Solution of Benchmark Problems for CO2 Storage

1
Solution of Benchmark Problems for CO2 Storage

• Min Jin, Gillian Pickup and Eric Mackay
• Heriot-Watt University
• Institute of Petroleum Engineering

2
Outline

• Introduction
• Problem 1
• Leakage through an abandoned well
• Problem 2
• Enhanced methane recovery
• Problem 3
• Storage capacity in a geological formation
• Conclusions

3
Numerical Simulation

• Simulation is a very important tool for CO2
  storage
• Can give estimates of
• migration of CO2 gas
• dissolution in brine
• build-up of pressure around injection well
• etc

4
Reliability

• Depends on
• Input data
• geological structure
• rock permeability/porosity measurements
• laboratory measurements
• Also depends
• Adequate computer models
• flow equations
• representation of physical processes

5
Reservoir Simulation

• Codes are complex
• Various different versions available for
• gridding model
• calculating fluid properties
• solving equations
• May get slightly different answers

6
Benchmark Problems

• Compare solutions using different codes
• If results are the same
• gives confidence in simulation results
• If they are different
• indicates where more work is needed

7
Stuttgart Workshop, April 2008

• Aim
• Discuss current capabilities of mathematical and
  numerical models for CO2 storage
• Compare results of 3 benchmark problems
• Focus model development on open questions and
  challenges
• 12 groups participating

web site http//www.iws.uni-stuttgart.de/co2-work
shop/
8
Heriot-Watt Entry

• Solutions to all 3 problems
• Eclipse 300
• Reservoir simulation software package
• Compositional simulation
• Schlumberger

9
Outline

• Introduction
• Problem 1
• Leakage through an abandoned well
• Problem 2
• Enhanced methane recovery
• Problem 3
• Storage capacity in a geological formation
• Conclusions

10
Problem 1

• CO2 plume evolution and leakage through an
  abandoned well

leaky well
k 200 mD, f 0.15
aquifer
k 0 mD, f 0.0
aquitard
k 200 mD, f 0.15
aquifer
1000 m
11
Problem 1

• CO2 plume evolution and leakage through an
  abandoned well

leaky well
CO2 injector
aquifer
aquitard
aquifer
12
Problem 1

• CO2 plume evolution and leakage through an
  abandoned well

leaky well
?
CO2 injector
aquifer
aquitard
aquifer
13
Model Details

• Lateral extent of model 1000 m x 1000 m
• Separation of wells 100 m
• Aquifer thickness 30 m
• perm 200 mD, poro 0.15
• Aquitard thickness 100 m
• impermeable
• Abandoned well
• model as thin column of 1000 mD, poro 0.15

14
Details of Fluid Properties

• Problem 1.1
• Reservoir is very deep, 3000 m
• Simplified fluid properties
• constant with P and T
• Problem 1.2
• Shallower reservoir, lt800 m
• CO2 can change state when rising
• More complex fluid properties

15
Other Inputs to Simulation

• Constant injection rate
• 8.87 kg/s
• Pressure should stay constant at the edges of the
  model
• No-flow boundaries top and bottom

16
Challenges

• Gridding
• Coarse over most of model
• Fine near wells

17
Close-up of Grid Centre
leaky well
injector
18
Challenges

• Modelling of abandoned well
• Model as high perm column
• Model as closed well
• output potential production

high perm cells
closed well
19
Challenges

• Maintaining pressure constant at boundaries
• Eclipse designed for oil reservoirs
• assumes sealed boundaries
• leads to build up of pressure
• We added aquifers to sides of the model
• fluids could move into the aquifer
• prevented build up of pressure

20
Challenges

• Fluid properties in Problem 1.2
• User-defined
• Specified as functions of pressure and
  temperature
• We used constant T 34 oC
• Tuned equations
• density and pressure similar to specified values

21
CO2 Distribution after 100 Days, Problem 1.2
Injector
Leaky well
Gas Sat
0.0
0.2
0.4
0.6
0.8
22
CO2 Distribution after 2000 Days, Problem 1.2
Inj
leaky well
23
Results

• Leakage rate for Problem 1.2

leaky well modelled as high perm cells
24
Summary of Problem 1

• Successfully predicted well rate
• Using high perm cells for leaky well
• well model overestimated leakage
• Our results similar to others
• Leakage rate 0.1 injected volume

25
Outline

• Introduction
• Problem 1
• Leakage through an abandoned well
• Problem 2
• Enhanced methane recovery
• Problem 3
• Storage capacity in a geological formation
• Conclusions

26
Problem 2

• Enhanced recovery of CH4 combined with CO2 storage

27
Model Details

• Two versions
• homogeneous
• layered
• Temperature 66.7 oC
• Depleted reservoir pressure 35.5 bar
• Molecular diffusion 6 x 10-7 m2/s

28
Model for Problem 2.2
29
Other Inputs to Simulation

• Constant injection rate for CO2
• 0.1 kg/s
• inject into lower layer
• produce from upper layer
• Constant pressure at production well
• P 35.5 bar
• No-flow across model boundaries

30
Challenges

• Mixing of gases
• Changes in physical properties of gas mixture
  with composition
• can be modelled in Eclipse 300
• Numerical diffusion
• will artificially increase the molecular diffusion

31
Result for Problem 2-1
32
Results Homogeneous Model

• Mass Flux of CH4 and CO2

33
Results Layered Model

• Mass Flux of CH4 and CO2

34
Results and Summary

• Assume well is shut down when CO2 production
  reaches 20 by mass
• Relatively easy problem

Problem Model Shut-in time (days) Recovery Efficiency ()
2.1 homogeneous 1727 59
2.2 layered 1843 64
35
Outline

• Introduction
• Problem 1
• Leakage through an abandoned well
• Problem 2
• Enhanced methane recovery
• Problem 3
• Storage capacity in a geological formation
• Conclusions

36
Problem 3

• Storage capacity in a geological model

37
Model Details

• Lateral dimensions
• 9600 m x 8900 m
• Formation thickness
• between 90 and 140 m
• Variable porosity and permeability
• Depth 3000 m
• Temperature 100 oC

38
Challenges

• Simulation of system after injection has ceased
• CO2 continues to rise due to buoyancy
• Brine moves into regions previously occupied by
  CO2
• Brine can occupy small pores, trapping CO2 in
  larger pores
• additional trapping mechanism
• hysteresis

39
Challenges

• Trapping of CO2 by hysteresis

after Doughty, 2007
40
CO2 Distribution after 25 Years
X
with hysteresis
fault
Y
Gas Sat
0.0
0.2
0.5
0.8
41
CO2 Distribution after 50 Years
X
with hysteresis
fault
Y
Gas Sat
0.0
0.2
0.5
0.8
42
Results

• Mass of CO2 in formation over time

(kg)
43
Results

• Leakage of CO2 across the boundaries

no hysteresis
with hysteresis
44
Summary of Problem 3

• CO2 did not move towards the fault
• moved up-dip
• leaked across model boundary
• Hysteresis did make difference, but not much
  difference in this example
• About 0.2 of the injected CO2 dissolved after 50
  years

45
Outline

• Introduction
• Problem 1
• Leakage through an abandoned well
• Problem 2
• Enhanced methane recovery
• Problem 3
• Storage capacity in a geological formation
• Conclusions

46
Conclusions

• Benchmark solutions highlight difficulties
• Adaptation of simulator for oil/gas reservoirs to
  CO2 storage
• Difficulties are surmountable
• Schlumberger created new module for CO2 storage
• Participation in the workshop
• Giving us confidence in simulations

47
Acknowledgements

• We thank Schlumberger for letting us use the
  Eclipse simulation software

48
Solution of Benchmark Problems for CO2 Storage

• Min Jin, Gillian Pickup and Eric Mackay
• Heriot-Watt University
• Institute of Petroleum Engineering