Measurement of leakage and estimation of force coefficients in a short length seal supplied with a

liquid/gas mixture (stationary journal)Luis San Andrés,

Xueliang Lu and Qing Liu

1

TRC project 32513/1519NS

Year IIIMay 2015

2

Bubbly & Foamy Annular Pressure Seals

• As oil fields deplete compressors work off-design with liquid in gas mixtures, mostly inhomogeneous.

• Similarly, oil compression station pumps operate with gas in liquid mixtures.

• The flow condition affects compressor or pump overall efficiency and reliability.

• Little is known about seals operating under 2-phase conditions, except that the mixture affects seal leakage, power loss and rotordynamic force coefficients; perhaps inducing random vibrations that are transmitted to the whole rotor-bearing system.

JustificationSeals operate with either liquids or gases, but not both……

Subsea Application: Wet Gas Compression

Brenne, L., Bjørge, T., Bakken, L., and Hundseid, Ø., 2008, "Prospects for Sub Sea Wet Gas Compression," ASME Paper GT2008-51158.

Brenne, L., Bjørge, T., Gilarranz, J., Koch, J., and Miller, H., 2005, “Performance Evaluation of A Centrifugal Compressor Operating under Wet Gas Conditions,” Proc. 34th

Turbomachinery Symposium, Houston, TX, pp. 111~120

Bertoneri, M., Duni, S., Ransom, D., Podesta, L., Camatti, M., Bigi, M., and Wilcox, M., 2012, “Measured Performance of Two-stage Centrifugal Compressor under Wet Gas Conditions,” ASME Paper GT2012-69819.

Vannini, G., Bertoneri, M., Del Vescovo, G., and Wilcox, M., 2014, “Centrifugal Compressor Rotordynamics in Wet Gas Conditions,” Proc. 43rd Turbomachinery & 30th

Pump Users Symposia, Houston, TX.

Bertoneri, M., Wilcox, M., Toni, L., and Beck, G., 2014, “Development of Test Stand for Measuring Aerodynamic, Erosion, and Rotordynamic Performance of a Centrifugal Compressor under Wet Gas Conditions,” ASME Paper GT2014-25349.

Announce / assess the effects of two phase flow in wet gas compression: drop in efficiency and reliability, excessive vibration and rotordynamic stability.

3

4

Experimental – Seals (two phase)

Iwatsubo, T., and Nishino, T., 1993, “An Experimental Study on the Static and Dynamic Characteristics of Pump Annular Seals,“ 7th Workshop on Rotordynamic Instability Problems in High Performance Turbomachinery, Texas A&M University.

Computational – Seals (two phase)

Literature: Two Phase Flow in Annular Seals

Hendricks, R.C., 1987, "Straight Cylindrical Seals for High Performance Turbomachinery," NASA TP-1850.

Arauz, G., and San Andrés, L., 1998, “Analysis of Two Phase Flow in Cryogenic Damper Seals, I: Theoretical Model, II: Model Validation and Predictions,” ASME J. Tribol., 120.

Beatty, P.A., and Hughes, W.F., 1987, "Turbulent Two-Phase Flow in Annular Seals," ASLE Trans., 30.

Arghir, M., Zerarka, M., Pineau, G., 2009"Rotordynamic Analysis of Textured Annular Seals with Mutiphase (Bubbly) Flow, “Workshop : “Dynamic Sealing Under Severe Working Conditions” EDF – LMS Futuroscope, Universite de Poitiers.San Andrés, L., 2012, “Rotordynamic Force Coefficients of Bubbly Mixture Annular Pressure Seals,” ASME J. Eng. Gas Turbines Power, 134.

NEED EXPERIMENTAL

validation.

.

4

5

Experimental – Seals (two phase)Iwatsubo, T., and Nishino, T., 1993, “An Experimental

Study on the Static and Dynamic Characteristics of Pump Annular Seals,“ 7th Workshop on Rotordynamic Instability

Problems in High Performance Turbomachinery.

NO description of water lubricated seal (L, D, c) or gas type…..

Tests at shaft speed 1.5-3.5 krpm and supply pressure=1.2 - 4.7 bar.

Gas/liquid volume fraction =0, 0.25, 0.45, 0.70

Mxx

Cxx

Kxx

, gas volume fraction increases

Stiffness, damping and mass coefficients decrease

as gas volume fraction increases.

6

• Application–Subsea compression and pumping(wet compressors must operate with LVF up to 5%)

• Effect of two component flow on seal– Leakage rate– Dynamic force coefficients– Stability

• Status of current research: – Bulk-flow model available– Lack of experimental validation

Motivation

• Prepared test rig for measurements and improved lubrication system.

• Made oil-gas mixtures with liquid volume fraction (LVF) at inlet plane: gas to liquid.

• Measured flow rates with pure oil, all air, and with mixture.

• Designed (manufactured) new test rig.

7

Past TRC work 2013-14 – Year I

• Revamped vertical test rig and troubleshoot.• Generated mixtures and measured flow rate.

– LVF = 0.02% (air) to 100% (liquid)– Supply pressure to 50 psig (3.5 bar (abs))– Rotor speed: stationary

• Conducted dynamic load tests and estimated force coefficients.– Installed shakers and performed single frequency

dynamic load measurements– Identification of force coefficients

• Visual inspection of bubbly mixture. 8

Completed work 2014-15 Year II

9

Wet Seal Test Rig

Main frame

Support pipe

Thermocouple

Accelerometer

Load cell

Eddy current sensor

Dynamic pressure sensor

Bearing cartridge

Rotor

Top plate

Air-Oil Mixture

Shaker Rotor

Bearing

Shaker

Shaker

Stinger

seal clearance

X

ZY

10

Seal cartridge & fluid propertiesSeal

Diameter 127 mm (5 in)Length 46 mm (1.8 in)clearance 0.127 mm(5 mil)

Supply pressure 1.0~3.5 bar (abs)

Oil ISO VG 10density 830 kg/m3

viscosity 18 cP at 20ºCAir density 1.2 kg/m3 at

1barviscosity 0.02 cP at 20ºC,

1bar

Flow

Seal

rotor & journal

L/D=0.36

X

ZY

11

Flows & Mixture

Air InletOil Inlet

(ISO VG 10)

Test seal section

Valve

ValveSparger(mixing) element

Oil

Air

Ps/Pa=1.5, inlet LVF= 2%

Flow rate of mixture vs. supply pressure 1.1 ~3.5bar (abs)Room temperature ~20ºC– Seal inlet liquid volume fraction (at supply pressure):

12

Flow Rate Measurement

Mass fraction is constant as there is no mass transfer between oil and air.

LVF= Qlβ =inlet PaQ +Ql g Ps

Oil

Air

0.012

0.014

0.016

0.018

0.02

0.022

0 0.2 0.4 0.6 0.8 1

Mass flow ra

te of m

ixture [k

g/s]

seal inlet LVF

Mass Flow Rate in Wet SealSupply PS=3.5 bar(abs), ambient Pa=1 bar(abs), (non-rotating) journal.

Liquid mass fraction () is large even for small LVF because of large density ratio (oil/air)~200 .Flow ranges from turbulent (pure air) to laminar (pure oil).Predicted flow ~ 6% lower than measured.

L/D=0.36 c=0.127 mm

Ps/Pa=3.5

Reynolds #

Liquid mass fraction

mass flow rate

Measurement

Prediction

Rapid drop

Flow visualization in Wet Sealambient pressure Pa=1 bar(abs),

inlet temperature 20oC. Non-rotating journal.

Inlet LVF=0.45

Smaller bubbles with

higher supply

pressureL/D=0.36 c=0.127 mm

Seal land length: 46mm

Top

Bottom

Top

Bottom

PS=1.5bar(abs)

=0.45

PS=3.0 bar(abs)

=0.45

Bubble expansion

15

Force Coefficients in Annular SealsSeal reaction force is a function of the fluid

properties, flow regime, operating

conditions and geometry.

For small amplitudes of rotor motion: the force

is linearized with stiffness, damping

and inertia force coefficients:

c

rotor

L

D

Axial pressure field (liquid)

stator

Pa

PS Pe

W Axial velocity

X

Y

Film thickness H=c+eX coseY sin

rotor

X XX XY XX XY XX XY

Y YX YY YX YY YX YY

F K K C C M Mx x xF K K C C M My y y

Model system (2-DOF): structure + sealEOM: Time Domain

EOM: Frequency Domain

Measure:

Estimate Parameters:

K C M

K, C, M

Parameter Identification

S SEAL S SEAL BC SEAL(K +K )z+(C +C )z+(M +M )z=F

2iω ω K + C - M z = F

Load F=Fo sin(t) Displacement z,

R e ( )

Im ( ) (?)

2ω

ω

H K - M

H C

1( ) ( ) 2iω ω H z = F H K + C - M = F z

0

5

10

15

0 20 40 60 80 100

Dis

plac

emen

t (um

)

Frequency (Hz)

inlet LVF 0%inlet LVF 4%

-90

-60

-30

0

30

60

90

0 0.02 0.04 0.06 0.08 0.1

Load

(N)

Time (s)

inlet LVF 0% inlet LVF 4%0

10

20

30

40

50

0 0.02 0.04 0.06 0.08 0.1

Dis

plac

emen

t (um

)

Time (s)

inlet LVF 0% inlet LVF 4%

Applied load and seal motion X direction, frequency = 30 Hz

FFT

0

5

10

15

20

0 20 40 60 80 100

Load

(N)

Frequency (Hz)

inlet LVF 0%

inlet LVF 4%

FFT

Load Displacement

Unexplained drift at low frequencies

Seal coefficients

(K, C, M)SEAL = (K, C,M) – (K, C, MBC)S

SEALTest system(lubricated) Structure (dry)

•Instrumental Variable Filter Method (IVFM) (Fritzen, 1985, J.Vib, 108)

2Re H K M Im (?) H CComplex stiffness

Parameter Identification

-16

-12

-8

-4

0

4

0 50 100 150 200

Rea

l(H) (

MN

/m)

Frequency (Hz)

Hᵪᵪ

Hᵧᵧ

K - Mω²0

4

8

12

16

20

0 50 100 150 200

Imag

inar

y(H

) (M

N/m

)

Frequency (Hz)

Hᵪᵪ

Hᵧᵧ

Ps/Pa=2, inlet LVF 2%

Test System Dynamic Stiffness2.0 = Supply pressure PS/ambient pressure Pa

inlet temperature 20oC. (non-rotating) journal.

βinlet=0.02

βinlet=0.04

liquid volume fraction at inlet:

K-2M fits well the test

data. -16

-12

-8

-4

0

4

0 50 100 150 200

Rea

l(H) (

MN

/m)

Frequency (Hz)

Hᵪᵪ

Hᵧᵧ

K - Mω²

Uncertainty ±0.7MN/m

-16

-12

-8

-4

0

4

0 50 100 150 200

Rea

l(H) (

MN

/m)

Frequency (Hz)

Hᵪᵪ

Hᵧᵧ

K - Mω²

Test System Quadrature Stiffness2.0 = Supply pressure PS/ambient pressure Pa

inlet temperature 20oC. (non-rotating) journal.

liquid volume fraction at inlet (

Ima(H) shows damping is frequency

dependent. Very little

damping with all gas.

βinlet=0.0

0.0 (all air)

(mm=6±0.8 g/s)

βinlet=0.02

0.90

(mm=6±0.2 g/s)

βinlet=0.04

0.95

(mm=7±0.2 g/s)

20

0

0.2

0.4

0.6

0.8

1

0 50 100 150 200

Imag

inar

y(H

) (M

N/m

)

Frequency (Hz)

Hᵪᵪ

Hᵧᵧ

0

4

8

12

16

20

0 50 100 150 200

Imag

inar

y(H

) (M

N/m

)

Frequency (Hz)

Hᵪᵪ

Hᵧᵧ

Hᵪᵪ Hᵧᵧ

Inlet LVF 0%

Inlet LVF 4%

Uncertainty: ±0.7MN/m

10

100

1,000

10,000

100,000

0 50 100 150 200

Cyy

_sea

l (N

-s/m

)

Frequency (Hz)

10

100

1,000

10,000

100,000

0 50 100 150 200

Cxx

_sea

l (N

-s/m

)

Frequency (Hz)

Wet Seal Direct Damping C=Im(H)/2.0 = Supply pressure PS/ambient pressure Pa

Stationary (non-rotating) journal. βinlet=0.04

βinlet=0.02

βinlet=0 (all air)

βinlet=0.02

βinlet=0

βinlet=0.04

liquid volume fraction at inlet 0% (all gas), 2%, 4%.

Damping is frequency dependent.

Small LVF causes damping to increase 20

fold w/r to air only.

CXX & CYY differ due to uneven clearance and

inhomogeneous flow

seal stiffness & mass coefficients

LVF at seal Liquid mass KXXseal KYYseal MXXseal MYYseal

inlet fraction MN/m MN/m kg kg

0 0 1.4 1.0 0 0

2% 90% 1.5 1.3 0.3 0.7

4% 95% 2.5 1.3 1.2 0.9

With mixture, CXX>CYY due to uneven clearance and mixture not homogenous (one phase travels faster than the other).

Wet

23

Conclusion• Mass flow rate through seal dictated by liquid

content (density ratio ~ = 830/1.2=692).

• System dynamic stiffness K-2M fits well with testdata.

• Damping coefficients are frequency dependent.For operating conditions with a small volume fractionof oil in gas (4%), damping can be up to twenty timeslarger than that obtained for the pure gas condition.

• The effect of a few droplets of liquid on affectingthe test system forced response is overwhelming.

LEAKAGE AND FORCE COEFFICIENTS IN AWET ANNULAR SEAL: INFLUENCE OFJOURNAL ROTATIONAL SPEED.

24

2015-2016 Continuation Proposal Year III

• Redesign stingers and mount shakers rigidly to avoid low frequency drift.

• Install a bubble eliminator to remove air content from the bubbly lubricant.

• Improve DAQ process to perform ensemble averages over multiple periods of external excitation.

• Benchmark bulk-flow model predictions1.• Extend existing predictive model1 to consider

inhomogeneous flow.

25

Proposed Work

San Andrés, L., 2012, “Rotordynamic Force Coefficients of Bubbly Mixture Annular Pressure Seals,” ASME J. Eng. Gas Turbines Power, 134 (2), p. 022503.

2015-2016 Year III

26

TRC Budget 2015-2016 Year III

Support for GS(20 h/week) x $ 2,200 x 12 months $ 26,400

Support for UGS (10 h/week) x 9 months $ 2,250

Fringe benefits (2.7%) & medical insurance ($150/month) $ 2,574

Registration and travel to technical Conference $ 1,250

Tuition three semesters ($ 363 credit hour x 24 ch/year) $ 8,712

Supplies: Bubble eliminator $1,050+flow meter repair $790+ oil $600

$ 2,440

PC upgrade ($1,150) + Mathcad ($105) + storage ($100) $ 1,355

Total BUDGET $ 44,981

Acknowledgments

Questions (?)

27

Thanks to the Turbomachinery Research Consortium for its support