Stability evaluation of fault in hydrocarbon reservoir-based underground gas storage: A case study of W gas storage

Fault stability evaluation is crucial for ensuring safety of underground gas storage (UGS) reconstructed from depleted oil and gas reservoirs. To make assessment result more reasonable, a novel fault stability evaluation method is proposed, and four major faults in W gas storage are studied as a case to detailly describe this method. This method considers both shear rupture failure and frictional slip failure of fault based on Mohr-Coulomb criterion and Amonton's law. In addition, dynamic variation in rock shear strength parameters under cyclic stress is understood and taken into account in fault stability evaluation by rock mechanics experiments. Meanwhile, specificity of fault friction coefficient in different UGSs is also considered in this method, and fault friction coefficient of W gas storage is determined by an empirical model. Fault stability evaluation results are quite disparate when different fault failure types are considered. For W gas storage, max operation pressure leading to fault failure is 62.10 MPa when shear rupture is regarded as fault failure type, and it reduce to 35.91 MPa when frictional slip is considered as fault failure type. Therefore, it is necessary to consider both two failure types, and further determine the key failure type in fault stability evaluation. Rock shear strength parameters vary with stress cycle, but the variation range is not significant in W gas storage. Rock cohesion shows a tendency that it fast declines first with stress cycles, then slow declines, and eventually tends to a constant. And variation in internal friction angle is contrary. Above both parameters orderly decrease by 3.08% and 8.75% after 53 cyclic stress cycles. The specificity of fault friction coefficient cannot be neglected in different UGSs. Otherwise, it could result in significant misestimate on fault stability of UGS. For W gas storage, fault friction coefficient is deter-mined of 0.71 by empirical model, which is significant disparate to previous works. This work proposes an effective method which can help scholars and engineers to accurately evaluate fault stability in UGS.