Relative Permeability and Capillary Pressure Controls on CO2 Migration and Brine

Displacement

Sally M. Benson1

Ljubinko Miljkovic2, Liviu Tomutsa2 and Christine Doughty2

1Energy Resources Engineering Dept., Stanford University2Earth Sciences Division, Lawrence Berkeley National Laboratory

Acknowledgements

• Funded by DOE Fossil Energy through the Zero Emissions Research and Technology Program (ZERT)

• Outstanding co-authors from Lawrence Berkeley National Laboratory– Ljubinko Miljkovic– Liviu Tomutsa– Christine Doughty

Some Key Issues for CO2 Storage in Deep Saline Aquifers

• What fraction of the pore space can be filled with CO2?• How big will the CO2 plume be?• How much CO2 will be dissolved?• How much will capillary trapping immobilize CO2?• Can accurate models be developed to predict CO2 fate and

transport?

Courtesy of Christine Doughty, LBNL

Answering these questions depends on the complex interplay of viscous, capillary, buoyancy forces and heterogeneity and structure on CO2 plume migration.

GravityViscous and

capillary forces Heterogeneity Structure

Core-flood Set-Up for Relative Permeability Measurements

*Brine composition: CO2 saturated brine with 0.5 molar potassium iodide

CO2 Brine

Constant Displacement

Pumps

Overburden Pressure: 100 bars

DifferentialPressureTransducer

Pressure Data Acquisition

CO2

Constant Pressure(65 bars)

Brine

Room Temperature: 16.5o C

75 mm38 mm

3.142

=CO

wμ

μ

Core-Scale Imaging of CO2 Distributions

High Pressure PumpsCore Holder

In Scanner

CT Scans Measure Core Porosity

7.8cm

3.8cm

Φ=0.22Φ=0.22

0.12 0.330.22

0.16

0.22

0.28

0 7.82.0 4.0 6.0Core Length (cm)

Por

osity

Calculation of Permeability

PorosityPorosity

Kozeny-Carmen

ki = Φi3

S(1-Φi)2

PermeabilityPermeability

Core Permeability Distribution

30 1000

k=301mDk=301mD

7.8cm3.8cm

500

50

300

550

0 7.82.0 4.0 6.0Core Length (cm)

Per

mea

bilit

y (m

D)

Laboratory Injections of Various CO2-Brine Proportions

• Experimental Setup:5%, 10%, 20%, 50%, 80%, 90%, 100% CO2 injections

3mL/min constant flow-rate

6.89MPa constant back-pressure

16 ±2°C lab temperature

Brine contains dissolved CO2

CO2 contains dissolved water

• Measure CO2 Saturation with CT ScannerDigitally reconstruct image

Relative Permeability CurvesR

elat

ive

Per

mea

bilit

y

Brine Saturation

CO2 Brine

Small-scale CO2 Saturation Variations

5% CO2 10% CO2

20% CO2 50% CO2 80% CO2

90% CO2 100% CO2

CO2 Saturation:0% 100%50% 75%25%

Sub-corescale saturation variations generally overlooked in relative permeability measurements.

Simulated Injection of Various CO2-Brine Proportions

• Simulation Cases10%, 90%, 100% CO2 injections

3mL/min constant flow-rate

6.89MPa constant back-pressure

16°C constant temperature

Brine contains dissolved CO2

CO2 contains dissolved water

• Core CharacterizationPorosity/permeability “map”coarsened

Relative permeability/capillary pressure curves matched to experimental curves

• TOUGH2 (Pruess, LBNL)

180px

60px

36px

Homogeneous Simulations

10%CO2

90%CO2

100%CO2

Variable Φ, k Simulations

CO2 Saturation:0% 70%

Simulated CO2 SaturationsConstant Pc Produces Homogeneous CO2 Saturations

PorosityLab Data

Fitting Capillary Pressure CurveP

c (P

a)

Brine Saturation

1000

10,000

100,000

Simulation Input Curve*

*Silin et al. (submitted, 2007

Hg Injection Data Curve

2000 8000Pcap (Pa)given 20% CO2

Pcap = 4500PaPcap =

4500Pa

Φi

ki√Pc,i ∝

10%CO2

90%CO2

100%CO2

Variable Φ, k Simulations

CO2 Saturation:0% 70%

Lab Data

Simulated CO2 SaturationsVariable Pc Produces Small-scale CO2 Saturation Variations

Variable Pc Simulations

Capillary Pressure CurveP

cap (P

A)

Brine Saturation

CO2 Saturation:

0% 70%10% CO210% CO290% CO290% CO2100% CO2100% CO2

Avg. Pc Φ=.22 k=301mD Pc Envelope

Why should we care?

Why Should We Care?

Average CO2 saturation is:

‣ Decreased by sub-corescale heterogeneity

‣ Flow-rate dependent

‣ Affected by simulation grid resolution

Subcore-scale Heterogeneity Decreases CO2 Saturation

0 7.84.02.0 6.0Length Along Core (cm)

CO

2S

atur

atio

n

100% CO100% CO22

InjectionInjection

90% CO90% CO22

InjectionInjection

10% CO10% CO22

InjectionInjection

Effects of Flow Rate on CO2 Saturation90% CO2 Injection Simulation

Frac

tiona

l Flo

w (C

O2

)CO2 Saturation

Inje

ctio

n R

ate

(mL/

min

)

Distance from Well (m)Brine Saturation

Rel

ativ

e P

erm

eabi

lity

CO

2 S

atur

atio

n

Length of Core (cm)

3.09.0

mL/min

0.3

0.0650.03

0.001

Buckley-Leverett

0 87621 5

Capillary Pressure Distribution at Different Flow RatesC

apill

ary

Pre

ssur

e (P

a)

Capillary Pressure Curves:Average: ϕ=0.22 k=206mD Upper Bound: ϕ=0.12 k=35.7mDLower Bound: ϕ=0.25 k=444mD

9 mL/min3 mL/min0.3 mL/min0.065 mL/min0.03 mL/min0.01mL/min

Brine Saturation

90% CO2, 10% Brine InjectionVariable Simulation Resolutions

Grid Size: 0.6x0.6x3mmGrid Count: 67,350

CO2 Saturation:

0% 55%30%

Grid Size: 1x1x3mmGrid Count: 23,400

Grid Size: 2x2x3mmGrid Count: 5,400

Finer Simulation Grids Decrease CO2 Saturation

Length Along Core (cm)

CO

2S

atur

atio

n

Conclusions

• Core-scale multi-phase flow experiments reveal strong influence of sub core-scale heterogeneity

• Spatial variations in capillary pressure behavior control CO2saturations

• CO2 saturation:– Decreases due to bypass of low porosity regions– Decreases at lower flow rates– Predictions depend on grid size

• Similar phenomena are expected at all spatial scales• Fundamental research needed to improve model predictions

– Fundamental process understanding based on lab and field experiments– Up-scaling strategies that accurately include the effects of sub-grid scale

heterogeneity– Calibration and validation of predictive models