Matrix Shrinkage and Permeability Reduction with Carbon Dioxide Injection

Transcription

1 Matrix Shrinkage and Permeability Reduction with Carbon Dioxide Injection Coal-Seq II Forum Presented by: Lawrence J. Pekot Advanced Resources International, Inc. Washington D.C. March 6-7,

2 Presentation Outline Background Comparison of Matrix Shrinkage Models Model Equivalence Routine CBM Applications Undersaturated Reservoirs Swelling Not Proportional to Concentration Differential Swelling Field Evidence Conclusions 2

3 Background Coal matrix shrinkage occurs in CBM reservoirs as production causes pressure decline and reduction in in-situ gas content. Matrix shrinkage causes cleat porosity and permeability to increase during depletion. Matrix shrinkage effects are observable in more mature CBM areas such as Warrior Basin and San Juan Basin. Matrix swelling causes porosity and permeability to decline during injection/repressuring with CO 2. 3

4 Background Coal porosity and permeability are related by: k φ = k i ( ) m φ i where m usually equals 3 4

5 Background In a depletion scenario, matrix shrinkage is a late-time phenomenon. Shrinkage is related to gas concentration and more rapid change in concentration occurs after reservoir pressure is reduced, increasing permeability Small Rate of Concentration Change Concentration, SCF/T Large Rate of Conc. Change Low Pressure, Late Time High Pressure, Early Time Pressure, psia 5

6 Background In contrast, most pore volume compressibility occurs in early time. Compressibility is related to pore pressure change and more rapid change occurs at higher reservoir pressure High Rate of Pressure Change 500 PRESSURE Early Time, High Pressure Late Time, Low Pressure Low Rate of Press. Change YEARS 6

7 Background To summarize: Changes in coal porosity and permeability are the net effect of matrix shrinkage and pore volume compressibility, which act in opposite directions. Compressibility dominates in early time, shrinkage dominates in late time. 7

8 Background Q: Why is shrinkage and swelling important for carbon sequestration? A: Injection of CO 2 into coalbeds causes large changes in pressure, in-situ gas concentration and gas type therefore it causes large changes in reservoir permeability. Technical evaluations and economic decisions rely on our ability to model these changes. Our understanding of this behavior remains limited. 8

9 Presentation Outline Background Comparison of Matrix Shrinkage Models Model Equivalence Routine CBM Applications Undersaturated Reservoirs Swelling Not Proportional to Concentration Differential Swelling Field Evidence Conclusions 9

10 Comparison of Matrix Shrinkage Models Two models describe shrinkage and swelling. ARI, Sawyer, et. al. CIM/SPE , 1990 P φ = φ i [1 + c p (P-P i )] c m (1 - φ i ) ( i ) (C-C i ) C i φ Palmer and Mansoori, SPE 36937, 1996 A m ε L K β P β P = φ i [1 + φ (P - P i i )] + ( M - 1) ( - i ) 1+β P 1+β P i 10

11 Comparison of Matrix Shrinkage Models The similarities: Both models use the same relationship between φ porosity and permeability k = k i ( ) m Both models describe pore volume compressibility the same way. C p φ (P-P i ) φ i 11

12 Comparison of Matrix Shrinkage Models The differences: The ARI model describes shrinkage in terms of matrix shrinkage compressibility and gas concentration change -c m (1 - φ i ) P i C i The P & M model describes shrinkage in terms of rock mechanics moduli and a Langmuir strain function 1 1+ υ P [ ( ) -1] ε ( L + P ) 3 1- υ m V L P 12

13 Presentation Outline Background Comparison of Matrix Shrinkage Models Model Equivalence Routine CBM Applications Undersaturated Reservoirs Swelling Not Proportional to Concentration Differential Swelling Field Evidence Conclusions 13

14 Model Equivalence Routine CBM Applications In routine CBM applications the ARI and P&M models yield equivalent results 14

15 Model Equivalence Variation in coal permeability with pressure. Results of Palmer and Mansoori model. Reproduced from SPE

16 Model Equivalence Perm ratio, k/ko Cp = 806 E-06 1/psi Cp = 225 E-06 1/psi Cp = 4032 E-06 1/psi Cp = 1124 E-06 1/psi Initial Conditions Pressure, psi Variation of coal permeability ratio with pressure. Results of ARI model. 16

17 Model Equivalence Undersaturated Coal Reservoirs The ARI model is based on gas concentration change. If no gas desorption occurs, no shrinkage occurs. The P & M model is based on a change in the Langmuir strain relationship and appears to allow matrix shrinkage even when no gas is desorbed. 17

18 Model Equivalence Matrix shrinkage affects P & M results even in undersaturated conditions Porosity, fraction Saturated Undersaturated Pressure, psi P&M model ARI model Variation in coal porosity with pressure. Undersaturated reservoir. ARI and P&M models, initial pressure 1100 psi 18

19 Model Equivalence Swelling not proportional to gas concentration The ARI model assumes swelling is proportional to gas concentration The P & M model uses a separate Langmuir strain function to model swelling If swelling does not follow a Langmuir function, both models would be inaccurate 19

20 Model Equivalence Strain (swelling) proportional to CH 4 concentration is a reasonable assumption Strain Concentration Volumetric Strain, in/in Concentration, SCF.Ton Pressure, psia Volumetric strain and methane concentration vs. pressure. Data of Levine, Ref. 5. Strain axis range 0.0 to EL (0.0101); Concentration axis range 0.0 to VL (1053). 20

21 Presentation Outline Background Comparison of Matrix Shrinkage Models Model Equivalence Routine CBM Applications Undersaturated Reservoirs Swelling Not Proportional to Concentration Differential Swelling Field Evidence Conclusions 21

22 Differential Swelling CO 2 causes more swelling than CH 4 at a given pressure due to higher CO 2 concentration. CO 2 also appears to cause more swelling than CH 4 at a given concentration level, i.e. the degree of swelling is dependent on the type of gas, not just gas concentration. This departure from CH 4 swelling behavior is termed Differential Swelling. 22

23 Differential Swelling Methane Carbon Dioxide Methane, Extrapolated CO2, Extrapolated Volumetric Strain, in/in Concentration, SCF/Ton Volumetric strain vs. gas concentration for methane and carbon dioxide. Data replotted from Levine. 23

24 Differential Swelling Previously described swelling models did not account for this behavior ARI model has been extended to account for differential swelling φ = φ i [ 1 - C p (p - p i ) c m (1-φ i ) ( P i C i ) [(C-C i ) + c k (C t -C)] 24

25 Differential Swelling Methane Carbon Dioxide Methane, Extrapolated CO2, Extrapolated ARI Model Volumetric Strain, in/in Concentration, SCF/Ton Volumetric strain vs. gas concentration for methane and carbon dioxide. Data replotted from Levine. ARI extended model for CO2 also shown using ck=

26 Differential Swelling Perm ratio, k/ko Pressure, psi Methane CO2, no Diff. Swelling CO2 with Diff. Swelling Variation in ratio of coal permeability with pressure. Effect of CO2 with and without differential swelling. 26

27 Presentation Outline Background Comparison of Matrix Shrinkage Models Model Equivalence Routine CBM Applications Undersaturated Reservoirs Swelling Not Proportional to Concentration Differential Swelling Field Evidence Conclusions 27

28 Field Evidence of Shrinkage and Swelling Burlington Resources Allison Unit ECBM pilot in the San Juan Basin provides evidence Well tests show over 90% loss of permeability in CO 2 injection wells Early loss of injectivity in CO 2 injectors Long term improvement in injectivity due to pressure decline and shrinkage 28

29 Field Evidence of Shrinkage and Swelling Rate Pressure Jan-95 Apr-95 Jul-95 Oct-95 Jan-96 Apr-96 Jul-96 Oct-96 Jan-97 Apr-97 Jul-97 Oct-97 Jan-98 Apr-98 Date CO2, Mcf/mo BHP, psi Jul-98 Oct-98 Jan-99 Apr-99 Jul-99 Oct-99 Jan-00 Apr-00 Jul-00 Oct-00 Injection/Pressure History for CO 2 Injection Well, Allison Unit, San Juan Basin

30 Field Evidence of Shrinkage and Swelling 250 Permeability, md Continued Injection & Depletion Primary Depletion Displace w/ CO2 Start Pressure, psi Permeability History for CO 2 Injection Well 30

31 Presentation Outline Background Comparison of Matrix Shrinkage Models Model Equivalence Routine CBM Applications Undersaturated Reservoirs Swelling Not Proportional to Concentration Differential Swelling Field Evidence Conclusions 31

32 Conclusions CO 2 injection can cause large permeability changes in ECBM and sequestration projects. Matrix swelling and permeability change is affected not only by the concentration of gas but also by the type of gas. (Differential Swelling) Differential swelling is observed in limited laboratory data but is not well documented for CO 2 and not at all for other gasses. 32

33 Conclusions The ARI and P & M formulations provide equivalent results in modeling matrix shrinkage and swelling routine CBM applications. The P & M form is better where swelling is poorly related to gas concentration. The ARI form appears more accurate for undersaturated reservoirs. The ARI shrinkage and swelling model has been extended to account for differential swelling and replicates laboratory data. 33