Mechanisms of hydraulic fracturing induced seismicity

点击翻译按钮获取中文摘要

Hydraulic fracturing on a massive scale has become ubiquitous in allowing to economically tap vast tight oil and gas reserves in North America. Microseismicity, which usually accompanies growth of hydraulic fractures, is a manifestation of dynamic slippage on pre-existing fractures and faults in the stimulated reservoir rock volume. This slip may lead to dilation of sheared fractures, and associated increase of permeability and connectivity of the reservoir rock to the hydraulic fracture (HF), and, therefore, improvement to eventual production. On rare occasions, dynamic slip induced by hydraulic fracturing is known to grow out into a small-to-moderate size earthquake. This has played a role in some countries' decision to ban fracturing. It is therefore important to study mechanisms of both dynamic and quasi-static fault slip due to stress and pore pressure perturbations introduced by a propagating HF to be able to better understand a) how it impacts reservoir properties and "stimulated rock volume"; b) the transition from aseismic to dynamic slip, and potential for dynamic rupture run-away (an earthquake). We will model slip on pre-existing fractures which lie in or near the path of a propagating HF. Slip is induced by one of the following mechanisms: a) stress perturbation at the front of the HF as it propagates past a pre-existing fracture; and b) pressurization of pre-existing fractures intersected by the hydraulic fracture. Due to a viscous fluid pressure drop along the propagating HF, we expect the two slip-­inducement mechanisms to be separated in time and space, with the former taking place near the advancing HF front, while the latter at some distance behind the front where the pressure in the hydraulic fracture has recovered to the levels sufficient to diffusively pressurize intersected shear fractures. Existing studies of earthquake instability - the transition of fault slip from initially slow rate driven by tectonic loading to seismic rates - are based on the assumption of unstable weakening of the fault gouge friction with slip or slip-rate. Since induced seismicity often takes place in provinces devoid of natural seismicity, it is possible that the earthquake nucleation mechanism there differs from the frictional instability. In the proposed work, we will model earthquake nucleation using a) traditional frictional instability, and b) a new model in which a fault slip transient activated by a pore pressure perturbation and associated frictional heating, leads to thermal pressurization of pore fluid driving the fault to instability. We will compare conditions for nucleation and run-­out distances of dynamic slip in the two models, as a function of the HF attributes, in situ stress, and orientation of pre-existing fractures and their proximity to the hydraulic fracture plane; and corroborate the results using existing laboratory and seismological observables.