In this work, numerical modeling was used to investigate how fluid injection influences the statistical properties of injection-induced earthquake sequences. The numerical model calculated the coupled processes of fluid flow in aquifers and faults, mechanical deformation, and quasidynamic earthquake rupture. Injection-induced earthquake sequences were simulated by modeling fluid injection into a porous aquifer that contained many faults, which were represented explicitly within an embedded fracture modeling framework. The earthquake rupture process was modeled using a rate-and-state constitutive model of fault friction. The Oklahoma fault map database provided by Holland (2015) was used to develop distributions of two-dimensional fault structures for four synthetic study areas. A sensitivity study was performed to characterize the relationship between the statistical properties of earthquake sequences and both geological and operational model parameters. Sensitivity studies were performed to assess the effect of both dynamic friction and variability of injection rate over time on statistics of simulated earthquake sequences. It was found that at lower dynamic friction, earthquake nucleation occurred with less injected volume, the maximum magnitude increased, and total number of earthquakes increased. The effects of a variable injection operational schedule were investigated by comparing constant-rate injection to a scenario where the well cycled between periods of injection and shut-in on a monthly basis. This study was performed using fixed pressure and no-flow boundary conditions for each of the four study areas. For the fixed pressure boundary cases there was an increase in seismicity when fluid was cycling monthly, likely due to higher pressures that occurred near the pumping site. For the no-flow boundary condition, the earthquake sequences were nearly identical in terms of the earthquake magnitude distribution, event timing, and total number of events. The Oklahoma Corporation Commission has placed mandates on disposal wells that are thought to be linked to significant earthquake events (Baker, 2016). However, there still is not a clear procedure for how to most efficiently manage injection well operation in order to reduce the seismic hazard. Therefore, the use of a numerical modeling system is fundamental in order to create a better understanding of induced seismicity and help shape impactful regulation.

Press the Back button in your browser, or search again.

Copyright 2017, Stanford Geothermal Program: Readers who download papers from this site should honor the copyright of the original authors and may not copy or distribute the work further without the permission of the original publisher.