Numerical Simulation of Basal Aquifer Depressurization in the Presence of Dissolved Gas – An Update

Rudy Maji, Ph.D. Karl Lawrence, Ph.D. Ken Baxter, M.Sc., P.Geol. Scott Morton, P.Geol.

What to Take Away From this Presentation

Reservoir modelling typically considers multiphase flow;

Groundwater modelling has only dealt with single phase flow;

A regional scale reservoir model is required to address the dissolved gas issues during basal aquifer depressurization in advance of mining;

Originally, this challenge was addressed through a modified application of MODFLOW (finite difference approach);

This presentation provides the results of a better approach developed using FEFLOW (finite element approach).

April-16-14 2

Outline

Background

Key Questions to be Answered

Ground Water Modelling Approach Model Construction Model Calibration Prediction Results

Summary and Conclusions

April-16-14 3

Background

April-16-14 4

Background – Project Location

Calgary, Alberta, Canada

Fort McMurray, Alberta, Canada

April-16-14 5

Background - CNRL Horizon Oil Sands Project

Located about 75 km north of Fort McMurray in northeastern Alberta

Open pit mining operation and oil sands plant for extracting and upgrading bitumen from the McMurray Oil sands deposit.

April-16-14 6

Background - CNRL Horizon Oil Sands Project

Prior to mining the piezometric elevations within primary aquifer varied between 276 m to 280 m asl

An aquifer depressurization plan consisting of 6 pumping wells were planned around the South Pit to depressurize the basal aquifer

April-16-14 7

Background - CNRL Horizon Oil Sands Project

DP network operational since 2009

In early 2012, the DP well network consisted of 35 vertical in-pit pumping wells and 2 horizontal pumping wells

Depressurization has been achieved in the area of active mining.

2006 DP Well Network Planning

2009 DP Well Network Operational

2010-2011 Mine Sequence Excavation

April-16-14 8

Background – Why? Gas Locking

If there are substantial amounts of dissolved gas in the groundwater, “gas-locking” may occur during the depressurization of aquifer

Pore pressure reductions lower the gas solubility;

When pressure reduces below critical value (“bubble pressure point”), exsolution occurs;

Presence of gas ↑ effective specific storage or ↓ effective permeability;

Permeability reduction results in steeper drawdown curves and reduced depressurization extent.

April-16-14 9

Background – Effects of Gas Locking

Reduction in DP well efficiency – depressurization is limited locally;

Additional DP wells may be required to achieve the target water level elevations.

Cost and schedule implications.

April-16-14 10

Background – Basal Aquifer Isopach

Mining starts from the west edge of the pit

South Pit Contours

April-16-14 11

Background – Key Indicators

0

5

10

15

20

25

30

35

40

45

50

55

60

65

700 30 60 90 120 150 180 210 240 270 300 330 360 390 420 450 480 510 540 570 600 630 660 690 720 750 780 810 840

Elapsed Time (min)

Dra

wd

ow

n (

m)

DP-03 Datalogger OBW @ DP-03 Datalogger OBW @ DP-03 Manual DP-03 Manual

Separation between pumping well and observation well = 31.0 m

Q1 = 157 m3/d (24 igpm)

Q2 = 333 m3/d (51 gpm)Q3 = 655 m3/d (100 igpm)Q4 = 982 m3/d (150 igpm)

Available drawdown to pump intake = 65.1m

Q1

Q2

Q3

Q4

April-16-14 12

Key Questions

April-16-14 13

Key Questions to be Answered

How many wells are required to be installed in the DP network?

How much water to be handled?

April-16-14 14

Modelling Approaches

April-16-14 15

Groundwater Modelling Approaches

Develop Regional Model in Multiphase Flow Modelling Framework. Numerically complex; Computationally memory extensive.

Develop Alternative Model to Build the Solution for Gas Locking Problem

in the Single-Phase Modelling Framework; Easy to solve using an existing regional model; Computationally less intense.

Regional Model in Visual MODFLOW; Regional Model in FEFLOW.

April-16-14 16

Single Phase Flow Modelling - MODFLOW

MODFLOW model was revised to account for dissolved gas issues The hydraulic conductivity values were revised based on the

simulated pressure at discrete time steps.

April-16-14 17

MODFLOW Approach – Discrete Reduction

210

220

230

240

250

260

270

280

290

300

0 50 100 150 200 250 300 350

Corr

ecte

d H

ead

(mas

l)

Time (min)

SIM-NO GAS LOCKING

DP-03-MEASURED

DP-03-SIM

DISCRETE STEPS

OBW4-MEASURED

OBW4-SIM

18 April-16-14

Single Phase Flow Modelling - FEFLOW

An external module was developed in FEFLOW Automated update of hydraulic conductivity during each timestep

based on: Aquifer pressures; and Relative permeability relationship.

April-16-14 19

FEFLOW Approach - Continuous Reduction

20 April-16-14

0

200

400

600

800

1000

12000

10

20

30

40

50

60

70

0 100 200 300

Time [min]

Pum

ping

Rat

e [m

3/da

y]

Draw

dow

n [m

]

DP-03

DP-03 - SIM

OBW

OBW - SIM

Pumping Rate

Single Phase Flow Modelling - FEFLOW

Gas characterization Pressurized sampling

Multi-phase flow (reservoir) modelling CNRL

Relative permeability relationship 30% reduction in permeability for every 10 m reduction in head below the bubble point pressure (BPP)

April-16-14 21

Single Phase Flow Modelling - FEFLOW

𝐾𝐾 = 1 − 𝑝𝐻−ℎ𝐷 × 𝐾 … obtained from the relative permeability and

pressure saturation relationship

p% of permeability reduction was found for D m of pressure drop

H = 240 masl (BPP) h = head in each node K = Original Hydraulic Conductivity

April-16-14 22

Calibration to a Pumping Test

April-16-14 23

With Gas-Locking – 30% reduction in K per 10 m

0

200

400

600

800

1000

12000

10

20

30

40

50

60

70

0 100 200 300

Time [min]

Pum

ping

Rat

e [m

3/da

y]

Draw

dow

n [m

]

DP-03

DP-03 - SIM

OBW

OBW - SIM

Pumping Rate

Calibration (Well Head)

Observed

Simulated

Comparison Between Approaches

FEFLOW Model

Reduction in effective permeability is continuous (i.e. at every timestep) Implemented through external module (IFM) which is linked to FEFLOW at run-time separate module created to

reduce K

shorter runtimes

minor user interaction (only need specify K reduction function)

Visual MODFLOW Model

Reduction in effective permeability is discrete (i.e. not at every timestep) The K field was revised at discrete timesteps based on the simulated heads and bubble point pressure model run was interrupted to

reduce K

longer runtimes

major user interaction

April-16-14 24

History Match

April-16-14 25

History Match – Pumping Wells

Pumping schedule irregular

Horizontal wells are

modelled

40 scheduled stress (pumping) periods

43 pumping wells

April-16-14 26

History Match – DP Obs Wells (with gas-lock)

April-16-14 27

History Match – DP Obs Wells (no gas-lock)

April-16-14 28

History Match – DP Obs Wells (with gas-lock)

April-16-14 29

History Match – DP Obs Wells (no gas-lock)

April-16-14 30

History Match - Summary

Detailed history match of operational volumes completed with irregular pumping schedule; 2 horizontal wells; 41 vertical wells; 40 scheduled stress (pumping) periods

Suitable match obtained given approach and uncertainties

April-16-14 31

Prediction Results

April-16-14 32

Estimation of Flow Volumes and Number of Wells

North South Total [m3/day [m3/day [m3/day 2012 711.23 98.46 809.69 2013 655.55 292.44 947.99 2014 482.81 555.90 1038.71 2015 457.22 662.96 1120.18

No Gas Locking No Gas Locking – Estimate # Wells

With Gas Locking – Estimate # Wells

North South Total

8 1 9

7 3 11

5 6 12

5 7 12

Number of Wells Operating

North South Total

24 3 27

22 10 32

16 19 35

15 22 37

Number of Wells Operating

April-16-14 33

Summary and Conclusions

April-16-14 34

Summary and Conclusions

Characterization of dissolved gas during the early stage of the project is critical;

A predictive tool has been developed to design the DP well network for gassy aquifers;

Greater number of wells are required to depressurize a given aquifer in the presence of dissolved gas;

The presence of dissolved gas could significantly impact the overall project schedule and budget.

April-16-14 35

Numerical Simulation of Basal Aquifer Depressurization in the Presence of Dissolved Gas – An UpdateWhat to Take Away From this PresentationOutlineBackgroundBackground – Project LocationBackground - CNRL Horizon Oil Sands ProjectBackground - CNRL Horizon Oil Sands ProjectBackground - CNRL Horizon Oil Sands ProjectBackground – Why? Gas LockingBackground – Effects of Gas LockingBackground – Basal Aquifer IsopachBackground – Key IndicatorsKey QuestionsKey Questions to be AnsweredModelling ApproachesGroundwater Modelling ApproachesSingle Phase Flow Modelling - MODFLOWSlide Number 18Single Phase Flow Modelling - FEFLOWFEFLOW Approach - Continuous ReductionSingle Phase Flow Modelling - FEFLOWSingle Phase Flow Modelling - FEFLOWCalibration to a Pumping TestComparison Between ApproachesHistory MatchHistory Match – Pumping WellsHistory Match – DP Obs Wells (with gas-lock)History Match – DP Obs Wells (no gas-lock)History Match – DP Obs Wells (with gas-lock)History Match – DP Obs Wells (no gas-lock)History Match - SummaryPrediction ResultsEstimation of Flow Volumes and Number of WellsSummary and ConclusionsSummary and Conclusions