Application of a convolutional neural network for seismic phase picking of mining-induced seismicity

Abstract

Summary Monitoring mining-induced seismicity (MIS) can help engineers understand the rock mass response to resource extraction. With a thorough understanding of ongoing geomechanical processes, engineers can operate mines, especially those mines with the propensity for rockbursting, more safely and efficiently. Unfortunately, processing MIS data usually requires significant effort from human analysts, which can result in substantial costs and time commitments. The problem is exacerbated for operations that produce copious amounts of MIS, such as mines with high-stress and/or extraction ratios. Recently, deep learning methods have shown the ability to significantly improve the quality of automated arrival-time picking on earthquake data recorded by regional seismic networks. However, relatively little has been published on applying these techniques to MIS. In this study, we compare the performance of a convolutional neural network (CNN) originally trained to pick arrival times on the Southern California Seismic Network (SCSN) to that of human analysts on coal-mine-related MIS. We perform comparisons on several coal-related MIS datasets recorded at various network scales, sampling rates, and mines. We find that the Southern-California-trained CNN does not perform well on any of our datasets without retraining. However, applying the concept of transfer learning, we retrain the SCSN model with relatively little MIS data after which the CNN performs nearly as well as a human analyst. When retrained with data from a single analyst, the analyst-CNN pick time residual variance is lower than the variance observed between human analysts. We also compare the retrained CNN to a simpler, optimized picking algorithm, which falls short of the CNN's performance. We conclude that CNNs can achieve a significant improvement in automated phase picking although some dataset-specific training will usually be required. Moreover, initializing training with weights found from other, even very different, datasets can greatly reduce the amount of training data required to achieve a given performance threshold.

Transfer learning: improving neural network based prediction of earthquake ground shaking for an area with insufficient training data

“…There has been some application of TL to seismology (e.g., Chai et al 2020, Titos et al 2019, Johnson et al 2021, Otović et al 2021) although its use is not yet widespread for applications of ML in seismology. In this study, we explore the use of TL for improving our IM prediction algorithm.…”

Study on the Mechanism and Response Law of Fracture Movement on the Super-High Position Hard-thick Strata

“…(3) After the initial breakage of the hard-thick strata, periodic breakage of the hard-thick strata will occur as the 221 upper 03 working face continues to advance. According to the calculation of "cantilever beam" structure [27][28] , the periodic breakage distance of the hard-thick strata is 25 m. In other words, the periodic breakage of the hard-thick strata will occur when the working face advances at least 138 m, and the cycle will continue until the end of the working face, as shown in Figure 5.…”

Impact of Dynamic Soil-Structure Interaction on Performance of a Single Span Footbridge with Overhangs Subjected to Mining-Induced Shocks

“…Mining-triggered tremors exhibit characteristic features that distinguish them from natural earthquakes. The differences relate mainly to the mechanism of shock generation, the wave propagation from the source to the receiver, the range of dominant frequencies, duration, and repeatability of occurrence [18][19][20]. In a typical record of mining-triggered vibrations [21], in close proximity to the source, amplitudes of vertical oscillations are comparable to or even higher than amplitudes of horizontal vibrations.…”