72 / 2021-09-16 08:45:42

Methodology to simulate triggered and induced seismic activities on a fault plane

triggered seismicity,induced seismicity,fault damage zone,crack tensor theory

Seismicity taking place in deep underground could pose serious issues for safety and production, in various engineering projects. Recently, it has increasingly become recognized that anthropogenic seismic activity is divided into two categories: triggered and induced seismicity. The former is the seismic activity caused by an infinitesimal stress change in the rock mass that had been intensely stressed prior to the onset of anthropogenic activity, whilst the latter is caused by a sufficient amount of stress change induced by human activity when compared to the ambient stress state. Although such seismic activities are recognized and observed in the field, knowledge has not yet been adequately accumulated to quantify this seismicity, owing to the difficulty in analytically and numerically investigating the complex, heterogeneous in-situ stress state. In fact, most of previous studies on the numerical simulation of anthropogenic seismic activity assume a homogenous stress state in the horizontal direction as an initial condition for the sake of simplicity, thus rendering the simulation of triggered seismicity and its quantitative analysis on seismic source parameters impossible.



To overcome this limitation, the present study aims to develop a methodology for the numerical simulation of triggered and induced seismicity by reproducing a heterogeneous stress state in a fault damage zone. Specifically, a fracture network in a fault damage zone is explicitly created as a first step with a stochastic approach as a discrete fracture network. Then, equivalent compliance tensor considering the elasticity of the fractures is computed with the crack tensor theory (Oda, 1986) for each zone in a three-dimensional numerical model constructed in the framework of finite difference method, i.e., the elastic matrix is anisotropic and spatially differs in the model. In addition, a fault core is explicitly modelled with interface elements at the center of the model. Subsequently, a numerical simulation is performed to simulate an initial stress state at a depth of 2,000 m whilst applying stresses on the model boundary. This simulation method generates a complex stress state in the fault zone whilst generating intensely stressed regions not only in the damage zone but also on the fault core. Finally, the effective stress on the fault core is gradually decreased in a stepwise manner at an interval of 0.5 MPa in a circular region with a diameter of 25 m whilst assuming the reduction of confining stress acting on the fault due to fluid injection or rock mass excavation.



The analysis result demonstrates that a slight decrease in the confining stress acting on the fault plane from the initial state gives a rise to spatially distributed seismic events, indicating the occurrence of triggered seismicity, as shown in Figure 1(a). Then, the number of slip patches increases as the confining stress further decreases. Eventually, the slip patches coalesce to form a single seismic even with the large magnitude of shear displacement. This result gives an insight into a quantitative estimation of seismic source parameters of triggered seismicity and sheds light on the prediction of intense seismic events caused by pre-existing stress anomalies.