An algorithm for locating microseismic events

Abstract

Monitoring seismic activity in mines, produced by high stress faults in the vicinity of the mining operations, is an important issue for mine safety. A seismic event produces a short-time duration acoustic pressure wave that travels through the rock. This low-energy seismic activity in mines is typically referred to as microseismic events. The location of a microseismic event can be estimated using the pressure wave signals recorded at a set of sensors distributed throughout the mine. The classical process for locating a radiating source involves two steps: an estimation of the time difference of arrival between all sensor pairs followed by the localization, requiring the solution of a set of non-linear equations. This traditional localization process has limited success when applied to microseismic events, since they have a short-time duration, and thus generating accurate arrival time estimates is a challenging task in a noisy environment. An alternate approach to traditional localization, that avoids time-delay estimation, is to search over a grid of hypothesized source locations to find the one that best explains the observed measurements. Here, this approach is used with a performance function that is the greatest energy calculated from the sum of the sensor signals, each of which is time-shifted by an amount consistent with the hypothesized location of the event. The result is a very robust algorithm that works well with short-time duration signals and given the recent advances in low-cost computational power can be implemented in real time. The paper describes the location algorithm. Results are presented for computer generated signals as well as actual signals produced by a microseismic event that occurred one kilometer below the surface in a potash mine near Saskatoon, Saskatchewan.