Microseismic monitoring plays an important role in many energy-related and environmental industries. The microseismic event catalog and seismic structure of subsurface are two of the primary outputs of the microseismic monitoring system. Though rough locations of microseismic events can be estimated automatically, obtaining high-resolution microseismic event locations requires a significant amount of human labor especially on seismic phase picking. Unlike traditional automatic pickers that are usually less precise than human analysts, a few recently proposed algorithms based on deep neural networks (DNN) were able to match or surpass human performance for earthquake signals. Due to differences in the spatial scale of the study area, sensor sampling rate, and geometry of the monitoring system, it is not clear whether these deep neural network models can be used to speed up microseismic data processing. In this paper, we adapted the DNN based technique for automatic phase picking of microseismic signals. We used microseismic data recorded at the experiment 1 site of the enhanced geothermal system (EGS) Collab project and designed a workflow that we call transfer-learning aided double-difference tomography (TADT), that combines transfer learning and seismic tomography. We re-train an existing DNN with our data to obtain a new model using around 2400 seismograms and associated manual phase picks. This transfer learned model is able to reach human performance but much faster than human analysts. The transfer-learning-derived phase picks were used to improve microseismic event locations and image the subsurface. The results are similar to or slightly better than those obtained with manual phase picks.

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

Copyright 2020, 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.