A displacement-dependent moment tensor method for simulating fault-slip induced seismicity

Fault-slip induced by underground longwall mining could trigger dynamic failure to nearby mine openings. A better understanding of fault-slip behavior and the associated seismicity is significant for estimating the triggered damage and finding a feasible way to predict potential larger events. In this study, a displacement-dependent moment tensor method is introduced and embedded in the FLAC3D code to simulate both small-scale direct shear test and field-scale mining-induced fault slip. For the direct shear test model, the numerical results agree well with reported experiments in respect to mechanical behavior and seismicity characteristics, such as moment magnitude, frequency-magnitude distribution (FMD), and variation of b-value. Numerical results show that a clustering process is essential to produce realistic FMD, and associated parameters are given. Material heterogeneity of the fault influences the event magnitude. Based on a case study, a field-scale model is built to investigate the fault-slip behavior as a longwall face approaches a fault. For individual contacts, ΔCFF (Coulomb Failure Function) is a useful indicator to evaluate the likelihood of slip. It gradually increases, followed by a quick rise before slip occurs, but sharply drops after slipping. Numerical simulation shows that shear slip initializes on the roof segment of the fault and gradually migrates to the seam and floor segments. Seismicity starts on the roof segment when the longwall face is 100 m away from the fault, but it increases strongly when the distance is less than 50 m and gradually moves toward the seam and floor segments, which is consistent with field observations. Simulated magnitude of the events and the b-value (0.8) also show reasonable agreement with field measurements. The numerical simulation also indicates that a significant drop in b-value usually precedes larger events triggered in the following advancement of the longwall face, which can be regarded as a precursor for larger events.