Waveform-based Microseismic Location Using Stochastic Optimization Algorithms: A Parameter Tuning Workflow

Lei Li,Jingqiang Tan,Yujiang Xie,Yuyang Tan,Jan Walda,Zhengguang Zhao,Dirk Gajewski

DOI: https://doi.org/10.1016/j.cageo.2019.01.002

IF: 5.168

2019-01-01

Computers & Geosciences

Abstract:A fast and accurate source location estimation is the foundation for passive seismic processing and interpretation. Waveform-based location methods become more and more popular for analysis of both natural and induced seismicity. We utilize stochastic optimization algorithms to speed up microseismic location. Two waveform-based location methods (i.e. diffraction stacking and cross correlation stacking) are adopted to test the performance of three algorithms (i.e. particle swarm optimization, differential evolution, and neighbourhood algorithm). In order to enhance the algorithmic performance, we propose a parameter tuning workflow which consists of two types of repeated tests. One type is multiple independent tests for a single event and the other involves tests of multiple events. The success rate, speedup, location uncertainty and bias are investigated to assess the algorithmic performances. We apply the workflow to a field dataset of mining induced seismicity and obtain preferential algorithm(s) with optimized ranges of control parameters. Synthetic tests are also conducted to demonstrated the feasibility of the proposed parameter tuning workflow. Given the two imaging operators, differential evolution is demonstrated to be the preferential one accounting for both algorithmic robustness and efficiency. Meanwhile, the workflow also examines the characteristics of different imaging operators. Cross correlation stacking proves to be simpler and more robust than its counterpart. Though the workflow is developed for microseismic location, it can also be adapted for other seismic inversion problems (e.g., source mechanism inversion) and ensure the algorithmic robustness and efficiency.