Institute of Geophysics, Academy of Sciences Prague, Czech Republic

1. Shear-tensile/implosion source model vs. momenttensor: benefit in single-azimuth monitoring Simulation of Cotton Valley configuration Jan Šílený Institute of Geophysics, Academy of Sciences Prague,Czech Republic

2. Outline: Motivation: hydrofracturing of oil/gas wells Inversion for MT: Cotton Valley MT summary: advantages, pitfalls MT alternative: STI (shear-tensile/implosion) reprocessing of Cotton Valley Conclusion

3. Hydrofracturing of oil/gas wells monitoring well monitoring well treatment well

4. Hydrofracturing of oil/gas wells Possibility to recover MT (6 components) far field: Vavryčuk (2007) 3 (or more) wells: OK from P, P+S rarely 2 wells: OK from P+S occasionally 1 well: Not enough ! standard • additional constraint needed Jechumtálová & Eisner(2008) • near-field needed Song& Toksoz(2011) • a simpler source model needed less parameters, e.g. shear +off-plane comp.

5. Source mechanism by fiting seismograms/amplitudes MT Troubles causing the uncertainty • noise in the waveforms • inexact hypocenter location improper Green’s function • inexact earth model • sparse station distribution

6. Micro-earthquakes induced by hydrofracturing Cotton Valley gas field hydrofractureexperiment Rutledge et al. 2004

7. Data examples: deviatoric vs. volumetric source P S DC 35% CLVD(P) 18% ISO(expl) 47% G1 DC 48% CLVD(T) 47% ISO(expl) 5% G2 P S R7 P S DC 28% CLVD(P) 11% ISO(impl) 61%

8. tensile fracturing successful treatment Cotton Valley mechanisms Šílený, J., Eisner, L., Hill, D. & Cornet, F., 2009. Non-double-couple mechanisms of microearthquakes induced by hydraulic fracturing. J. Geophys. Res., 114, doi:10.1029/2008JB005987.

9.  66% CLVD(T)  66% CLVD(P) + 33% ISO(impl) + 33% ISO(expl) Hydrofracturing of oil/gas wells Cotton Valley case study Far field P+S from 2 wells = even-determined inversion for MT No over-determination ! Viewpoint of fracture mechanics: MT is unnecessarily complex Question of particular interest: Was permeability of the reservoir increased? i.e, were tensile cracks created?  discriminate between shear and tensile modes of fracturing shear slip tensile crack crack closure  100% DC

10. traditional intuitive explanation (opening crack with fluid influx) is incorrect ! tensile crack / cavity closure 66% CLVD(P) 66% CLVD(T) + 33% ISO(impl) + 33% ISO(expl) physical sources of interest are combinations of ISO and CLVD Prospects: MT … general dipole source advantage:linearity an alternative BUTtoo general to the MT  includes unphysical sources Traditional decomposition ISO+DC+CLVD motiv: • split isotropic and deviatoric part ISO DC+CLVD BUTthere is something • single out DC in addition, namely CLVD an unphysical source benefits:  shear – tensile / implosion source model • physical source Dufumier & Rivera 1997 Vavrycuk 2001 • less parameters  advantage in inversion a + / disadvantage: • non-linearity off-plane (4 angles + magnitude) angle a shear slip + tensile crack / cavity closure

11. Shear-tensile/implosion source model • DC angles dip, strike, rake Parameters: • off-plane angle a • magnitude (scalar moment) Inversion method: 2-step grid search • coarse grid global search • fine grid local refinement advantageous mapping of model space: at ‘each’ point (in terms of the sampling) information available on the goodness of fit: chi-square possibility to construct a confidence region: subspace in the model space around estimated solution with a priori specified probability content

12. Tensile/implosion part DC part of the mechanism Projection of T,P,N axes Histogram of values of the off-plane angle a of the mechanisms of the mechanisms with NRMS < NRMSthreshold with NRMS < NRMSthreshold • estimate of uncertainty in the tensile/implosion part • estimate of uncertainty in the DC part orientation a positive  tensile fracturing negative  implosion Shear-tensile/implosion source model Plots of ‘confidence zones’

13. Synthetic testing of STI: two-well vs. single well monitoring Cotton Valley gas field hydrofractureexperiment Simulation of a single-well monitoring Source model: • vertical strike-slip • 450 dip-slip

14. vertical strike-slip off-plane angle shear slip tensile crack

15. 450 dip-slip off-plane angle shear slip tensile crack

16. MT: noise, velocity mismodeling, mismodeling+mislocation T Inversion of data from two monitoring wells dip-slip + tensile (50 off-plane) strike-slip + tensile (50 off-plane) T model model P T P T T P P P P T T mismodeling mismodeling + mislocation mismodeling mismodeling + mislocation noise noise 10% 30% 10% 30% T T T T P P P P model model mismodeling mismodeling mismodeling + mislocation mismodeling + mislocation

17. STI: velocity mismodeling T model model dip-slip + tensile (50 off-plane) strike-slip + tensile (50 off-plane) P T P T T two wells two wells T single well single well T P P P P weak mismodeling weak mismodeling T T T two wells two wells single well single well T P P P P strong mismodeling strong mismodeling 120,150,200% NRMS 120,150,200% NRMS

18. STI: velocity mismodeling + mislocation T model model dip-slip + tensile (50 off-plane) strike-slip + tensile (50 off-plane) P T P two wells two wells single well single well P P P P strong mismodeling strong mismodeling T T T T 120,150,200% NRMS 120,150,200% NRMS

19. STI: noise contamination 10%, 20% T model model dip-slip + tensile (50 off-plane) strike-slip + tensile (50 off-plane) P T P two wells two wells single well single well T T P P P P T noise 10% noise 10% T two wells two wells single well single well T T P P T P P T noise 20% noise 20%

20. STI: noise contamination 30%, 40% T model model dip-slip + tensile (50 off-plane) strike-slip + tensile (50 off-plane) P T P two wells two wells single well single well T T P P T noise 30% noise 30% P T P two wells two wells single well single well T T P P P T P T noise 40% noise 40%

21. Re-processing of Cotton Valley: G1 map view depth view MT STI

22. G1 composite T,N,P axes off-plane NRMS < 0.9 NRMS < 0.8 NRMS < 0.7

23. R5 map view depth view MT STI

24. G2 composite map view depth view MT STI

25. G3 composite map view depth view MT STI

26. G4 composite MT STI map view depth view

27. Conclusions: MT: general description of a dipole source but too general in case of a simple fracturing: hydrofracturing in oil/geothermal industry: opening/closing of tensile cracks additionalconstraint helpful  STI model less parametersthan MT (5 vs. 6) non-linear model  exploration  advantage in estimate of model space in estimate of uncertainty (e.g., grid search) of the solution beneficial in deficient configurations in particular, single-well monitoring robust even with: • reasonable noise in data • realistic mislocation • slight velocity mismodeling

28. Acknowledgements: Institute of Geophysics, Academy of Sciences, Prague Czech Republic EC FP7 EC FP7 IAPP Project AIM “Advanced Industrial Microseismic Monitoring” Grant agreement No.230669 Grant Agency of the Czech Republic Project “Non-double-couple mechanisms – – a tool for monitoring the mode of fracturing” Grant agreement No.205/09/0724