)Shipping (7*109.kg)
Coastal Discharges (10*109.kg)Waste/Dredge Dump (13*109.kg)
Atmospheric (13*109.kg)Offshore Oil (18*109.kg)
River/Run-off/Sewage (50*109.kg)
Natural Seepage (260*109.kg)
Whilst the oil and gas industry makes a significant contribution to the total hydrocar-bon input into the North Sea, the same is not true for other contaminants. Emissionsto the atmosphere are much greater - figures 10 & 11 were taken from a majoroperators 1996 Annual Health Safety and Environmental Report on their Explorationand Production Operations. These large values contrast with the some 3 *109kg tonnesof oil that were discharged to the marine environment with produced water.
Figure 9
Estimated inputs of
Hydrocarbons to the North
Sea.
2
Department of Petroleum Engineering, Heriot-Watt University 17
210Oil and Gas Field Operations10
1992
1993
1994
1995
1996
16.014.012.010.0
8.06.04.02.00.0
Hyd
roca
rbon
s fla
red
(109
kg/g
r)
1992
1993
1994
1995
1996
0
50.0
100.0
150.0
Car
bon
Dio
xide
em
issi
ons
(109 kg
/gr)
4.2. Potential Environmental Impacts of Production OperationsThe sources of potential Environmental Impact of Production Operations are summa-rised below:
Oil Spillage Cooling Water Production Water Noisecontains oil
Cuttings Discharge Thermal Discharge Pipeline Discharges Sewage
Atmospheric Physical Ballast/Crude Washing / cleaning / Emissions Disturbances storage displacement & main drainage (flares/vents/turbine water water exhausts etc.)
The relative proportion that each of the sources contributes to oil discharge to themarine environment is summarised in figure 12. Some discharges e.g. pipelinedischarges when very large volumes of treated water are discharged over a shortperiod of times e.g. after pressure testing the pipeline during initial commisioning.Historically, by far the greatest input of hydrocarbons comes from the discharge ofdrilling cutting; while the impact of oil in produced water is now dominant.
Figure 11
Total Carbon Dioxide
Emissions by a major
operators E and P
operations.
Figure 10
Mass hydrocarbons flared
during a major operators E
and P operations.
Table 5
Sources of potential
environmental impact from
Exploration and
Production operations
1
18
Cuttings Produced Water Spills
94.3% 76.3%
4.2%
39.6% 86.9%
1.5%
3.2%20.5%
1.8%58.6%
13.1%
1982 1987
1992 1997
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
40
35
30
25
20
15
10
5
0
Oil
cont
ent (
ppm
)
Figure 12(a)
Changes in relative
quantities of discharges to
the sea from offshore oil
operations 1982-1997
Figure 12(b)
Long term trend for oil
content in discharged
producted water for the UK
continental shelf.
2
Department of Petroleum Engineering, Heriot-Watt University 19
210Oil and Gas Field Operations10
4.2.1. Produced WaterProduced water consists of a mixture of formation water (found with the oil) andbreakthrough injection sea water (used to push the oil through the rock strata). Thewater discharged contains oil, the statutory limit being set at a monthly average of 40mg/l, with no sample having a value greater than 100 mg/l. During 1998, the conceptof a lower (currently voluntary) company wide allowable average discharge limit of30 mg/l was introduced within U.K. waters; rather than the above (higher) individualfield value. Different limits apply elsewhere e.g. 32 ppm maximum average oilconcentration in the Gulf of Mexico (USA).
As a well ages, there is greater (injection) water breakthrough resulting in greaterquantities of produced water being discharged. This trend, combined with theincrease in number of wells drilled, has led to a substantial rise in the quantity ofhydrocarbons released into the North Sea from produced water {Figure 12(a)}. Thequantity of oil discharged with production water is now dominant especially since thedischarge of oil containing cuttings has now been banned. The discharge of drillcuttings from water based muds is still allowed; as is the discharge of cuttings frommuds based on organic esters, ethers etc. if suitable permits are obtained. These specialchemicals are allowed as a replacement for oil in some circumstances. Cuttingsdischarge from these latter mud types have been recently disallowed in the North Seabut are still allowed in other offshore drilling areas.
Large effects are being made to reduce the oil content in the discharged producedwater. Figure 12(b) records the progress made in the United Kingdom ContinentalShelf). Produced water re-injection, where the produced water, is substituted forspecially prepared injection water is also being pursued as a means to reduce the longterm environmental impact of production operations.
Production chemicals are also used on a large scale - e.g. more than 25,000 tonnes inthe UK sector during 1988. It is unknown how much of this is eventually dischargedwith the production water since the (relative) oil/water/gas phase solubility of all thechemical and their subsequent reaction products is unknown - but it is suspected that20-30% is discharged via the produced water. The quantity of production chemicalsreaching the environment will be reduced as re-injection of produced water into aformation requiring water injection or into a dedicated injection formation becomesmore common.
4.2.2. Other Aqueous DischargesProduced water is not the only source of aqueous discharges. This can come frommany sources - process cooling water, excess treated (sea) water for injection, waterused for pipeline testing, drainage water etc. Some of these sources have a (relatively)small volume while others e.g. water used for pipeline testing, (treated with at leasta bactericide and corrosion inhibitor) may have a large volume but only occur onceor twice in the life of the project. Others, such as cooling water (treated by filtrationand chlorination coupled with a discharge temperature 25° - 30° C above seatemperature) have large continuous flows (up to 500,000 bwpd or 3,300 m3 / hr).
Other well treatment fluids are emitted as temporary, one-off events e.g. acidic brinesproduced after a well stimulation with acid, may also be discharged with the producedwater.
1
20
4.2.3. Solids DisposalTwo organic solid types - wax and ashphaltenes - may separate from the crude oilduring the production process. The point at which they form - in the well or thefacilities - and the quantities of solid, will depend on the properties of the crude oil andthe production conditions. They have to be removed in an appropriate manner alongwith any separated inorganic solids e.g. clay fines, formation sand grains etc. Theamount deposited would build up in the separator if it were not regularly removed.This would reduce residence time and oil/water separation efficiency, as well as blockflow lines.
Formation sand particles and "fines" are also produced from the wells along with theoil and gas. As discussed in the Sand Control module , these are collected in theseparators et. and may, after suitable washing procedures to remove any adhering oil,be discharged. Alternatively, they may be transported to a licensed disposal site.
A special class of solids are inorganic scales e.g. CaCO3, BaSO
4, formed by pressure
/ temperature changes during the production process or the mixing of incompatiblewaters. Carbonate scales are normally removed from tubing by acid but sulphatescales requre removal by drilling them out. Care should be taken if dry drilling of scalefrom tubing recovered from the well is practised. Barium and strontium scales oftencontain contaminants giving them a low level of radioactivity (which can be a healthhazard if the dust, from the drilling out of dry samples of these scales, is deposited inthe lungs).
4.2.4. Atmospheric EmissionsProduction stations and (offshore) platforms are often sited in remote locations.However, the environmental concerns (acid rain, ozone, greenhouse gasses etc.) aresimilar to other industrial activities. Significant amounts of electrical power (50 - 100kW) may be generated for use at the production location if electrical supply from aNational Grid is not feasible or economic. Gas turbines are frequently employed.Often, heat recovery is not employed since fuel gas is available at relatively low cost.
Large quantities of volatile organic compounds are emitted to the atmosphere duringExploration and Production Operations. (See figure 13 for figures from Shell’sExploration and Production Operations). This emission of Volatile Organic Com-pounds (VOC) is due to deliberate venting (65%), un-burnt gas during flaring (25%)and fugitives from tanker loading operations, valves, tank roof vents etc. (10%). Theyare equivalent to:
1992 1993 1994 1995 1996
30002500200015001000
5000
VO
C e
mis
sion
(10
9 kg/g
r)
Figure 13
Volatile Organic
Compound emitted by a
major E&P operator
2
Department of Petroleum Engineering, Heriot-Watt University 21
210Oil and Gas Field Operations10
(i) 20% of the hydrocarbon mass flared or
(ii) 1000 times (three orders of magnitude) greater than that of the volume of oildischarged with produced water or
(iii) 200 times the volume of oil lost due to oil spills.
Apart from carbon dioxide, flare gasses may contain nitrogen oxides, carbon monox-ide, sulphur dioxide, particulate carbon as well as un-burnt hydrocarbons.
The large scale hydrocarbon losses e.g. venting, tanker loading operations etc. areeasiest to identify and can be controlled with appropriate technology. The largernumbers of small scale emissions e.g. diffusion / leakage through valves, gaskets,flexible hoses and connections, isolation of process equipment during maintenanceetc. are more difficult to manage. The following statistics put these emissions incontext - The offshore oil industry contributes less than 4% of the UK greenhouse gasemissions, while for Norway the corresponding figure is 20%.
4.2.5. Accidental Spills and FlaringOil spills tend to have a high public profile. This is because they are immediatelyvisible and the damage they cause to wildlife such as birds is obvious and distressingto witness. The amount of oil that enters the environment through accidental spillageis small (figure 12). Oil also enters the sea from flaring due to incomplete hydrocar-bons combustion. Studies have shown that as much as 30% of the hydrocarbons arebeing flared remain un-burnt. A new generation of flare tips (burners) with designsthat give more complete combustion have been developed to address this problem.
4.2.6. Other Forms of Environmental ImpactThe environment around offshore developments may also be disturbed by noise andvibration and by thermal pollution. Thermal pollution results from the discharge ofcooling water and by local warming around risers and oil storage cells. The effect isnot necessarily deleterious, the localised warming encourages fish, although it mayalso enhance growth of fouling organisms.
Physical disturbance of the seabed by oil and gas explorations is largely confined tothe area immediately around fixed installations and pipelines. Installations and theirexclusion zones cover around 0.1% of the area of the North Sea, accounting for twicethe area closed off by ship wrecks. By far the greatest disturbance of the seabed iscaused by fishing activity (54%).
4.3. A Waste Management StrategyA Waste Management strategy in any business area, not just Exploration andProduction, will address:
(i) REDUCTION generate less waste through more efficient practices
(ii) REUSE reuse material in original form
1
22
(iii) RECYCLE convert waste back to usable material
(iv) RECOVER extract material or energy from waste for other uses
(v) DISPOSAL dispose of final residue in most environmentallybenign manner possible.
Many companies are developing strategies that reduce the environmental load belowthe legal requirement and are developing “zero discharge options”. In every case,when developing a Waste Management Strategy, it is important to look at the totalimpact of the complete system rather than concentrating on a particular pollutant e.g.is it always sensible to gas strip hydrocarbons out of produced water if the net effectis that the pollution load becomes air - rather than water-borne and CO
2 emissions are
increased?
4.4. Produced Water
4.4.1. Regulatory RequirementsDry oil (no water production) is produced when an oil field is initially brought intoproduction. However, the level of water production rapidly increases as the field ages;with many wells only being abandoned when the water cut reaches 95% or evenhigher. In 1995 the Major Oil Companies typically produced as much water as oil intheir world-wide operations. Their water production is expected to increase rapidly;a doubling within 5 years not being unusual in their corporate projections. This waterproduction entails a high economic penalty - it comes with a typical lifting, treatingand disposal cost of some 25-75 US cents/bbl water.
The requirement for produced water discharges to have a (monthly average) maxi-mum of 40 ppm oil was set by the Paris Commission (PARCOM) in 1978 and hasn’tbeen significantly revised since then. The analytical technique specified essentiallymeasures the aliphatic hydrocarbons, ensuring that dispersed rather than total oil ismeasured. This standard is not based on an underlying environmental consideration;but is an equipment based standard chosen on the basis of “what is achievable” with“best available technology” driven by a belief that “less oil is better”.
The regulatory value of 40 ppm, which has been in existence for nearly 20 years, hasbeen discussed at regular intervals during this period. A general reduction to say, 30ppm, would increase the fraction (maybe 25-30%) of the North Sea fields that alreadyhave difficulty in reaching the regulation criteria. A voluntary, company wide targetfor an average discharge level of 30 ppm oil has now been accepted for UK continentalshelf waters.
Although the quantities of oil discharged into the surface water environment aresignificant, they are a small proportion of the total oil produced, e.g. a figure of 10goil discharged per tonne of product for Shell’s Exploration and Production Opera-tions. For comparison, these 3,000 tonnes are 25% of the amount of oil accidentallylost by Shell during oil spills.
In Shell’s case, more than 50% of the produced water is re-injected into undergroundreservoirs while the average concentration of oil in discharged production water was28 ppm.
2
Department of Petroleum Engineering, Heriot-Watt University 23
210Oil and Gas Field Operations10
4.4.2. Minimisation of Produced Water VolumesIt must be remembered that large volumes of water, several times that of the oil,frequently have to be produced during a fields lifetime due to the physics of theproduction process associated with flow in porous media. Advanced Well Design,such as horizontal wells, minimise the draw-down and hence extend the productionperiod prior to water breakthrough. Such well designs ensure that the volume of“extraneous” water due to coning, high permeability (or thief) zones showing earlywater breakthrough etc. are minimised.
���������
yyyyyyyyy
���������
yyyyyyyyy
Water Producing Zone
Water Producing Zone
Water Producing
Zone
Polymer Blocking Gel
e.g. Maraseal TM
Pumped IntoFormation
CementPlug
Bridge Plug
CementSqueezed
IntoPerforations
ScabLinerStopsWaterInflow
Improved techniques for selective Down-hole Water Shut-Off of water producingzones minimises this “extraneous” water production. Not only does this reduce thevolumes of water (and its entrained oil) discharged, but also reduces the (artificial)lifting costs, potentially increases the well’s (oil) production capacity and extends thewell’s producing life above the “economic limit”. Figure 14 illustrates mechanicalshut-off options (e.g. scab liners, bridge plugs etc.) which place a mechanicalrestriction across the perforation of the wellbore to stop the water flow have beenemployed for many years. Their effective application requires identification of thewater producing zone. Similar knowledge is required when placing a cement plug inthe well bore to stop water production . An alternative is to perform a chemicalpolymer (gel) shut off operations where the gel can be pumped into the formation.Both of these can be carried out, rapidly, (relatively) cheaply and easily via coiledtubing. These latter, chemical options are being intensively researched to developimproved, more environmentally friendly, materials and better placement techniques.
Figure 14
Water shut off options
1
24
Development of reliable procedures to selectively reduce water production viabullhead techniques using the so-called “relative permeability modifiers” would findwide application.
Equipment for Down-hole Separation and Injection has been developed (figure 15).The concept is based on hydrocyclone technology for the separation of “free water”followed by down-hole injection into a disposal zone, typically with the aid of anelectric submersible pump. The oil and remaining water is lifted to surface forconventional separation and treatment. Apart from reducing lift energy and waterseparation and disposal costs; Down-hole Separation and Injection provides newproduction options such as dynamic control of the oil-water contact. Practicalproblems such as monitoring, control and maintenance need to be resolved prior towidespread application.
A final option is Wellhead (rather than down-hole) Separation and Injection at asubsea production manifold followed by re-injection.
± 975 m KB
± 10 m
± 25 m
Hydro-Cyclone
Shale Streak
Packers
Submersible Motor
Upstream Full Flow Pump
Concentrated "Emulsion" Pump
Production toSurface Reduced Watercut
(Factor 10 Less Water)
Produc ing Interval
Figure 15
Downhole separation and
injection: Oil/water
2
Department of Petroleum Engineering, Heriot-Watt University 25
210Oil and Gas Field Operations10
4.4.3. Alternatives to Surface DisposalAlthough water treatment techniques have improved significantly, there are noavailable methods for the complete removal of oil. Furthermore, with increasingconcern over the environmental effects of the dissolved, chemical components ofproduced water; operators are being challenged to develop alternatives by their ownmanagement (drivers: company reputation, “license to operate”), as well as by theregulators,.
Produced Water Re-Injection has been trialed in several fields with variable success.It has usually proved to be more difficult (and expensive) to retain high, long term wellinjectivity with produced water than with sea or aquifer water. Well injectivity iseffected by produced water ‘quality’ (composition, oil and solid content, tempera-ture), formation permeability, fracture gradient and down-hole scaling potential.Further, corrosion and souring problems have been observed. However, it cannormally be successfully introduced once sufficient experience has been gained andthe influence of the above factors controlling the process has been understood.
Separation andWater Re-injection
Install in 2004
Field BStart up 2004
Field AStart up 1998
10 km
Water Injection
Production
5-50 km
WaterB
Oil BOil A
1998 2004 2010 2016
WaterAG
ross
Flu
ids
Field A and B
1998 1999 2000 2001 2002 2003 2004
500m -1500mWater Depth
Seabed separation and reinjection is being developed as a means of increasingcapacity of Floating Production, Storage and Offtake (FPSO) vessels by reducing thevolume of produced water to be processed. This is illustrated in figure 16 whereproduction of a satellite field is accelerated through use of the FPSO facilities once theproduction from the main (field A) field has peaked. The technology requiredseparation , wellheads and power distribution and pumping and control are all fieldproven but require "repackaging" for seabed operation.
Produced Water should ideally be used to replace “fresh” injection water duringmature water floods e.g. BP’s Ula field currently meets 95% of its water injection
Figure 16
Development of a satellite
field using subsea
separation and injection
1
26
requirements with Produced Water. However if it is being injected into a disposalzone, the re-injection could result in an increase of energy consumption equivalent to250 - 500 tonnes of carbon dioxide emissions for every tonne of oil prevented frombeing discharged into the sea. The total impact on the environment must be consideredwhen considering such options.
It should be noted that surface discharge of produced water is not an option for onshoreoperation where contamination of surface water by discharged oil is not acceptable.Further, selection of disposal zones must also ensure that (potentially) potableaquifers are not contaminated.
4.5. Atmospheric EmissionsApart from CO
2 (the major component by weight - see Figure 11) this covers Volatile
Organic Compounds (figure 11) that are emitted to the atmosphere due to deliberateventing, un-burnt gas during flaring and fugitives from tanker loading operations,leaking valves and pump seals, tank roof vents etc. Reduction in hydrocarbonemissions via reduced flaring is now receiving the full management attention that itdeserves. The scale of the problem - and the opportunity with respect to improvingrecovery - was summarised in Figure 10.
Being able to usefully use the gas co-produced with the oil depends on the availabilityof a suitable infrastructure (transport pipeline and end user) to allow the economicevacuation of the gas. Gas fields remote from existing infrastructure, or thosecontaining only small quantities of gas, make the finding of economic solution for theso called "stranded gas" problem particularly challenging. This is particularly truenow that large scale continuous flaring of excess gas is now seen as unacceptable byboth the regulatory authorities and many company managements. Solutions employedto date have included:
(i) injecting the gas back into the reservoir to increase liquid recovery (either byinitiating a gas drive or condensate recycling mechanism);
(ii) temporarily storing the gas in an underground formation while awaiting thecommissioning of the sales pipeline;
(iii) developing a market for the gas e.g. Liquefied Natural Gas scheme, Synfuel/methanol synthesis plants, power generation (though offshore location of thepower plant apart from local platform requirements has not proved economicto date);
(iv) use of the gas for artificial lift (gas lift);
(v) disposal of the gas in a suitable reservoir to avoid the “carbon tax” penaltylevied in some countries.
Other innovative solutions will no doubt be developed in the coming years sinceseveral research projects to solve the “stranded gas” problem are now in progress.
2
Department of Petroleum Engineering, Heriot-Watt University 27
210Oil and Gas Field Operations10
OIL AND GAS FIELD OPERATIONS TUTORIAL 1
List 3 key uncertainties faced by Production Technology/Operations/Engineering.
SOLUTION
(a) Reserves.Surface facilities design, in terms of handling the production rates, will dependlargely on the size of reserves associated with that particular field. Real field’s(potential) plateau production and reserves may be much larger or smaller than thefacility design values.
(b) Reservoir Description.Wellhead conditions are subject to uncertainty until the formation is actually drilled.Reservoir pressure conditions and fluid type characteristics will affect, for example,the design and specifications for wellhead equipment and surface flow lines.
(c) Reservoir Drive Mechanism.The production rates of the individual fluids will change and then decline as the fieldmatures. The rate at which these fluids change, and the long term productionperformance will depend on the reservoir production mechanism i.e. the drivemechanism which forces oil from a distant location in the reservoir into the well andsubsequently to the surface.
OIL AND GAS FIELD OPERATIONS TUTORIAL 2
Explain the statement “Production Planning and Optimisation occurs at different timescales”.
SOLUTION
Reservoir and well conditions change throughout the life of a field. Productionmanagement is a continuous process of surveillance and reaction to these changes.The processes designed to meet the objectives of the Production optimisation modeland the contingency operations are executed as follows:
(a) “Maximise Economic Production Rate” => Immediate.The impact of temporary operational difficulties can be minimised by immediateaction (re-routing wells, adjustment of producing conditions, etc).
(b) “Minimise Production Downtime” => Hours.This requires an effective surveillance program to rapidly identify and repair wellproblems.
(c) “Identify Production Restrictions” => Short/medium term.Identify those factors limiting production (e.g. compressor capacity, pipeline restric-tions, well productivity, etc) and perform an economical analysis to asses remedialactions.
1
28
(d) “Maintenance Planning” => Short/medium term.Operating conditions such as field lift gas allocation, production rate setting, etc.need to be reset in order to maximise production when equipment has to be taken outof service for maintenance.
(e) “Planning for the future” => Long term.Production forecasts can be made by simulating expected future production conditions.The optimum time for the installation/implementation of artificial lift, additionalcompression capacity, drilling of new wells or development of near field potential,etc, can then be identified.
The Haggis Field Exercise
For Distance Learning
Combined Question & Blank Answer Form1
VERSION 2.2
1 The % mark noted at the end of each question refer to its relative weighting with respect to the complete
exercise Figures, which are referred to in the text, are placed at the end of the exercise.
DECLARATION
I …………………………………confirm that this work submitted for assessment is my own and expressed in my own words. Any uses made within it of the works of other authors in any form (e.g. ideas, equations, figures, text, tables, programs) are properly acknowledged at the point of their use. (A list of the references employed should be included.)
Signed : ………………………………………….. Date : ………………
Name in Capitals……………………………………………………..
-----------------------------------------------------------------------------------------------------------------
Guidance Notes for Students
Heriot-Watt University Ordinance 9, para 2.4 (c) states: A student may be deemed to be in breach of discipline if he should: During an examination or other test copy from or communicate with another person or be found in possession of books or any printed or written papers or any other material containing information relevant to the subject of the examination other than those allowed in the examination or use any other unfair means.
20.1.3 The words highlighted in this extract refer to “other test” and “other
unfair means”. These guidance notes are intended to draw the attention of students to the importance of avoiding inadvertent use of what might be regarded as unfair means.
“Other test” is construed in the widest possible sense of any assessed work that contributes to the award of a degree or other qualification of the University.
“Unfair means” is construed in the widest possible sense of any practice that is intended to gain a dishonest advantage.
20.1.4 The particular unfair means noted in this paper are copying, plagiarism
and collusion, although this is by no means an exhaustive list. 20.2 COPYING 20.2.1 Copying the work of others, including that of other students in the class
or group, is an indication of unfair means, whereby one person gains credit for work undertaken by another.
The Haggis Field
20.2.3 Recommendation to students Make clear in your submission any permitted
reproduction that you have carried out. Check the rest of the work to ensure that it is your own. Working with other students in informal study groups is a desirable part of the academic experience, but be sure that the work that you eventually submit is yours and not that of others. Keep copies of material such as working notes, or sketches of diagrams, or drafts of essays that show that the work is your own effort.
20.3 PLAGIARISM
20.3.1 Plagiarism is the act of stealing from the writings or ideas of another. This is regarded as a very serious offence in academic works. It is generally accepted that such “stealing” occurs where there is no acknowledgement that the writings or ideas belong to someone else. Most academic scholarship involves building on the work of others, while acknowledging their contribution. There are accepted conventions for making that acknowledgement, although the conventions may vary marginally from one subject to another.
20.3.3 Recommendation to students When you undertake an assessed work which involves drawing on the writings or ideas of others, make sure that you acknowledge each contribution. Use a style of acknowledgement that is good practice in the academic discipline. If you are not sure what is good practice, read the guidance provided by your School or seek advice from academic staff. When the work is completed, check carefully that you have not overlooked acknowledgement of any source used.
20.4 COLLUSION
20.4.1 Collusion involves a secret agreement to deceive. This means that more than one person is involved in the deception. Where an accusation of collusion is added to an accusation of, for example, copying, there must be clear evidence of the involvement of each party. One person may copy the work of another without the knowledge or approval of that person.
20.4.3 Recommendation to students Be careful about lending your completed work to
other persons. You may think that you are helping them to meet a deadline, but it may result in problems for you if they copy your work without telling you. What started out as a friendly action might risk an accusation of collusion.
The Haggis Field
The Haggis Field
The Haggis field was discovered in June 1994 in the UK sector of the North Sea in a water depth of 300 ft. The field was developed using 5 wells and reached peak production in 1996. Since then, oil production has decreased rapidly due to an increase in water production. Reservoir Properties: The Haggis sand was deposited in a turbidite environment. It is quite homogeneous with an average porosity and permeability of 22% and 200 mD, respectively. The reservoir sand, however, is isotropic with a Kv/Kh ratio of 0.1. The top of the sand was encountered at 6400 ft TVDSS, and the oil-water contact is at 6500 ft TVDSS. The reservoir is normally pressured with an initial reservoir pressure of 3300 psia and little or no aquifer support. Reservoir pressure has declined with production to 2800 psia at present. Pressure maintenance was not considered when the field was being developed. For information and for those wishing to run this exercise using a completion design software package, Table 1 lists the PVT data for the Haggis fluids at current reservoir conditions.
Reservoir Temp. 150° F Oil API Gravity 40° API
Gas Relative Density 0.80 G.O.R. 550 scf/STB
Pb 2030 psia Bo 1.27
Oil Viscosity 0.66 cp Bg 0.0046
Gas Viscosity 0.022 cp Bw 1.023
Gas Z-Factor 0.73 Water Salinity 200000 ppm
Water Viscosity 0.67 cp Table 1: Haggis PVT Data
Haggis wells: The Haggis field wells have an economical limit of 1500 STB Oil/d/well; i.e. producing at rates lower than that is not feasible. Haggis-3 was drilled in May 1995. It is taken to be the case study for this field as it has average parameters for Haggis wells. Figure 1 is the completion diagram for Haggis-3. Above the wellhead, the well was completed with the same 5 1/2" OD production tubing encased in a mud line (no insulation). The mud line connects the wellhead (on the sea-bed) to the Xmas tree on the platform. For information, results of a recent pressure survey from Haggis-3 are listed in Table 2.
Depth 650 1605 2590 3600 4590 5587 6490 ft TVD
Pressure 525 735 990 1292 1629 1920 2266 psia Table 2: Haggis-3 Pressure Survey
The Haggis Field
Haggis-3's well parameters, and results from both well testing and production logging are summarised in Table 3. Table 3 also contains data on Haggis-1. Haggis-3 and Haggis-1 are essentially twin wells, except that completion damage has resulted in Haggis-1 being considered the worst well in the field.
Haggis-3 Haggis-1 Oil Production Rate 4730 3930 STB/d Water Cut 30 28 % WH Flowing Temperature 65 61 °F Pressure at Xmas tree 445 psia Skin (Well Test)
2.92 7.17
P.I. (J) (Well Test) 12.36 9.11 STB/d/psi Damaged Zone Relative Permeability 50 25
%
Damage Zone Thickness 12 in
Crushed Zone Skin 0.100 Drainage Radius 4000 ft
Table 3: Well Data
The Scenario: The rate of oil production decline in the Haggis field is alarming, and if no action is taken, Haggis will become uneconomical by the end of this year. The Operator of the field, Big Kahuna Oil Inc., does not accept this situation and has fired the field's former team leader for improper management of the field. Big Kahuna has hired you to improve production from the Haggis field. Your Mission: Big Kahuna Oil Inc. has asked you to study the field’s potential. A model has been created using the company approved software (EPS's FloSystem). A variety of production proposals have been modelled and the outcome of these simulations has been provided in graphical form. It is your job to evaluate these proposals using the all the data provided. The report should outline: A) the model used in the study, B) the potential of the base case scenario, C) your assessment of production enhancement proposals from the Haggis engineers and D) your recommendation for a project which will enhance production from Haggis.
Note: Wellflo graphs show gross production rates e.g. total production rates. Net production rates are required for the well analysis.
The Haggis Field
A) Develop a Well Model for Haggis-3:
a) Using Haggis-3 as your case study, complete the missing data in Figure 2.
b) The well Haggis-3 is used as the base case well for the Haggis field throughout this exercise. To minimise the computer time involved in simulations, the model contains only those components that contribute significantly to the pressure drop along Haggis-3. These are shown on Table 4.
Node No.
Component Name Measured Depth (ft)
1 Outlet node / Xmas tree 0 2 Riser 350 3 Wellhead 350 4 5.5” Tubing 850 5 S.C.S.S.S.V 850 6 5.5” Tubing 4000 7 5” Tubing 5600 8 7” Liner 6530.5
Table 3: Well Data
Compare this model to the completion design in Figure 1. When would it be necessary to include the nipples and tubing constrictions in the model design? __________________________________________________________________________________________________________________________________________________________
[2.5%]
c) The Outflow curve for Haggis-3 was modelled using a number of well-known correlations. The
sensitivity analysis is shown in Figure 3. The correlation that is most appropriate is: __________________________________________________________________________
[2%] There are two main reasons why this correlation was chosen: 1) __________________________________________________________________________
__________________________________________________________________________
2) ____________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
[2.5%]
d) The Big Kahuna Oil Inc. is unfamiliar with the concept of nodal analysis used in this flow simulation. You are asked to explain the process and the conditions for flow stability.
The inflow-outflow diagram below is for Haggis-3 at 30% water cut and utilising an appropriate flow correlation.
The Haggis Field
Curve A
Curve B
Point D
Point B
Point C
Point APoint E
Diagram 1: Haggis-3 base case Curve A represents the _________________________________________________. Curve B represents the _________________________________________________. The reservoir pressure is _____________psia (Point______).
The_________________ IPR model was used to produce this curve. Why was this chosen?
___________________________________________________________________________________________________________________________________.
Point ______ represents the operating point. The operating point is [stable/unstable]. Explain:
_____________________________________________________________________________________________________________________________________________________________________________________________________.
Nodal analysis may be carried out at any point in the producing system. In this report, the analysis was carried out to find the operating point at the sandface (see above diagram). Other typical examples of nodes selected during completion design are: 1)__________________________________________________________to evaluate _________________________________________________________________________________
2)__________________________________________________________to evaluate _________________________________________________________________________________
[11.5%]
The Haggis Field
B) Base Case Analysis: As a good manager, the first thing you have to do is evaluate the potential of what you have at the moment. To achieve that, you must determine what effect the decline in reservoir pressure and the increase in water cut will have on Haggis-3's production if nothing is done to improve its production. In other words, determine the reservoir pressure and the water cut at which Haggis-3 will becomes uneconomical to produce under the current production scenario.
a) In order to evaluate the viability of producing Haggis-3 as water-cut increases and reservoir
pressures reduces, a number of sensitivity analyses were performed on the production from Haggis-3 using WellFlo. These sensitivities are illustrated by Figures 4a, b, c and d, and their results should be summarised in Table 5:
PRes.
WC 2800 2700 2600 2500 psia
30% 4770 35% 40% 45%
Table 5: Haggis-3 Production Forecast [4%]
b) You discuss these figures with the field’s engineers. Since artificial lift can not be supported by
the production facilities on the Haggis platform, you agree with your engineers that you have to start a water injection scheme to maintain the reservoir pressure at 2800 psia. Figure 5 shows the well sensitivity to water cut. Under these circumstances, Haggis-3 will produce economically (1500+ BOPD) at maximum water cut of ____________, and at an oil production rate of ____________ BOPD. This is considered to be the Base Case scenario, against which all other schemes in section C will be compared.
Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
Base Case
c) With a water injection scheme in place, you expect to face even more severe water production
from Haggis-3. One way of dealing with such a problem is to plug-off “watered-out” perforations. List two advantages and two disadvantages of this scheme.
Advantages: 1)___________________________________________________________________________________________________________________________. 2)___________________________________________________________________________________________________________________________.
Disadvantages: 1)___________________________________________________________________________________________________________________________. 2)___________________________________________________________________________________________________________________________.
[5%]
The Haggis Field
d) A sensitivity study is outlined in Figures 6a, b and c for three plugging-off policies. Based on the advantages and disadvantages you identified above, the most advantageous plugging off policy is:____________________________________________________. The minimum open interval at which Haggis-3 will produce economically under this policy is ________________________ft.
[2.5%]
C) Further Projects: You phone management and convince them to fund the pressure maintenance scheme. In addition, they agree to provide a budget for an extra project to improve Haggis’s production. In this section you must select the best project from your engineers’ suggestions below. Since production, and thus revenue, from Haggis is greatly affected by water production form the field, Big Kahuna Oil Inc. defines the best production enhancement project for the Haggis field as the one that sustains economical production form Haggis-3 at the highest water-cut. Therefore, this should be the criterion you use to select the project you recommend to management.
C.1) Production Technology Solutions:
You walk into the team’s senior production technologist office, and give him the good news from management. He congratulates you, and immediately suggests that the project should be a production technology project. He goes on to explain that as the Haggis wells are extremely damaged, acidising all the wells make the field very profitable.
a) Determine the benefit from acidising Haggis-3 in terms of the maximum water cut at which the
acidised Haggis-3 will sustain economic production, if acidising restores the original rock permeability. Figure 7 shows the sensitivity of Haggis-3 to water cut after acidising. The maximum water cut at which the well can produce economically is ________%.
Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
Acidising
(Use this result to compare the benefits from this project to the other proposed projects, and as a basis for your recommendations in section D).
b) A production model of well Haggis-1 is already available and shows that Haggis-1 will produce
economically at maximum water cuts of 50 % after acidising. What are the implications of this if it is decided to carry out a campaign in which all the Haggis wells are to be acidised? __________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
[3.5%]
The Haggis Field
c) The diagram below shows the inflow-outflow curves for Haggis-3 at the maximum economic water cut after acidising. Assuming that acidising succeeds in restoring the original rock permeability in both wells, sketch on the diagram the inflow-outflow curves for Haggis-1.
Diagram 2: Haggis-3 after acidising
[2%] C.2) Suggestions from the Drilling Engineer:
The drilling engineer walks into your office and says that he heard that you were going around the different departments asking for ideas on how to improve the field's production. He says that he could have saved you all the trouble because he has the perfect solution. "Side-track the well", he says. He goes on to explain that, he would have gone for a horizontal well. However, since Big Kahuna have recently had bad luck with drilling horizontal sections longer than 400 ft, a 75° deviated well through the reservoir is probably the better option from a drilling point of view. a) Given the drilling department’s recent experience with horizontal wells, should you side-track
Haggis-3 to a 400 ft horizontal well running through the middle of the reservoir or propose a 75° deviated well penetrating the whole of the reservoir height?
The production prognosis from Haggis-3 sidetrack as a 400ft horizontal well is shown in Figure 8 while a 75° deviated well is illustrated in Figure 9. Summarise the results in the table below:
Option Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
C2.a1 400ft Horizontal C2.a2 75° Deviated
Option C2.a2 produces the better results because: ____________________________________________________________________________________________________________________________________________________________
[3%]
The Haggis Field
b) Given the engineering and economic factors and assumptions below, design the optimum horizontal well (to the nearest 500-ft) for the Haggis-3 side-track: Length of build-up section is: 2000 ft-MD Pay-back Time: 6 months Side-track cost: US$850 /ft-MD Price of Oil: US$15/bbl Cost of Processing and Shipping: US$8 / bbl Assumptions: 1) Economics based on un-dicounted cash flows
2) Production rate is constant for the first 6 months
3) Use Profit to Investment Ratio as measure of value, see below.
(Profit to Investment (PI) Ratio = TCS/MCO where TCS is Terminal Cash Surplus and MCO is Maximum Capital Outlay. The PI ratio is a measure of the cash surplus or profit generated for every unit of currency invested.)
Table 6 below can be completed using the results from Figure 10 and taking in account the above engineering and economic factors for drilling horizontal wells.
Horizontal length (ft) 500 1000 1500 2000 2500 3000 3500 4000
Parameterand unit Formula
Table 6:Economic analysis of sidetracking Haggis-3.
(In the space above, enter the parameter, formulas used and intermediate working) Table 6 indicates that the optimum horizontal section length for Haggis-3 is _____ ft. However, by carefully analysing the economic outcome and technical aspects of the design, what other conclusion could be drawn? _______________________________________________________________________________________________________________________________________________________________________________________________________________________________________
The Haggis Field
c) Determine the benefit from side-tracking Haggis-3 in terms of the maximum water cut you can economically produce the well with if it was side-tracked to the optimal horizontal length determined above.
If Haggis-3 is sidetracked to the optimal horizontal length, then the maximum water-cut at which the well will produce economically is ____%. (Figure 11)
Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
Horizontal Well
(Use this result to compare the benefits from this project to the other proposed projects, and as a basis for your recommendations in section D).
[15%]
C.3) Another Production Technology Suggestion: Artificial Lift
After lunch, Haggis’s senior production technologist steps into your office. He says that while he was having lunch, he remembered that Big Kahuna has been quite successful using Electric Submersible Pumps (ESP’s) in the nearby Tatties field. Additionally, he says that gas lift may be an option as the well has a high PI and the Big Kahuna has sufficient gas supply from another nearby field. He says that he is aware of the drilling department suggesting a slanted well, but installing artificial lift will be quicker and cheaper than side tracking the well. He suggests investigating the installation of an artificial lift scheme with a target liquid rate of 9000 STB/d.
C.3.1: Electrical Submersible Pumps
a) The Big Kahuna has had success with the following Centrilift pumps in the North Sea.
Pump Motor Cable Size GC 8200 562 Series #1 HC 7000 562 Series #1 HC 9000 562 Series #2 KC 12000 562 Series #2
Given the current conditions choose the optimum pump for Haggis-3 given the details and
engineering assumptions below,
Pump Types: Centrilift Setting Depth: 5000ft Minimum Equipment OD: 5” Maximum Equipment OD: 6.8” Platform Electricity Supply Frequency: 60Hz
Assumptions: 1) Production tubing is unaltered. 2) No wear on Pump or motor i.e. wear
factor is 1. 3) Efficiency of the gas separator is 100%
i.e. separator efficiency is 1. 4) Viscosity and gassiness corrections are
used.
Figures 12a & b show the performance plots of the pumps at current conditions. The most suitable pump of Haggis-3 given the present conditions is ______________.
The Haggis Field
This pump is the optimum choice because ________________________________________________________________________________________________________________________________________________________________________________________________________________________________.
[3%]
b) As water production is the limiting factor, the production technologist suggests halting water injection and allowing the reservoir pressure to drop. Determine which pump would be suitable in these conditions. Figures 13 a-d and Table C.3 are designed to assist you with that determination.
PRes.
Pump 2800 2600 2400 2200 Psia
GC 8200 HC 7000 HC 9000
KC 12000 Table C.3: Haggis-3 Production Forecast with ESP installed,
** denotes rate outwith the operating range of the pump.
According to Table C.3 the optimum ESP for Haggis-3 for declining reservoir pressure is ________________, at a water cut of 30%.
This pump is the optimum choice because ________________________________________________________________________________________________________________________________________________________________________________________________________________________________.
[5%] c) Determine the benefit from installing an ESP in Haggis-3 using Figure 14 in terms of the
maximum water-cut at which the optimised pump will sustain economic production prior to suspending water injection (i.e. no depletion, reservoir pressure 2800psia).
Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
Optimised ESP
(Use this result to compare the benefits from this project to the other proposed projects, and as a basis for your recommendations in section D). [1.5%]
The Haggis Field
C.3.2: Gas Lift Design a) A gas lift design for Haggis-3 is undertaken based on current conditions and engineering
assumptions/details outlined below,
Max casing head pressure: 1200 psi Gas available for injection: 20MM scf/d Injected gas gravity: 0.6 Setting depth: 3900 ft Valve differential pressure: 100 psia Minimum spacing: 450 ft “Kill” brine density: 0.465 psi / ft Minimum safety margin (see A on diagram) 50 psi
Assumptions: 1) Production tubing is unaltered. 2) Unload the tubing full of static fluid
against the well head pressure (i.e. static fluid to 0ft MD).
3) No transfer margin is required. The gas lift design is shown below (Diagram 3). Note briefly on the diagram the roles of the different valves in the design. The upper valves should be OPEN/CLOSED when assessing gas lift capabilities during field life.
A
Diagram 3: Gas lift design for Haggis-3 [3%]
The Haggis Field
b) Determine the optimum injection rate as the reservoir pressure declines from Figure 15 and summarise the results in Table C.4.
PRes.
2800 2600 2400 2200 2000 Psia
Optimum injection rate
MMscf/day
Table C.4: Optimum gas injection rate for Haggis-3. The required gas injection rate for the design production rate is __________________________.
The criteria used to choose the optimum injection rate are _________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________.
Higher injection rates do not improve production as the reservoir declines because ___________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
[8%]
c) The gas lift scheme is redesigned using the optimum gas injection rate for a reservoir pressure of 2800 psia. Determine the benefit from installing Gas Lift in Haggis-3 in terms of the maximum water-cut at which the optimised injection rate will sustain economic production using Figure 16.
Scenario Maximum Economic
Water Cut Production Rate @
30% Water Cut Optimised Gas Lift
(Use this result to compare the benefits from this project to the other proposed projects, and
as a basis for your recommendations in section D). [2%]
C.4) Suggestions from the Facilities Engineer:
The gentleman behind you introduces himself as the Haggis field's facilities engineer when you are walking to your car after a long day at the office. He shares with you a very interesting discovery he made today. He explains that if you agree to provide a dedicated line from the Haggis wells to the low-pressure separator, you can lower the average Xmas tree pressure to only 100 psia. He points out that suggestions from the other department have large uncertainties associated with them since they deal with the subsurface. His suggestion is simple, neat and will solve the field's problems.
The Haggis Field
a) Consider the base case inflow-outflow diagram for Haggis-3 below. Sketch on the diagram how the base case curves would alter if the well head pressure were decreased.
Diagram 4: Haggis-3 base case
[2%]
b) Evaluate the benefits of lowering the Xmas tree pressure, in terms of the maximum water-cut you can economically produce Haggis-3 with after the Xmas tree pressure is lowered to 100 psia.
If the production facilities on the Haggis platform can be modified to allow the Xmas tree pressure of Haggis-3 to be lowered to 100 Psia, then the maximum water-cut at which the well will produce economically will become _____%. (See Figure 17)
Scenario Maximum Economic
Water Cut Production Rate @ 30%
Water Cut Lowering Xmas Tree Press
(Use this result to compare the benefits from this project to the other proposed projects, and as a basis for your recommendations in section D).
c) List 2 advantages and 2 disadvantages of this scheme.
Advantages 1)__________________________________________________________________________________________________________________________________________________________ 2)__________________________________________________________________________________________________________________________________________________________ and disadvantages 3)__________________________________________________________________________________________________________________________________________________________ 4)_________________________________________________________________________________________________________________________________________________________
[5%]
The Haggis Field
D) Recommendations to Management:
a) Assess the production enhancement projects proposed by the Haggis field engineers in section C above and compare them to one another and to the base case scenario. Bear in mind that Big Kahuna Inc. has set the ranking criteria for these projects to be the maximum water-cut at which Haggis-3 can sustain economic production (i.e. > 1500 STB oil/d).
Table 8 below summarises the results from the various simulations carried out on the Haggis-3 well.
Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
Base Case
Acidising
75° Deviated Well
Optimum Horizontal Well
Optimum ESP
Optimum Gas Lift
Lowering Xmas Tree Pressure to 100 psia
Table 8: Haggis-3 Production Forecast [2.5%] b) Recommend to management a plan of action which either recommends maintaining the base case
scenario or executes one of the proposed projects.
Based on the WellFlo simulations and my assessment of them, I recommend that Big Kahuna Inc. invest in a water injection scheme to maintain Haggis’s reservoir pressure at 2800 psia. In addition, I recommend that Big Kahuna adopts [the _______________________ project /none of the projects investigated above] because: _____________________________________________________________________________________________________________________________________________________________.
[5%] 3) A number of risks have been overlooked in this assessment since the maximum water cut at which
the wells will flow at an economic rate has been used as the ranking criteria for the above projects. These risks add to the uncertainty of achieving the results on which your recommendation was based. As the Haggis field team leader it is your duty to report and account for these risks to management.
Complete Table 9 below which should identify three major risks that have been overlooked by this assessment. Briefly explain how each one could add to the uncertainty of the assessment and prescribe steps that need to be taken to account for their effects.
The Haggis Field
Risk/Uncertainty How this adds uncertainty to
above assessment? Steps that can be taken to account for / minimise this uncertainty
1)
2)
3)
Table 9: Haggis-3 Production Forecast [9%]
The Haggis Field
Figures:
The Haggis Field
The Haggis Field
The Haggis Field
Figure 3: Sensitivity to flow correlation.
The Haggis Field
Figure 4a: Layer pressure =2800 psia, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 4b: Layer pressure = 2700 psia, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 4c: Layer pressure = 2600 psia, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 4d: Layer pressure = 2500 psia, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 5: Layer pressure = 2800 psia, maximum economic water cut.
Operating point region
The Haggis Field
Plugging off policies Figure 6a: Plug off at 30% water cut, sensitivity to open interval.
Operating point region
The Haggis Field
Figure 6b: Plug off at 40% water cut, sensitivity to open interval.
Operating point region
The Haggis Field
Figure 6c: Plug off at 48% water cut, sensitivity to open interval.
Operating point region
The Haggis Field
Figure 7: Haggis-3 after acidising (e.g. damage zone permeability 200mD),
sensitivity to water cut.
Operating point region
The Haggis Field
Figure 8: 400ft horizontal well, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 9: 75° deviated well, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 10: Horizontal well, sensitivity to effective length.
Operating point region
The Haggis Field
Figure 11: Optimum horizontal well, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 12a: Performance curves for ESP GC8200 and HC7000
Figure 12b: Performance curves for ESP HC9000 and KC12000
The Haggis Field
Figure 13a: GC 8200, sensitivity to reservoir pressure
Operating point region
The Haggis Field
Figure 13b: HC 7000, sensitivity to reservoir pressure.
Operating point region
The Haggis Field
Figure 13c: HC 9000, sensitivity to reservoir pressure.
Operating point region
The Haggis Field
Figure 13d: KC 12000, sensitivity to reservoir pressure.
Operating point region
The Haggis Field
Figure 14: Optimum ESP, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 15: Performance analysis of the gas lift design, sensitivity to reservoir pressure and gas injection rate.
The Haggis Field
Figure 16: Optimum gas lift design, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 17: 100psia well head pressure, sensitivity to water cut.
Operating point region
The Haggis Field Exercise
For Distance Learning
Combined Question & Blank Answer Form1
VERSION 2.2
1 The % mark noted at the end of each question refer to its relative weighting with respect to the complete
exercise Figures, which are referred to in the text, are placed at the end of the exercise.
DECLARATION
I …………………………………confirm that this work submitted for assessment is my own and expressed in my own words. Any uses made within it of the works of other authors in any form (e.g. ideas, equations, figures, text, tables, programs) are properly acknowledged at the point of their use. (A list of the references employed should be included.)
Signed : ………………………………………….. Date : ………………
Name in Capitals……………………………………………………..
-----------------------------------------------------------------------------------------------------------------
Guidance Notes for Students
Heriot-Watt University Ordinance 9, para 2.4 (c) states: A student may be deemed to be in breach of discipline if he should: During an examination or other test copy from or communicate with another person or be found in possession of books or any printed or written papers or any other material containing information relevant to the subject of the examination other than those allowed in the examination or use any other unfair means.
20.1.3 The words highlighted in this extract refer to “other test” and “other
unfair means”. These guidance notes are intended to draw the attention of students to the importance of avoiding inadvertent use of what might be regarded as unfair means.
“Other test” is construed in the widest possible sense of any assessed work that contributes to the award of a degree or other qualification of the University.
“Unfair means” is construed in the widest possible sense of any practice that is intended to gain a dishonest advantage.
20.1.4 The particular unfair means noted in this paper are copying, plagiarism
and collusion, although this is by no means an exhaustive list. 20.2 COPYING 20.2.1 Copying the work of others, including that of other students in the class
or group, is an indication of unfair means, whereby one person gains credit for work undertaken by another.
The Haggis Field
20.2.3 Recommendation to students Make clear in your submission any permitted
reproduction that you have carried out. Check the rest of the work to ensure that it is your own. Working with other students in informal study groups is a desirable part of the academic experience, but be sure that the work that you eventually submit is yours and not that of others. Keep copies of material such as working notes, or sketches of diagrams, or drafts of essays that show that the work is your own effort.
20.3 PLAGIARISM
20.3.1 Plagiarism is the act of stealing from the writings or ideas of another. This is regarded as a very serious offence in academic works. It is generally accepted that such “stealing” occurs where there is no acknowledgement that the writings or ideas belong to someone else. Most academic scholarship involves building on the work of others, while acknowledging their contribution. There are accepted conventions for making that acknowledgement, although the conventions may vary marginally from one subject to another.
20.3.3 Recommendation to students When you undertake an assessed work which involves drawing on the writings or ideas of others, make sure that you acknowledge each contribution. Use a style of acknowledgement that is good practice in the academic discipline. If you are not sure what is good practice, read the guidance provided by your School or seek advice from academic staff. When the work is completed, check carefully that you have not overlooked acknowledgement of any source used.
20.4 COLLUSION
20.4.1 Collusion involves a secret agreement to deceive. This means that more than one person is involved in the deception. Where an accusation of collusion is added to an accusation of, for example, copying, there must be clear evidence of the involvement of each party. One person may copy the work of another without the knowledge or approval of that person.
20.4.3 Recommendation to students Be careful about lending your completed work to
other persons. You may think that you are helping them to meet a deadline, but it may result in problems for you if they copy your work without telling you. What started out as a friendly action might risk an accusation of collusion.
The Haggis Field
The Haggis Field
The Haggis field was discovered in June 1994 in the UK sector of the North Sea in a water depth of 300 ft. The field was developed using 5 wells and reached peak production in 1996. Since then, oil production has decreased rapidly due to an increase in water production. Reservoir Properties: The Haggis sand was deposited in a turbidite environment. It is quite homogeneous with an average porosity and permeability of 22% and 200 mD, respectively. The reservoir sand, however, is isotropic with a Kv/Kh ratio of 0.1. The top of the sand was encountered at 6400 ft TVDSS, and the oil-water contact is at 6500 ft TVDSS. The reservoir is normally pressured with an initial reservoir pressure of 3300 psia and little or no aquifer support. Reservoir pressure has declined with production to 2800 psia at present. Pressure maintenance was not considered when the field was being developed. For information and for those wishing to run this exercise using a completion design software package, Table 1 lists the PVT data for the Haggis fluids at current reservoir conditions.
Reservoir Temp. 150° F Oil API Gravity 40° API
Gas Relative Density 0.80 G.O.R. 550 scf/STB
Pb 2030 psia Bo 1.27
Oil Viscosity 0.66 cp Bg 0.0046
Gas Viscosity 0.022 cp Bw 1.023
Gas Z-Factor 0.73 Water Salinity 200000 ppm
Water Viscosity 0.67 cp Table 1: Haggis PVT Data
Haggis wells: The Haggis field wells have an economical limit of 1500 STB Oil/d/well; i.e. producing at rates lower than that is not feasible. Haggis-3 was drilled in May 1995. It is taken to be the case study for this field as it has average parameters for Haggis wells. Figure 1 is the completion diagram for Haggis-3. Above the wellhead, the well was completed with the same 5 1/2" OD production tubing encased in a mud line (no insulation). The mud line connects the wellhead (on the sea-bed) to the Xmas tree on the platform. For information, results of a recent pressure survey from Haggis-3 are listed in Table 2.
Depth 650 1605 2590 3600 4590 5587 6490 ft TVD
Pressure 525 735 990 1292 1629 1920 2266 psia Table 2: Haggis-3 Pressure Survey
The Haggis Field
Haggis-3's well parameters, and results from both well testing and production logging are summarised in Table 3. Table 3 also contains data on Haggis-1. Haggis-3 and Haggis-1 are essentially twin wells, except that completion damage has resulted in Haggis-1 being considered the worst well in the field.
Haggis-3 Haggis-1 Oil Production Rate 4730 3930 STB/d Water Cut 30 28 % WH Flowing Temperature 65 61 °F Pressure at Xmas tree 445 psia Skin (Well Test)
2.92 7.17
P.I. (J) (Well Test) 12.36 9.11 STB/d/psi Damaged Zone Relative Permeability 50 25
%
Damage Zone Thickness 12 in
Crushed Zone Skin 0.100 Drainage Radius 4000 ft
Table 3: Well Data
The Scenario: The rate of oil production decline in the Haggis field is alarming, and if no action is taken, Haggis will become uneconomical by the end of this year. The Operator of the field, Big Kahuna Oil Inc., does not accept this situation and has fired the field's former team leader for improper management of the field. Big Kahuna has hired you to improve production from the Haggis field. Your Mission: Big Kahuna Oil Inc. has asked you to study the field’s potential. A model has been created using the company approved software (EPS's FloSystem). A variety of production proposals have been modelled and the outcome of these simulations has been provided in graphical form. It is your job to evaluate these proposals using the all the data provided. The report should outline: A) the model used in the study, B) the potential of the base case scenario, C) your assessment of production enhancement proposals from the Haggis engineers and D) your recommendation for a project which will enhance production from Haggis.
Note: Wellflo graphs show gross production rates e.g. total production rates. Net production rates are required for the well analysis.
The Haggis Field
A) Develop a Well Model for Haggis-3:
a) Using Haggis-3 as your case study, complete the missing data in Figure 2.
b) The well Haggis-3 is used as the base case well for the Haggis field throughout this exercise. To minimise the computer time involved in simulations, the model contains only those components that contribute significantly to the pressure drop along Haggis-3. These are shown on Table 4.
Node No.
Component Name Measured Depth (ft)
1 Outlet node / Xmas tree 0 2 Riser 350 3 Wellhead 350 4 5.5” Tubing 850 5 S.C.S.S.S.V 850 6 5.5” Tubing 4000 7 5” Tubing 5600 8 7” Liner 6530.5
Table 3: Well Data
Compare this model to the completion design in Figure 1. When would it be necessary to include the nipples and tubing constrictions in the model design? __________________________________________________________________________________________________________________________________________________________
[2.5%]
c) The Outflow curve for Haggis-3 was modelled using a number of well-known correlations. The
sensitivity analysis is shown in Figure 3. The correlation that is most appropriate is: __________________________________________________________________________
[2%] There are two main reasons why this correlation was chosen: 1) __________________________________________________________________________
__________________________________________________________________________
2) ____________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
[2.5%]
d) The Big Kahuna Oil Inc. is unfamiliar with the concept of nodal analysis used in this flow simulation. You are asked to explain the process and the conditions for flow stability.
The inflow-outflow diagram below is for Haggis-3 at 30% water cut and utilising an appropriate flow correlation.
The Haggis Field
Curve A
Curve B
Point D
Point B
Point C
Point APoint E
Diagram 1: Haggis-3 base case Curve A represents the _________________________________________________. Curve B represents the _________________________________________________. The reservoir pressure is _____________psia (Point______).
The_________________ IPR model was used to produce this curve. Why was this chosen?
___________________________________________________________________________________________________________________________________.
Point ______ represents the operating point. The operating point is [stable/unstable]. Explain:
_____________________________________________________________________________________________________________________________________________________________________________________________________.
Nodal analysis may be carried out at any point in the producing system. In this report, the analysis was carried out to find the operating point at the sandface (see above diagram). Other typical examples of nodes selected during completion design are: 1)__________________________________________________________to evaluate _________________________________________________________________________________
2)__________________________________________________________to evaluate _________________________________________________________________________________
[11.5%]
The Haggis Field
B) Base Case Analysis: As a good manager, the first thing you have to do is evaluate the potential of what you have at the moment. To achieve that, you must determine what effect the decline in reservoir pressure and the increase in water cut will have on Haggis-3's production if nothing is done to improve its production. In other words, determine the reservoir pressure and the water cut at which Haggis-3 will becomes uneconomical to produce under the current production scenario.
a) In order to evaluate the viability of producing Haggis-3 as water-cut increases and reservoir
pressures reduces, a number of sensitivity analyses were performed on the production from Haggis-3 using WellFlo. These sensitivities are illustrated by Figures 4a, b, c and d, and their results should be summarised in Table 5:
PRes.
WC 2800 2700 2600 2500 psia
30% 4770 35% 40% 45%
Table 5: Haggis-3 Production Forecast [4%]
b) You discuss these figures with the field’s engineers. Since artificial lift can not be supported by
the production facilities on the Haggis platform, you agree with your engineers that you have to start a water injection scheme to maintain the reservoir pressure at 2800 psia. Figure 5 shows the well sensitivity to water cut. Under these circumstances, Haggis-3 will produce economically (1500+ BOPD) at maximum water cut of ____________, and at an oil production rate of ____________ BOPD. This is considered to be the Base Case scenario, against which all other schemes in section C will be compared.
Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
Base Case
c) With a water injection scheme in place, you expect to face even more severe water production
from Haggis-3. One way of dealing with such a problem is to plug-off “watered-out” perforations. List two advantages and two disadvantages of this scheme.
Advantages: 1)___________________________________________________________________________________________________________________________. 2)___________________________________________________________________________________________________________________________.
Disadvantages: 1)___________________________________________________________________________________________________________________________. 2)___________________________________________________________________________________________________________________________.
[5%]
The Haggis Field
d) A sensitivity study is outlined in Figures 6a, b and c for three plugging-off policies. Based on the advantages and disadvantages you identified above, the most advantageous plugging off policy is:____________________________________________________. The minimum open interval at which Haggis-3 will produce economically under this policy is ________________________ft.
[2.5%]
C) Further Projects: You phone management and convince them to fund the pressure maintenance scheme. In addition, they agree to provide a budget for an extra project to improve Haggis’s production. In this section you must select the best project from your engineers’ suggestions below. Since production, and thus revenue, from Haggis is greatly affected by water production form the field, Big Kahuna Oil Inc. defines the best production enhancement project for the Haggis field as the one that sustains economical production form Haggis-3 at the highest water-cut. Therefore, this should be the criterion you use to select the project you recommend to management.
C.1) Production Technology Solutions:
You walk into the team’s senior production technologist office, and give him the good news from management. He congratulates you, and immediately suggests that the project should be a production technology project. He goes on to explain that as the Haggis wells are extremely damaged, acidising all the wells make the field very profitable.
a) Determine the benefit from acidising Haggis-3 in terms of the maximum water cut at which the
acidised Haggis-3 will sustain economic production, if acidising restores the original rock permeability. Figure 7 shows the sensitivity of Haggis-3 to water cut after acidising. The maximum water cut at which the well can produce economically is ________%.
Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
Acidising
(Use this result to compare the benefits from this project to the other proposed projects, and as a basis for your recommendations in section D).
b) A production model of well Haggis-1 is already available and shows that Haggis-1 will produce
economically at maximum water cuts of 50 % after acidising. What are the implications of this if it is decided to carry out a campaign in which all the Haggis wells are to be acidised? __________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
[3.5%]
The Haggis Field
c) The diagram below shows the inflow-outflow curves for Haggis-3 at the maximum economic water cut after acidising. Assuming that acidising succeeds in restoring the original rock permeability in both wells, sketch on the diagram the inflow-outflow curves for Haggis-1.
Diagram 2: Haggis-3 after acidising
[2%] C.2) Suggestions from the Drilling Engineer:
The drilling engineer walks into your office and says that he heard that you were going around the different departments asking for ideas on how to improve the field's production. He says that he could have saved you all the trouble because he has the perfect solution. "Side-track the well", he says. He goes on to explain that, he would have gone for a horizontal well. However, since Big Kahuna have recently had bad luck with drilling horizontal sections longer than 400 ft, a 75° deviated well through the reservoir is probably the better option from a drilling point of view. a) Given the drilling department’s recent experience with horizontal wells, should you side-track
Haggis-3 to a 400 ft horizontal well running through the middle of the reservoir or propose a 75° deviated well penetrating the whole of the reservoir height?
The production prognosis from Haggis-3 sidetrack as a 400ft horizontal well is shown in Figure 8 while a 75° deviated well is illustrated in Figure 9. Summarise the results in the table below:
Option Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
C2.a1 400ft Horizontal C2.a2 75° Deviated
Option C2.a2 produces the better results because: ____________________________________________________________________________________________________________________________________________________________
[3%]
The Haggis Field
b) Given the engineering and economic factors and assumptions below, design the optimum horizontal well (to the nearest 500-ft) for the Haggis-3 side-track: Length of build-up section is: 2000 ft-MD Pay-back Time: 6 months Side-track cost: US$850 /ft-MD Price of Oil: US$15/bbl Cost of Processing and Shipping: US$8 / bbl Assumptions: 1) Economics based on un-dicounted cash flows
2) Production rate is constant for the first 6 months
3) Use Profit to Investment Ratio as measure of value, see below.
(Profit to Investment (PI) Ratio = TCS/MCO where TCS is Terminal Cash Surplus and MCO is Maximum Capital Outlay. The PI ratio is a measure of the cash surplus or profit generated for every unit of currency invested.)
Table 6 below can be completed using the results from Figure 10 and taking in account the above engineering and economic factors for drilling horizontal wells.
Horizontal length (ft) 500 1000 1500 2000 2500 3000 3500 4000
Parameterand unit Formula
Table 6:Economic analysis of sidetracking Haggis-3.
(In the space above, enter the parameter, formulas used and intermediate working) Table 6 indicates that the optimum horizontal section length for Haggis-3 is _____ ft. However, by carefully analysing the economic outcome and technical aspects of the design, what other conclusion could be drawn? _______________________________________________________________________________________________________________________________________________________________________________________________________________________________________
The Haggis Field
c) Determine the benefit from side-tracking Haggis-3 in terms of the maximum water cut you can economically produce the well with if it was side-tracked to the optimal horizontal length determined above.
If Haggis-3 is sidetracked to the optimal horizontal length, then the maximum water-cut at which the well will produce economically is ____%. (Figure 11)
Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
Horizontal Well
(Use this result to compare the benefits from this project to the other proposed projects, and as a basis for your recommendations in section D).
[15%]
C.3) Another Production Technology Suggestion: Artificial Lift
After lunch, Haggis’s senior production technologist steps into your office. He says that while he was having lunch, he remembered that Big Kahuna has been quite successful using Electric Submersible Pumps (ESP’s) in the nearby Tatties field. Additionally, he says that gas lift may be an option as the well has a high PI and the Big Kahuna has sufficient gas supply from another nearby field. He says that he is aware of the drilling department suggesting a slanted well, but installing artificial lift will be quicker and cheaper than side tracking the well. He suggests investigating the installation of an artificial lift scheme with a target liquid rate of 9000 STB/d.
C.3.1: Electrical Submersible Pumps
a) The Big Kahuna has had success with the following Centrilift pumps in the North Sea.
Pump Motor Cable Size GC 8200 562 Series #1 HC 7000 562 Series #1 HC 9000 562 Series #2 KC 12000 562 Series #2
Given the current conditions choose the optimum pump for Haggis-3 given the details and
engineering assumptions below,
Pump Types: Centrilift Setting Depth: 5000ft Minimum Equipment OD: 5” Maximum Equipment OD: 6.8” Platform Electricity Supply Frequency: 60Hz
Assumptions: 1) Production tubing is unaltered. 2) No wear on Pump or motor i.e. wear
factor is 1. 3) Efficiency of the gas separator is 100%
i.e. separator efficiency is 1. 4) Viscosity and gassiness corrections are
used.
Figures 12a & b show the performance plots of the pumps at current conditions. The most suitable pump of Haggis-3 given the present conditions is ______________.
The Haggis Field
This pump is the optimum choice because ________________________________________________________________________________________________________________________________________________________________________________________________________________________________.
[3%]
b) As water production is the limiting factor, the production technologist suggests halting water injection and allowing the reservoir pressure to drop. Determine which pump would be suitable in these conditions. Figures 13 a-d and Table C.3 are designed to assist you with that determination.
PRes.
Pump 2800 2600 2400 2200 Psia
GC 8200 HC 7000 HC 9000
KC 12000 Table C.3: Haggis-3 Production Forecast with ESP installed,
** denotes rate outwith the operating range of the pump.
According to Table C.3 the optimum ESP for Haggis-3 for declining reservoir pressure is ________________, at a water cut of 30%.
This pump is the optimum choice because ________________________________________________________________________________________________________________________________________________________________________________________________________________________________.
[5%] c) Determine the benefit from installing an ESP in Haggis-3 using Figure 14 in terms of the
maximum water-cut at which the optimised pump will sustain economic production prior to suspending water injection (i.e. no depletion, reservoir pressure 2800psia).
Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
Optimised ESP
(Use this result to compare the benefits from this project to the other proposed projects, and as a basis for your recommendations in section D). [1.5%]
The Haggis Field
C.3.2: Gas Lift Design a) A gas lift design for Haggis-3 is undertaken based on current conditions and engineering
assumptions/details outlined below,
Max casing head pressure: 1200 psi Gas available for injection: 20MM scf/d Injected gas gravity: 0.6 Setting depth: 3900 ft Valve differential pressure: 100 psia Minimum spacing: 450 ft “Kill” brine density: 0.465 psi / ft Minimum safety margin (see A on diagram) 50 psi
Assumptions: 1) Production tubing is unaltered. 2) Unload the tubing full of static fluid
against the well head pressure (i.e. static fluid to 0ft MD).
3) No transfer margin is required. The gas lift design is shown below (Diagram 3). Note briefly on the diagram the roles of the different valves in the design. The upper valves should be OPEN/CLOSED when assessing gas lift capabilities during field life.
A
Diagram 3: Gas lift design for Haggis-3 [3%]
The Haggis Field
b) Determine the optimum injection rate as the reservoir pressure declines from Figure 15 and summarise the results in Table C.4.
PRes.
2800 2600 2400 2200 2000 Psia
Optimum injection rate
MMscf/day
Table C.4: Optimum gas injection rate for Haggis-3. The required gas injection rate for the design production rate is __________________________.
The criteria used to choose the optimum injection rate are _________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________.
Higher injection rates do not improve production as the reservoir declines because ___________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________________
[8%]
c) The gas lift scheme is redesigned using the optimum gas injection rate for a reservoir pressure of 2800 psia. Determine the benefit from installing Gas Lift in Haggis-3 in terms of the maximum water-cut at which the optimised injection rate will sustain economic production using Figure 16.
Scenario Maximum Economic
Water Cut Production Rate @
30% Water Cut Optimised Gas Lift
(Use this result to compare the benefits from this project to the other proposed projects, and
as a basis for your recommendations in section D). [2%]
C.4) Suggestions from the Facilities Engineer:
The gentleman behind you introduces himself as the Haggis field's facilities engineer when you are walking to your car after a long day at the office. He shares with you a very interesting discovery he made today. He explains that if you agree to provide a dedicated line from the Haggis wells to the low-pressure separator, you can lower the average Xmas tree pressure to only 100 psia. He points out that suggestions from the other department have large uncertainties associated with them since they deal with the subsurface. His suggestion is simple, neat and will solve the field's problems.
The Haggis Field
a) Consider the base case inflow-outflow diagram for Haggis-3 below. Sketch on the diagram how the base case curves would alter if the well head pressure were decreased.
Diagram 4: Haggis-3 base case
[2%]
b) Evaluate the benefits of lowering the Xmas tree pressure, in terms of the maximum water-cut you can economically produce Haggis-3 with after the Xmas tree pressure is lowered to 100 psia.
If the production facilities on the Haggis platform can be modified to allow the Xmas tree pressure of Haggis-3 to be lowered to 100 Psia, then the maximum water-cut at which the well will produce economically will become _____%. (See Figure 17)
Scenario Maximum Economic
Water Cut Production Rate @ 30%
Water Cut Lowering Xmas Tree Press
(Use this result to compare the benefits from this project to the other proposed projects, and as a basis for your recommendations in section D).
c) List 2 advantages and 2 disadvantages of this scheme.
Advantages 1)__________________________________________________________________________________________________________________________________________________________ 2)__________________________________________________________________________________________________________________________________________________________ and disadvantages 3)__________________________________________________________________________________________________________________________________________________________ 4)_________________________________________________________________________________________________________________________________________________________
[5%]
The Haggis Field
D) Recommendations to Management:
a) Assess the production enhancement projects proposed by the Haggis field engineers in section C above and compare them to one another and to the base case scenario. Bear in mind that Big Kahuna Inc. has set the ranking criteria for these projects to be the maximum water-cut at which Haggis-3 can sustain economic production (i.e. > 1500 STB oil/d).
Table 8 below summarises the results from the various simulations carried out on the Haggis-3 well.
Scenario Maximum Economic Water Cut
Production Rate @ 30% Water Cut
Base Case
Acidising
75° Deviated Well
Optimum Horizontal Well
Optimum ESP
Optimum Gas Lift
Lowering Xmas Tree Pressure to 100 psia
Table 8: Haggis-3 Production Forecast [2.5%] b) Recommend to management a plan of action which either recommends maintaining the base case
scenario or executes one of the proposed projects.
Based on the WellFlo simulations and my assessment of them, I recommend that Big Kahuna Inc. invest in a water injection scheme to maintain Haggis’s reservoir pressure at 2800 psia. In addition, I recommend that Big Kahuna adopts [the _______________________ project /none of the projects investigated above] because: _____________________________________________________________________________________________________________________________________________________________.
[5%] 3) A number of risks have been overlooked in this assessment since the maximum water cut at which
the wells will flow at an economic rate has been used as the ranking criteria for the above projects. These risks add to the uncertainty of achieving the results on which your recommendation was based. As the Haggis field team leader it is your duty to report and account for these risks to management.
Complete Table 9 below which should identify three major risks that have been overlooked by this assessment. Briefly explain how each one could add to the uncertainty of the assessment and prescribe steps that need to be taken to account for their effects.
The Haggis Field
Risk/Uncertainty How this adds uncertainty to
above assessment? Steps that can be taken to account for / minimise this uncertainty
1)
2)
3)
Table 9: Haggis-3 Production Forecast [9%]
The Haggis Field
Figures:
The Haggis Field
The Haggis Field
The Haggis Field
Figure 3: Sensitivity to flow correlation.
The Haggis Field
Figure 4a: Layer pressure =2800 psia, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 4b: Layer pressure = 2700 psia, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 4c: Layer pressure = 2600 psia, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 4d: Layer pressure = 2500 psia, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 5: Layer pressure = 2800 psia, maximum economic water cut.
Operating point region
The Haggis Field
Plugging off policies Figure 6a: Plug off at 30% water cut, sensitivity to open interval.
Operating point region
The Haggis Field
Figure 6b: Plug off at 40% water cut, sensitivity to open interval.
Operating point region
The Haggis Field
Figure 6c: Plug off at 48% water cut, sensitivity to open interval.
Operating point region
The Haggis Field
Figure 7: Haggis-3 after acidising (e.g. damage zone permeability 200mD),
sensitivity to water cut.
Operating point region
The Haggis Field
Figure 8: 400ft horizontal well, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 9: 75° deviated well, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 10: Horizontal well, sensitivity to effective length.
Operating point region
The Haggis Field
Figure 11: Optimum horizontal well, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 12a: Performance curves for ESP GC8200 and HC7000
Figure 12b: Performance curves for ESP HC9000 and KC12000
The Haggis Field
Figure 13a: GC 8200, sensitivity to reservoir pressure
Operating point region
The Haggis Field
Figure 13b: HC 7000, sensitivity to reservoir pressure.
Operating point region
The Haggis Field
Figure 13c: HC 9000, sensitivity to reservoir pressure.
Operating point region
The Haggis Field
Figure 13d: KC 12000, sensitivity to reservoir pressure.
Operating point region
The Haggis Field
Figure 14: Optimum ESP, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 15: Performance analysis of the gas lift design, sensitivity to reservoir pressure and gas injection rate.
The Haggis Field
Figure 16: Optimum gas lift design, sensitivity to water cut.
Operating point region
The Haggis Field
Figure 17: 100psia well head pressure, sensitivity to water cut.
Operating point region
Course:- 28117Class:- 289033b
HERIOT-WATT UNIVERSITYDEPARTMENT OF PETROLEUM ENGINEERING
Examination for the Degree ofMEng in Petroleum Engineering
Production Technology 2
Thursday 2X April 200X09.30 - 11.45
NOTES FOR CANDIDATES
1. This is a Closed Book Examination.
2. 15 minutes reading time is provided from 09.15 - 09.30.
3. Examination Papers will be marked anonymously. See separate instructions forcompletion of Script Book front covers and attachment of loose pages. Do not write yourname on any loose pages which are submitted as part of your answer.
4. This Paper consists of 1 Section:-Attempt 3 numbered Questions from 4
5. Marks for Questions and parts are indicated in brackets
6. This Examination represents 70% of the Class assessment.
7. State clearly any assumptions used and intermediate calculations made in numericalquestions. No marks can be given for an incorrect answer if the method of calculation isnot presented.
1.(a) Sketch the main components of a 3 phase (gas/oil/water) horizontal separator and briefly
(one sentence) explain the function of each of the main components.[8]
(b) Indicate how the export of the oil/water/gas flows are controlled and why the outlets aresituated at your indicated locations.
[3](c) Stokes Law (below) describes the velocity of separation (v) of one liquid from another
VkD d c
c
=−( )2 ρ ρ
µ
Where D is the droplet size, ρ the density, µ the viscosity and c and d refer to the continuous anddiscontinuous phases respectively. An oil/water 2-phase separator has been in use in a field formany years. The main producing zone (35˚ API, saline formation water) is now depleted and itis proposed to produce a shallower, subsidiary zone (17˚ API oil, fresh formation water). Therequired data are given in Table 1.
You are required to advise management as to whether the existing separator capacity is sufficientwhen the subsidiary zone is producing at 1% and 75% water cut.
Table 1.
Fluid properties at Production zone separator conditions Main Sand Subsidiary SandOil viscosity cp 2.5 40 density/g.cm-3 0.85 0.95Water viscosity cp 0.9 0.7 density/g.cm-3 1.1 0.99
N.B. The production rate from the subsidiary zone is only 10% of that achieved from the mainzone.
[9](d) An assumption has to be made in the above calculations. Indicate its impact on the
conclusion reached in the unfavourable (separator capacity insufficient) case and indicatetwo remedial actions that could be taken.
[5]2.
(a) You are the Production Technologist responsible for completion of a well in a new field.Briefly list what techniques you would use to help you in the decision as to whether sandcontrol measures need to be installed.
N.B. A core has been taken across the pay zone.[8]
(b) This field has been declared marginal and can only be economically developed withsubsea wells. Briefly describe how this will affect your decision:
(i) on the need for the installation of sand control measures and(ii) type of sand control measures installed.
[5](c) The field is developed with an oil well producing through a gravel pack. The (Darcy)
skin due to presence of the gravel pack and the resulting pressure drop (∆Ps) may becalculated from:
Sk k L
d nand
PB
KLS
Dqd n
or
P q SDq
d n
g
s
s
=( )
= +
= +
2
4 2
4 2
96
141 2
0 00539
µ
/
.
.
∆
∆
q
(see Table 2 for definition of the parameters and numerical values)
Calculate the (Darcy) skin value (S) and the resulting pressure drop for a perforation density of 4shots/ft.
[4]This is the target, allowable pressure drop in the well.
(d) Well testing found that the turbulent (non-Darcy) resulted in an unacceptably highpressure drop of 374 psi. You are required to advise management as to whether the nextwell should be completed with:
Case Cost Shot Density DiameterA Low 12 shots/ft 0.5 inB High 4 shots/ft 1.0 in
and whether it will meet the target, allowable pressure drop.[5]
(e) Briefly comment on which case you would have expected to give the better inflow, andwhy.
[3]Table 2
Well Production (q) 2500 STB/DTotal Production Height (h) 23 ftReservoir Permeability (k) 578 mDOil Viscosity (µo) 0.310 cpFormation Volume Factor (Bo) 1.636 bbl/STB20-40 Mesh Gravel Permeability 120,000 mDPerforation Penetration (L) 6 inPerforation Diameter (d) 0.5 inPerforation Density (n) 4 shots/ftNon-Darcy (turbulance factor) (D) 0.01
3.(a) List up to 6 key features for both Rod Pumps and Gas Lift that form the basis of the fol
lowing statement:
“Worldwide, 85% of Artificial Lift equipment installed is rod pumps. This is mainly in stripper wells while gas lift is the most popular artificial lift technique for higher rate wells”.
[6]
(b) Most gas lift fields have insufficient gas to lift all the wells at their (technical) maximumproduction. Briefly describe the process of optimal allocation of available lift gas;mentioning the key economic parameters involved.
[6](c) Design a gas lift installation for the following conditions:
Tubing 3.958 inRequired Production Rate 3000 STB/dayOil Cut 100%Gas Oil Ratio 100scf/bblGas Specific Gravity 0.65Average Flowing Temperature 150˚FReservoir Productivity Index 4 bpd/psiReservoir Depth 10,000 ftReservoir Pressure 3400 psiLift Gas Injection Gradient 20 psi/1000 ftMinimum flowing tubing head pressure totransfer fluids to facility 250 psiDead Oil Density 35˚ API or 0.368 psi/ftGas Oil Ratio 100 scf/bblBrine Density 0.44 psi/ftLift Gas Injection Rate 3,000,000 scf/d
A pressure traverse curve is provided as Figure 1.
Tubing size, in. : 3.958
Liquid rate, STBL/D : 3000
Water fraction : 0
Gas gravity : 0.65Oil API gravity : 35Water specific gravity : 1.07Average flowing temp, F. : 150
Assume that the well is closed in with dead oil in the tubing and brine in the casing/tubingannulus.
(i) does this well require artificial lift to produce?[2]
(ii) what depth should the gas lift valve be installed in a single valve lift installation in orderto achieve the required production?
[6]
HINT: Note that the relevant portions of the pressure traverse curve can be approximated bystraight lines.
(iii) what is the minimum surface gas injection pressure to kick the well off in theconfiguration described?
[4]
(iv) how does this change if dead crude oil was present in the casing/tubing annulus insteadof brine?
[1]4.
(a) Briefly contrast the generalised selection criteria for matrix acidising and fracturingtreatments when considering carrying out a stimulation treatment on a well.
[5](b) List 2 sources of formation damage encountered during drilling and completion
operations and 3 damage sources during production operations. Briefly indicate how thefluid selection for a (matrix) removal treatment will be influenced by the damage source(examples may clarify your answer).
[6]
(c) A well completed on 40 acre spacing (re = 745 ft) has a damaged region extending 1 ftbeyond the wellbore (rw = 0.328 ft).
The Hawkins formula may be used to calculate the skin due to formation damage:
Skk
rrd
o
d
d
w
= −
1
while the productivity ratio (Ji/Jd) of the well with and without the above formation damage isgiven by:
JiJd
In rr S
In rr
e
w
e
w
=
+
Use the above to illustrate the statement:
{“Formation Damage reduces well productivity greatly while the stimulation effect ofincreasing the near wellbore permeability above the initial value has limited effect”.}
HINT : estimate the relative well productivity with 95%, 75%, 50% formation damage and 10times increase in near wellbore formation permeability.
[6]
(d) Your service company has designed the following fracturing treatments:
Wellbore radius (rw): 0.328 ftReservoir height: 100 ft; bounded by competent shalesReservoir Permeability: 0.1 mDProppant available: 300,000 lb
Design Fracture Conductivity (kf*w):treatment A- 1500 mD.ft at 4 lb/ft2 proppant loading
treatment B- 850 mD.ft at 2 lb/ft2 proppant loading
(i) Use the accompanying graph (Figure 2) from Cinco-Ley and Samiengo to advisemanagement as to whether treatment A or B will give the highest well productivities.
[6]
(ii) Why would you expect one of these treatments to be preferred?[2]
End of Paper
Model Solutions to Examination
1
Date:
1. Complete the sections above but do not seal until the examination is finished.
2. Insert in box on right the numbers of the questions attempted.
3. Start each question on a new page.
4. Rough working should be confined to left hand pages.
5. This book must be handed in entire with the top corner sealed.
6. Additional books must bear the name of the candidate, be sealed and be affix ed to the first book by means of a tag provided
Subject:
INSTRUCTIONS TO CANDIDATES
8 Pages
PLEASE READ EXAM INATION RE GULATIONS ON BACK COVER
No. Mk.
NAM
E:REGISTRATION N
O.:
COURSE:
YEAR:
SIGNATURE:Complete this section but do not
seal until the examination
is finished
Production Technology 2
2
Model Solutions to Examination
3
Answer to Q1Answer to Q1Answer to Q1Answer to Q1Answer to Q1
1(a) (Standard) 3-phase horizontal separator diagram.
Gas
Mist Eliminator
Oil Outlet andLevel Control
Water Outlet andLevel Control To Oil Export
To ProducedWater Treatment
InletFrom ProductionManifold
Inlet Deflector/ Momentum Breaker
Pressure Control Valve
OIL and EMULSION
WATER
GAS
PC
Weir
OIL
Main componentsMain componentsMain componentsMain componentsMain components Main functionMain functionMain functionMain functionMain function
Inlet momentum breaker => to reduce inlet flow velocity (hence
helping disengage free gas)
Mist eliminator/extractor => to remove liquid drops from gas
(by allowing liquid droplets to impinge
on the wire, coalesce and flow down in
to the liquid phase)
Weir => to separate oil and water phases (oil
is collected here and skimmed off
from main settling area)
Liquid outlets under level controls => to evacuate oil and water (usually
equipped with a vortex breaker to
prevent re-entrainment of gas)
4
Bonus marksBonus marksBonus marksBonus marksBonus marks
- coalescer pack
- solids wash facility
- production chemical addition
- fresh water wash
1(b) In a gravity type separator => the separation process takes
place due to the action of gravitycombined with the difference in
densities of the phases/components of the produced fluid.
For a 3-phase separator (vertical/horizontal) the heavier fluids (water
and oil) go to the bottom of the vessel while the lighter phase (gas)
rises to the top. Therefore:
* Water outlet is located at the bottom of the vessel, remote from
the inlet to allow maximum residence time. It is controlled by the oil/
water interface level control.
* Oil outlet is also located at the bottom, remote from the inlet to
allow maximum residence time. It is connected to the oil level control.
* Gas will flash off and disengage from the liquid phases and exit from
the uppermost part of the vessel remore from the inlet (to allow maxi
mum residence time). Gas outlet is regulated by a pressure control
valve which controls the separator operating pressure.
Model Solutions to Examination
5
1(c) Assume droplet size distribution the same for subsiduary and
main production zones.
V
Vd c
d c
c
c
1
2
1 1
2 2
2
1
=−( )−( )
µµ
ρ ρρ ρ
.
75% water cut75% water cut75% water cut75% water cut75% water cut
• Water is the continuous phase
• Oil is the discontinuous phase => i.e. oil droplets are separating from
the water phase.
V
Vmain
subsiduary
om wm
os ws
ws
wm
=−( )−( )
µµ
= −( )−( )
= ××
=
ρ ρρ ρ
.
. .. .
...
. .
. ..
0 85 1 10 95 0 99
0 70 9
0 25 0 70 04 0 9
4 9
Separation of oil droplets from the water phase takes 5 times longer
for the more viscous oil. Under these conditions, an increase of
residence time of at least 5 times is required in order to achieve the
same efficiency of separation.
Since well production is reduced by factor 10, residence time is in
creased by factor 10 (i.e. 10 times). Therefore separator capacity is
sufficient and does not need upgrading.
1% water cut1% water cut1% water cut1% water cut1% water cut
• Oil is the continuous phase
• Water is the discontinuous phase => i.e. water droplets are separating
from the oil phase.
6
V
Vmain
subsiduary
wm om
ws os
os
om
=−( )−( )
µµ
= −( )−( )
= ××
=
ρ ρρ ρ
.
. .. .
...
.. .
1 1 0 850 99 0 95
4 02 5
0 25 400 04 2 5
100
Separation of water droplets from the oil phase takes 100 times
longer, i.e would require an increase of residence time of at least 100
times to carry on with the separation process
Despite production rate being reduced by factor 10, separation
capacity is still factor 10 too small. Therefore separation capacity is
insufficient and needs upgrading.
1(d) Separation rate increases as the square of the droplet
diameter - therefore an increase in the waterwaterwaterwaterwater droplet size by a factor
3 would be sufficient to compensate for above factor 10.
Possible remedial measures are:
• installation of a coalescer pack to reduce the distance oil and water
droplets have to separate
• add demulsifier chemicals to speed up the rate of droplet
aggregation by coalescing the small oil particles to form larger
particles which will separate faster (Stokes law - the D term is
squared & so the improvement is squared).
Model Solutions to Examination
7
Answer to Q2Answer to Q2Answer to Q2Answer to Q2Answer to Q2
2(a) The main techniques used when deciding whether to install sand
control measures are:
• Field Experience• Field Experience• Field Experience• Field Experience• Field Experience
History of any sand production problems in wells producing from the
same formation in the same field, or wells in the same or similar
formation but on other nearby fields, will be the best guide for
installing sand control measures in a new well. Evaluation of the
drawdown conditions and rock stresses under which sand production
started in those wells will be very useful.
Sanding tendencies may be detected by performing high drawdown
tests in exploration wells.
• Petrophysical and core analysis (sonic travel time)• Petrophysical and core analysis (sonic travel time)• Petrophysical and core analysis (sonic travel time)• Petrophysical and core analysis (sonic travel time)• Petrophysical and core analysis (sonic travel time)
Based on laboratory measurements on core material:
i) The sonic travel time is proportional to the porosity i.e. the higher
the porosity, the longer the travel time.
ii) The rock strength is inversely related to porosity i.e. the lower the
porosity, the greater the rock strength.
Sand production tendency is dependent on rock strength, which
depends on porosity and grain cementation. A measure of porosity and
8
grain cementation can be obtained from the sonic travel time through
the rock matrix. This can be done by acquiring an acoustic wireline log.
The well sonic travel time log can thus be processed to derive a
continuous estimate of the formation strength.
BonusBonusBonusBonusBonus: Shell petrophysical risk approach to sand failure prediction:
• Well site strength estimation• Well site strength estimation• Well site strength estimation• Well site strength estimation• Well site strength estimation (core recovery, onsite core tests and
completion experience)
Simple tests at well site can determine if core is friable (by scratching
it or simply scraping it with the fingernails) and therefore a potential
sand problem may be expected. Other simple measures include
observation of the core recovery status and completion experience e.g.
sand run into the casing when the drill pipe or tubing is out of the hole.
• Rock mechanical measurements and calculations• Rock mechanical measurements and calculations• Rock mechanical measurements and calculations• Rock mechanical measurements and calculations• Rock mechanical measurements and calculations (unconfined
compressive strength, brinnel hardness, thick wall cylinder
strength, triaxial rock strength)
Bonus marks for extra description of the Rock mechanical tests 1Bonus marks for extra description of the Rock mechanical tests 1Bonus marks for extra description of the Rock mechanical tests 1Bonus marks for extra description of the Rock mechanical tests 1Bonus marks for extra description of the Rock mechanical tests 1
to 4 below:to 4 below:to 4 below:to 4 below:to 4 below:
Model Solutions to Examination
9
1) Unconfined or Uniaxial Compressive Strength1) Unconfined or Uniaxial Compressive Strength1) Unconfined or Uniaxial Compressive Strength1) Unconfined or Uniaxial Compressive Strength1) Unconfined or Uniaxial Compressive Strength - In this test, an
unsupported cylinder of rock is loaded axially at a steady rate to
failure. The maximum stress reached is called the Unconfined or
uniaxial compressive strength (UCS).
2) Brinell Hardness Number2) Brinell Hardness Number2) Brinell Hardness Number2) Brinell Hardness Number2) Brinell Hardness Number - is the load required to press a standard
spherical indenter (metal disk) a constant distance into a slabbed core
face. The strength of rock is proportional to the hardness number i.e.
the stronger the rock, the greater the load required and the higher
the BHN
3) Thick wall cylinder collapse strength3) Thick wall cylinder collapse strength3) Thick wall cylinder collapse strength3) Thick wall cylinder collapse strength3) Thick wall cylinder collapse strength - A hollow, thick wall cylinder
is created by drilling a narrow hole in the middle of a rock and placed
in a rubber sleeve. Then, the hollow cylinder is loaded axially and
radially until it collapses {the Thick Wall Collapse Strength ((TWC)}.
4) Triaxial test4) Triaxial test4) Triaxial test4) Triaxial test4) Triaxial test - Similar to uniaxial test but a radial stress also
applied to the core. The axial stress imposed by the end pieces and the
radial stress are controlled separately. The maximum stress prior to
failure increases as the radial stress increased i.e. the rock sample
shows stronger behaviour as the radial, confining stress increases.
This test gives most information on the rock strength properties.
2(b) i) In a marginal field any type of workover operations is often
not economic. This means that the decision on whether or not to install
sand control has to be made correctly the first time. Further, if in
10
stalled, the sand control equipment should not impair the well
productivity below the economic limit since a corrective workover may
not be economically viable.
ii) In a Subsea well, “living with sand” is not an option since we are
unable to detect sand production can not manage its presence in subsea
flowlines.
Subsea wells have high intervention costs. Hence the need to minimise
workover requirements by correct, first time design and use of high
reliability equipment since remedial measures may not be affordable.
The two alternatives regarding sand control installation are:
- If sand control is installed but not neededinstalled but not neededinstalled but not neededinstalled but not neededinstalled but not needed it will probably cause
additional pressure drops and reduce the well Productivity Index.
Consequently well economics will be affected
- If sand control is not installed but is neededis not installed but is neededis not installed but is neededis not installed but is neededis not installed but is needed an intervention is
likely to be necessary to install sand control and/or repair blocked/
eroded equipment.
The specific case needs to be studied and a risk analysis carried out.
Sand control options include internal & external gravel packs, prepacks,
slotted liners, wire wrapped screens and chemical treatments. Gravel
packs give a higher pressure drop but also provide excellent sand
Model Solutions to Examination
11
control if designed correctly. Slotted liners are relatively cheap but
only stop coarse grained sand effectively. Wire wrapped screens are
more expensive and more effective, provide greater flow area and
lower pressure drop. Pre-packed screens are even more expensive with
good sand control but are susceptible to plugging during installation. It
is advisable to install a completion with high productivity greater
{e.g. horizontal well with liner} than that required based on
conventional well inflow analysis to allow for lower (e.g. gravel packed)
well productivities often observed in practice.
2(c) At a production rate of 2500 STB/D and a perforation density
of 4 shots/ft
S
P
P psi
s
s
= ( ) ( )( ) ( )
=
= ( ) +
=
96 578 120 000 60 5 4
2 77
0 00539 2500 2 770 01 2500
0 5 4
374
2
4 2
* / , *. *
.
. * * .. *
. *∆
∆
2(d) Case A
S
P
P
s
s
=
= ( ) +
=
0 92
0 00539 2500 0 920 01 25000 5 12
49 8
4 2
.
. * * .. *. *
.
∆
∆ psi
12
Case B
S = 0..
. * * .. *
*
.
69
0 00539 2500 0 690 01 2500
1 4
30 4
4 2∆
∆
P
P
s
s
= +
= psi
Case B perforating is recommended.
2(e) Doubling the perforation diameter increases the inflow area by
a factor 4 while the increase in shot density increases the inflow area
by a factor 3 only. Wide diameter perforations provide a much smaller
pressure drop and also create less non-darcy flow effects (no data
given in the question).
Wide diameter perforations are normally preferred for gravel pack
completions (easier placement of gravel without a screen out).
Answer to Q3Answer to Q3Answer to Q3Answer to Q3Answer to Q3
3(a) Rod PumpRod PumpRod PumpRod PumpRod Pump
• Typically low rates and moderate depths
• Relatively cheap to install & run
• Rod pumps are mechanically simple to operate and easy to repair/
maintain/replace. Can be operated by inexperienced personnel
• Sensitive to gas and solids (wax/scale/sand) - Solids can damage
moving parts
• Not suitable for (highly) deviated wells
Model Solutions to Examination
13
• Obtrusive in urban locations. Heavy equipment for location offshore
• Pump can be easily changed and performance monitored
• (reasonably) Viscous oil can be handled
Gas LiftGas LiftGas LiftGas LiftGas Lift
• capable of high production rates
• Suitable for water drive reservoirs with relatively high bottomhole
pressure gradients
• The above, coupled to high well Productivity Indices and high
formation permeabilities result in a high Flowing Bottom Hole Pressure,
limiting achievable reservoir depletion
• Gas has to be available
• Wireline serviceable up to 65˚ deviation
• Flexible - valve settings can be adjusted for optimum performance
based on actual, well conditions
• can be used off-shore
• Fully open tubing giving access for logging
• tubing, and annular surface controlled, subsurface safety valves
available
• Limited surface requirements once gas available
• Forgiving of poor design & operation, but difficult to run efficiently
• Can handle (tolerate) produced solids e.g. reasonable formation sand
concentrations
• High GOR => advantage rather than a drawback
N.B. Only 6 items are required from the above lists
14
3(b) Optimal allocation of Lift Gas
• Allocate each increment of gas to that well with the highest
incremental oil production until all gas allocated.
• Implies slope of (net oil/lift gas rate) should be the same for each well.
• A brief description of the process is as follows:
- Perform a gas lift tubing performance calculation with a range of
injected gas flow rates for each well in the field. Then, plot the
results on a curve of liquid flow rate achieved versus injected gas rate.
Each well has its own characteristics (geometry and flow capacity) and
will show a different rate of oil production increase with incremental
increase on the gas lift injection rate.
- Calculate the incremental oil production rate in each well for each
extra increment of gas injected.
- Allocate each increment of gas to that well with the highest
incremental oil production until all gas allocated.
- Implies slope of (net oil/lift gas rate) should be the same for each well.
- This is a complex process that has many variables such as gas
availability, number of wells, etc, and even becomes even more
complicated by those wells that require kick-off lift gas.
- There is software available to assist with this analysis.
Bonus pointsBonus pointsBonus pointsBonus pointsBonus points
• Check economic limit, when cost of incremental lift gas equals
income from incremental oil not exceeded
• Chosen rates should be sufficiently far from technical maximum
production (where extra gas decreasesdecreasesdecreasesdecreasesdecreases production rate) to avoid
Model Solutions to Examination
15
unstable well operation.
3(c) (i) Dead oil in tubing height 3,400 / 0.368 = 9,239 ft
Fluid level = 10,000 - 9,239 = 761 ft
Well is dead with fluid level at 761 ft
(ii)GOR after gas injection = 3,000,000 / 3,000 + 100 = 1,100 SCF/STB
From pressure traverse plot:
- Average flowing gradient afterafterafterafterafter lift gas injection:
between 6,000 ft and surface = 600 / 6,000 = 0.1 psi/ft
- Average flowing gradient priorpriorpriorpriorprior to lift gas injection
from 10,000 ft to 4,000 ft = (3,040-900)/(10,000-4,000) = 0.357 psi/ft
FWP + ∆ P above valve + ∆ P below valve + Draw down = Reservoir Pressure
Drawdown = Production Rate/PI = 3,000 / 4 = 750 psi
Let depth of valve = x
250 0 1 10 000 0 357 750 3 400
0 1 0 357 3 400 1 000 3 570
1 170 0 257
4 550
+ + −( ) + =− = − −
==
. * , * . ,
. . , , ,
, / .
,
x x
x x
x
x ft
16
Install gas lift valve at 4,550 ft.
(iii) Minimum surface gas injection pressure (P)
P + Gas gradient = Hydrostatic head at 4,550 ft + FWHP
P + (4,550) 0.02 = (4,550) 0.44 + 250
P = 2,252 - 91
P = 2,161 psi
(iv) Required injection gas pressure decreases by the ratio of the
brine and dead oil fluid densities.
If dead crude was in the annulus
P1 + 91 = (4,550) 0.368 + 250
P1 = 1,833 psi
Answer to Q4Answer to Q4Answer to Q4Answer to Q4Answer to Q4
4a) Matrix (Acidising) - removal of near wellbore damage
Hydraulic Fracturing - improving well inflow performance
Model Solutions to Examination
17
SkinSkinSkinSkinSkin PermeabilityPermeabilityPermeabilityPermeabilityPermeability TreatmentTreatmentTreatmentTreatmentTreatment
High High Matrix treatment
High Medium Matrix/frac and pack
High Low Fracturing (matrix possible)
Low Low Hydraulic fracturing
Low High Treatment economic?
Stimulation treatment should increase well Productivity Index:
Matrix acidising aims to increase the Productivity Index (PI) by
reducing the Skin (S) through dissolving formation damage components
and rock in the near wellbore region. It is particularly suitable for
medium/high perm reservoirs with high skin.
Fracturing increases PI by increasing effective wellbore radius throgh
the creation of high conductivity fractures from wellbore. Fractures
bypass the damaged zone and extend to a greater depth into the
reservoir than can be reached by acidising. It is suitable for low
permeability reservoirs with or without skin.
4b) Drilling and Completion FluidsDrilling and Completion FluidsDrilling and Completion FluidsDrilling and Completion FluidsDrilling and Completion Fluids
• solid block pore throats => mud solids invasion leading to blocked pore
throats => reduce permeability
• solids have many sources => drilling mud components, drilled
formation particles etc.
• Fluid loss increases liquid saturation
18
• incompatible formation and fluid loss reduces invaded zone
permeability
• Mud filtrate invasion => clay swelling => reduced permeability
• Water block => from increasing water saturation near wellbore due to
loss of drilling / completion fluid
Production Operations:
• Scale => due to temperature and pressure changes. Salts in
formation water become over saturated and precipitate e.g. calcium
carbonate.
• Wax
• Asphaltenes => Asphaltene precipitation due to pressure drop.
• Fines migration
• Sand production due to increased effective stress
Matrix treatment fluid should dissolve (or at least mobilise) the source
of the formation damage, e.g.
wax - hot, organic fluids
calcium carbonate scale - hydrochloric acid
clay particles - mud acid
bonus marksbonus marksbonus marksbonus marksbonus marks: choice of fluid has to maintain the materials in solution
and avoid later reprecipitation of the dissolved materials.
Model Solutions to Examination
19
4c) Skk
rr
JJ
rr S
rr
JJ
S
d
d
w
i
o
e
w
e
w
i
o
= −
=+
= +
1
7 737 73
ln
ln
ln
..
r = 0.328, rD = 1.328 ln r / r = 1.40
r = 0.328, r = 745 ln r / r = 7.73
w d w
w e e w
Damage removal
K/kd S Ji/Jo Jo/Ji
20 26.6 4.44 0.234 4.2 1.54 0.652 1.4 1.18 0.851 0 1.00 1.00
Stimulation
K/kd S Ji/Jo Jo/Ji
0.5 -0.70 0.91 1.100.1 -1.29 0.84 1.20
The above tables reflect the statement “Formation damage reduces
well productivity greatly while the stimulation effect of increasing thenear wellbore permeability above the intial value has limited effect.”
20
4d) i) Treatment A:Treatment A:Treatment A:Treatment A:Treatment A:
Fracture length (both wings) = Proppant
loading * frac. height
= 300,000 / (4*100) = 750 ft
Fracture half length (Xf) = 375 ft
FCD = Kf*w/k*Xf = 1,500 / (0.1*375) = 40
From graph
ln (Xf/rw) + Sf = 0.75
∴ Sf = 0.75 - ln (375/0.328)
Sf = - 6.29
Treatment B:Treatment B:Treatment B:Treatment B:Treatment B:
Fracture half length = 300,000 / (2*100*2) = 750 ft
FCD = 850/(0.1*750) = 11.3
From graph
Sf = 0.8 - ln (750/0.328)
= - 6.93
∴ Treatment B achieves a more negative skin and is recommended.
Model Solutions to Examination
21
ii) Long fractures are more effective at stimulating low permeability
wells providing the fracture conductivity is sufficient (FCD > 10 - 15).
22
Course:- 28117Class:- 289033a
HERIOT-WATT UNIVERSITYDEPARTMENT OF PETROLEUM ENGINEERING
Examination for the Degree ofMEng in Petroleum Engineering
Production Technology 1b
Friday 23rd April 199909.30 - 12.30
NOTES FOR CANDIDATES
1. This is a Closed Book Examination.
2. 15 minutes reading time is provided from 09.15 - 09.30.
3. Examination Papers will be marked anonymously. See separate instructions for completion ofScript Book front covers and attachment of loose pages. Do not write your name on any loosepages which are submitted as part of your answer.
4. This Paper consists of 2 Sections:- A and B.
5. Section A & B:- Attempt 4 numbered Questions from 7 with at least 1 Question from eachSection
6. Marks for Questions and parts are indicated in brackets
7. This Examination represents 55% of the Class assessment.
8 State clearly any assumptions used and intermediate calculations made in numerical questions.No marks can be given for an incorrect answer if the method of calculation is not presented.
9. Answers must be written in separate, coloured books as follows:-
Section A:- BlueSection B:- Green
SECTION A
A1. “Advanced wells and in particular horizontal and multi-lateral wells, can enhance the businesscase of a field development by any of 3 primary techno-economic drivers.” What are these?
[3]
“Multi-lateral well configurations can in the main be classified as stacked, opposed or planar, butthe selection of the optimum geometry must be based on the reservoir structure and flow characteristics.” Discuss this statement giving examples to illustrate the application of the variousoptions - use sketches as appropriate.
[7]
A2. Two subsea well completion designs are shown in Figures 1 and 2 - review and compare each ofthese designs for the following applications:
(a) 10,000ft T.V.D. oil producer, normally pressured with a GOR of 400scf/bbl.[5]
(b) Water injection completion for the same reservoir.[5]
A3. The well shown in Figure 3 is an overpressured oil producer. If pressure buildup is experiencedin the 103/
4” x 7” annulus at surface:
(a) What are the potential sources of the pressure?[3]
(b) What method(s) and tools could be used to identify the cause of the leakage(s)?[4]
(c) What corrective measures would you propose for the causes identified in (b) above?[3]
Typical InjectionCompletion Schematic
Typical ProducerCompletion Schematic
Tubing hanger
3.812" RQ' Landing nippleinjection valve
4 1/2 " 2.6lb/ft new vam tubing
4 1/2 " alloy 'MMG' SPM
4 1/2 " 12.5 lb/ftvam tubing
Seal unit
4 1.2"PBR
Tubing anchor latch
Millout extension
Wireline re-entry guide
2.750" XN nipple
2.313' XN' No-Go landing nipple
3.1 2/9' /AM perforated at
9 5/8' SAB-3 HYD set permanent packer
3.812' 'XN' No-Go lancing nipple
PBTD
PERFS
PBTD
PERFS
Annuluscheckvalve
Annuluscheckvalve
4 1/2" TRCHSV
4 1/2" 12,6 lb/ft VAM J55 Tubing
4 1/2" Alloy 'MMG' SPM
4 1/2" PBR
9 5/8" 40lb/ft NBS vam casing
9 5/8" SAB-3' HYD set permanentpacker O/W millout extension
Tubing anchor latch
Seal unit
3.912" 'XN' No-Go landing nipple
3.588" RN' No-Go landing nipple
2.756" XN' No-Go landing nipple
2.313" XN' No-Go landing nipple
Wireline re-entry guild
Figure 1 and 2
Perforations
7" CRA production liner
7" TR SSV
7" CRA tubing
9 5/8" liner
A-Ryte sealassemblew/ anochor latch
Liner top isolation packer
10 3/4" tie -back
Liner hanger
Liner hanger w/lower swal bore
9 5/8" shoe @10700'
Liner top isolation packerw/ upper seal bore andBi-directional slipes and w/ bulletseals on lower seal assembly
Full bore nipple profile
13 3/8" shoe @ 5700'
7" shoe @ 14500'
Figure 3
SECTION B
B4.(a) Sketch the main components of a 3 phase (gas/oil/water) horizontal separator and briefly (one
sentence) explain the function of each of the main components.[8]
(b) Indicate how the export of the oil/water/gas flows are controlled and why the outlets are situatedat your indicated locations.
[3]
(c) Stokes Law (below) describes the velocity of separation of one liquid from another
Vkd d c
c
= −2 ( )ρ ρµ
An oil/water 2-phase separator has been in use in a field for many years. The main producingzone (35˚ API, saline formation water) is now depleted and it is proposed to produce a shallower,subsidiary zone (17˚ API oil, fresh formation water). The required data are given in Table 1.You are required to advise management as to whether the existing separator capacity is sufficientwhen the subsidiary zone is producing at 1% and 75% water cut.
Fluid properties at Producing zoneseparator conditions Main sand Subsidiary sandOil viscosity cp 2.5 40 density/gcm-3 0.85 0.95Water viscosity cp 0.9 0.7 density/gcm-3 1.1 0.99
N.B. The production rate from the subsidiary zone is only 10% of that achievedfrom the main zone.
[9]
(d) An assumption has to be made in the above calculations. Indicate its impact on the conclusionreached in the unfavourable (separator capacity insufficient) case and indicate two remedialactions that could be taken.
[5]
B5.(a) You are the Production Technologist responsible for completion of a well in a new field. Briefly
list what techniques you would use to help you in the decision as to whether sand control measures need to be installed.
N.B. A core has been taken across the pay zone.[8]
(b) This field has been declared marginal and can only be economically developed with subseawells. Briefly describe how this will affect your decision:
(i) on the need for the installation of sand control measures and(ii) type of sand control measures installed.
[5]
(c) The field is developed with an oil well producing through a gravel pack. The (Darcy) skin due topresence of the gravel pack and the resulting pressure drop (∆Ps) may be calculated from:
Sk k L
d nand
PqB
KLS
Dq
d n
or
P q SDq
d n
g
s
s
=
= +
= +
96
141 2
0 00539
2
4 2
4 2
( / )
.
.
∆
∆
µ
(see Table 2 for definition of the parameters and numerical values)
Calculate the (Darcy) skin value (S) and the resulting pressure drop for a perforation density of 4shots/ft.
[4]This is the target, allowable pressure drop in the well.
(d) Well testing found that the turbulent (non-Darcy) resulted in an unacceptably high pressure dropof 374 psi. You are required to advise management as to whether the next well should becompleted with:
Case Cost Stort Density DiameterA Low 12 shots/ft 0.5 inB High 4 shots/ft 1.0 in
and whether it will meet the target, allowable pressure drop.[5]
(e) Briefly comment on which case you would have expected to give the better inflow, and why.[3]
Well production (q) 2500 STB/DTotal production height (h) 23 ftReservoir permeability (k) 578 mDOil viscosity (µo) 0.310 cpFormation volume factor (Bo) 1.636 bbl/STB20-40 mesh gravel pearmwability 120,000 mDPerforation Penetration (L) 6 inPerforations Diameter (d) 0.5 inPerforation Density (n) 4 shots/ftNon-Darcy (turbulence factor) (D) 0.01
B6.(a) List up to 6 key features for both Rod Pumps and Gas Lift that form the basis of the following
statement:
“Worldwide, 85% of Artificial Lift equipment installed is rod pumps. This is mainly in stripperwells while gas lift is the most popular artificial lift technique for higher rate wells”.
[6]
(b) Most gas lift fields have insufficient gas to lift all the wells at their (technical) maximumproduction. Briefly describe the process of optimal allocation of available lift gas; mentioningthe key economic parameters involved.
[6]
(c) Design a gas lift installation for the following conditions:
Tubing 3.958 inRequired Production Rate 3000 STB/dayOil Cut 100%Gas Specific Gravity 0.65Average Flowing Temperature 150˚FReservoir Productivity Index 4 bpd/psiReservoir Depth 10,000 ftReservoir Pressure 3400 psiLift Gas Injection Gradient 20 psi/1000 ftMinimum flowing tubing head pressureto transfer fluids to facility 250 psiDead Oil Density 35˚ API or 0.368 psi/ftBrine Density 0.44 psi/ftLift Gas Injection Rate 3,000,000 scf/d
A pressure traverse curve is provided as Figure 4.
Assume that the well is closed in with dead oil in the tubing and brine in the casing/tubingannulus.
(i) does this well require artificial lift to produce?[2]
(ii) what depth should the gas lift valve be installed in a single valve lift installation in order toachieve the required production?
[6]
HINT: Note that the relevant portions of the pressure traverse curve can be approximated bystraight lines.
(iii) what is the minimum surface gas injection pressure to kick the well off in the configurationdescribed?
[4]
(iv) how does this change if dead crude oil was present in the casing/tubing annulus instead ofbrine?
[1]
B7.(a) Briefly contrast the generalised selection criteria for matrix acidising and fracturing treatments
when considering carrying out a stimulation treatment on a well.[5]
(b) List 2 sources of formation damage encountered during drilling and completion operations and 3damage sources during production operations. Briefly indicate how the fluid selection for a(matrix) removal treatment will be influenced by the damage source (examples may clarify youranswer).
[6]
(c) A well completed on 40 acre spacing (re = 745 ft) has a damaged region extending 1 ft beyondthe wellbore (rw = 0.328 ft).
The Hawkins formula may be used to calculate the skin due to formation damage:
Sk
k
r
rdo
d
d
w
= −
1
while the productivity ratio (Ji/Jd) of the well with and without the above formation damage isgiven by:
Ji
Jd
rr S
rr
e
w
e
w
=
+
ln
ln
Use the above to illustrate the statement:
“Formation Damage reduces well productivity greatly while the stimulation effect of increasingthe near wellbore permeability above the initial value has limited effect”.
HINT : estimate the relative well productivity with 95%, 75%, 50% formation damage and10 times increase in near wellbore formation permeability.
[6]
(d) Your service company has designed the following fracturing treatments:
Wellbore radius (rw): 0.328 ftReservoir height: 100 ft; bounded by competent shalesReservoir Permeability: 0.1 mDProppant available: 300,000 lbDesign Fracture Conductivity (kf*w): treatment A
- 1500 mD.ft at 4 lb/ft2 proppant loading
treatment B- 850 mD.ft at 2 lb/ft2 proppant loading
(i) Use the accompanying graph from Cinco-Ley and Samiengo to advise management as towhether treatment A or B will give the highest well productivities.
[6]
(ii) Why would you expect one of these treatments to be preferred?[2]
0.1 1
1
2
010 100 1000
FCD =
Sf +
ln (
x f/r
w)
Kf.w
K.xf
End of Paper