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 University
2Earth 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?
Viscous and
capillary forces
Heterogeneity
Gravity
Structure
Answering these questions depends on the complex
interplay of viscous, capillary, buoyancy forces and
heterogeneity and structure on CO2 plume migration.
Courtesy of Christine Doughty, LBNL
Core-flood Set-Up for Relative Permeability
Measurements
Overburden Pressure: 100 bars
μw
μCO = 14.3
2
38 mm
CO2
Brine
75 mm
Differential
Pressure
Transducer
CO2
Brine
Pressure Data Acquisition
Constant
Displacement
Pumps
*Brine
o
Room Temperature: 16.5 C
composition: CO2 saturated brine with 0.5 molar potassium iodide
Constant
Pressure
(65 bars)
Core-Scale Imaging of CO2 Distributions
High Pressure Pumps
Core Holder
In Scanner
CT Scans Measure Core Porosity
3.8cm
Φ=0.22
7.8cm
0.12
0.22
4.0
Core Length (cm)
6.0
0.33
Porosity
0.28
0.22
0.16
0
2.0
7.8
Calculation of Permeability
Porosity
Kozeny-Carmen
Φ i3
ki =
S(1-Φi)2
Permeability
Core Permeability Distribution
3.8cm
k=301mD
Permeability (mD)
7.8cm
30
500
1000
550
300
50
0
2.0
4.0
Core Length (cm)
6.0
7.8
Laboratory Injections of Various
CO2-Brine Proportions
•
•
Experimental Setup:
¾
5%, 10%, 20%, 50%, 80%, 90%, 100% CO2 injections
¾
3mL/min constant flow-rate
¾
6.89MPa constant back-pressure
¾
16 ±2°C lab temperature
¾
Brine contains dissolved CO2
¾
CO2 contains dissolved water
Measure CO2 Saturation with CT Scanner
¾
Digitally reconstruct image
Relative Permeability Curves
Relative Permeability
CO2
Brine Saturation
Brine
Small-scale CO2 Saturation Variations
5% CO2
20% CO2
10% CO2
50% CO2
80% CO2
90% CO2
100% CO2
Sub-corescale saturation variations generally overlooked in relative
permeability measurements.
CO2 Saturation:
0%
25%
50%
75%
100%
Simulated Injection of Various CO2Brine Proportions
•
•
•
Simulation Cases
¾
10%, 90%, 100% CO2 injections
¾
3mL/min constant flow-rate
¾
6.89MPa constant back-pressure
¾
16°C constant temperature
¾
Brine contains dissolved CO2
¾
CO2 contains dissolved water
Core Characterization
¾
Porosity/permeability “map”
coarsened
¾
Relative permeability/capillary
pressure
curves matched to experimental
curves
TOUGH2 (Pruess, LBNL)
180px
60px
36px
Simulated CO2 Saturations
Constant Pc Produces Homogeneous CO2 Saturations
Porosity
Lab
Data
Homogeneous Simulations Variable Φ, k Simulations
10%
CO2
90%
CO2
100%
CO2
CO2 Saturation:0%
70%
Fitting Capillary Pressure Curve
100,000
Simulation Input Curve*
Hg Injection Data Curve
Pc (Pa)
Pc,i ∝
√
Φi
ki
Pcap =
4500Pa
10,000
2000
Pcap (Pa)
8000
given 20% CO2
1000
Brine Saturation
*Silin et al. (submitted, 2007
Simulated CO2 Saturations
Variable Pc Produces Small-scale CO2 Saturation Variations
Lab Data
Variable Φ, k Simulations
10%
CO2
90%
CO2
100%
CO2
CO2 Saturation:0%
70%
Variable Pc Simulations
Capillary Pressure Curve
Pcap (PA)
Avg. Pc Φ=.22 k=301mD
Pc Envelope
CO2 Saturation:
0%
70%
100% CO2
90% CO2
Brine Saturation
10% CO2
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
CO2 Saturation
100% CO2
Injection
90% CO2
Injection
10% CO2
Injection
0
2.0
4.0
Length Along Core (cm)
6.0
7.8
Effects of Flow Rate on CO2 Saturation
Relative Permeability
Fractional Flow (CO2 )
Injection Rate (mL/min)
90% CO2 Injection Simulation
CO2 Saturation
CO2 Saturation
Brine Saturation
Distance from Well (m)
Buckley-Leverett
9.0
3.0
0.3
0.065
0.03
0.001
mL/min
0
1
2
Length of Core (cm)
5
6
7
8
Capillary Pressure
(Pa)
Capillary Pressure Distribution at Different Flow Rates
Capillary Pressure Curves:
Average: ϕ=0.22 k=206mD
Upper Bound: ϕ=0.12 k=35.7mD
Lower Bound: ϕ=0.25 k=444mD
Brine Saturation
9 mL/min
3 mL/min
0.3 mL/min
0.065 mL/min
0.03 mL/min
0.01mL/min
90% CO2, 10% Brine Injection
Variable Simulation Resolutions
Grid Size: 0.6x0.6x3mm
Grid Count: 67,350
Grid Size: 1x1x3mm
Grid Count: 23,400
CO2 Saturation:
0%
30%
55%
Grid Size: 2x2x3mm
Grid Count: 5,400
CO2 Saturation
Finer Simulation Grids Decrease CO2 Saturation
Length Along Core (cm)
Conclusions
•
Core-scale multi-phase flow experiments reveal strong influence of
sub core-scale heterogeneity
•
Spatial variations in capillary pressure behavior control CO2
saturations
•
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