Microseismic monitoring involves the geophones placement within the rock mass of a mine to record its micro seismic activity. Prior to reducing the number of records required for the post processing and proving the hypocenter and automatic location, a microseismic waveforms classification technique using Hidden Markov Model has been studied in this article. Four different kind of microseismic signals were considered as most representative of coal mines in our study, including blasting, random noise, electromagnetic interference and rock fracture signals. Using the MATLAB module, feature extraction process was applied to obtain the MFCC characteristic of each kind of signal. The features were performed to reduce the number of them using an optimization process, and normalized to generate a feature vector at last. Then a Hidden Markov Model was implemented as a classifier to develop mining microseismic waveforms automatic classification system. And the EM method was applied to optimize the initial parameters. Our method was tested and verified on a dataset from Shandong province, provided by Longgu coal mine. The result indicates that the event classification method presented in this paper was precise and automatic. The described methodology can be used to classify more seismic signals to improve the study of the activity of this coal mine or to extend the study to other active mines.


Classification of microseismic waveforms is the important basis to improve the rate and accuracy about the microseismic location. Through identify the effective mining microseismic events accurately and effectively, we can eliminate a lot of useless background noise and mechanical vibration interference waveforms, and transmission signal interference in the process. This is the important foundation for the automatic location.

This content is only available via PDF.
You can access this article if you purchase or spend a download.