Experimental study on the characteristics of microseismic signals generated during granite rockburst events

Rockburst is a typical catastrophic phenomenon in deep underground excavation. Microseismic (MS) monitoring is an extremely significant technique for rockburst prediction. The effectiveness of rockburst warning using MS depends on the cognition of characteristics of the MS during the rockburst process, especially the MS precursor information before rockbursts. In the present study, the rockburst process in granite is reproduced in the laboratory using a true triaxial rockburst testing system. Some notable characteristics of the MS in terms of its quiescent period, waveform fractal dimension, activity level, b value, lgN/b, and frequency spectra were obtained. These MS characteristics are consistent with that obtained on-site, which indicates that the laboratory investigation of rockbursts using MS signals is feasible. In addition, a systematic comparison between MS and AE signals during rockbursts was performed. The obtained results indicate that there are essential differences between MS and AE signals in precursors of the rockburst. The quiescent period of the AE hits was much longer than that of the MS amplitude immediately preceding the rockburst. Besides, the change rate of the time-dependent fractal dimensions of the MS was much higher than that of the AE before a rockburst occurred. Furthermore, there was also a quiescent period in the evolution of the signal activeness index of the AE. However, the Microseismic activity index did not show an obvious quiescent period. On the eve of the rockburst, there were high-amplitude signals with low frequency appearing continuously in both kinds of signals, while the frequency range of MS signals is extremely lower than that of the AE. These different precursors enlighten us that the accuracy of rockbursts prediction can be effectively increased by integrated usage of MS and AE signals.