Various analysis techniques are used to determine the nature of a large seismic event that was triggered by a mass blast in an underground mine. This event was recorded by surface accelerometer arrays whose output was doubly integrated using a stable filter to determine the dynamic displacements. The peak dynamic displacements during the blast were 0.09 mm, and during its latter stages, this blast induced a seismic event having a peak dynamic displacement more than ten times larger at 1.26 mm. Polarisation analysis, based on the complex covariance matrix, showed that the seismic event was caused by movement of a known underground fault, close to the blast, which radiated most of its release motion as a shear-horizontal (SH) wave. The change in frequency content with time for the blast and seismic event was initially determined using the Fourier-based reassigned S-method. This analysis showed that the large amplitude seismic event had frequency content significantly lower than that of the blast. Nevertheless, this Fourier technique still retains the problems associated with short time windows as well as the fact that most realistic signals are non-stationary. In order to eliminate such influences, the more recently developed empirical mode decomposition (EMD) was used to obtain the intrinsic mode functions (IMF) of the seismic wave propagation. The EMD vibration analysis showed that each IMF generally exhibited distinct frequency changes throughout the blast and a significant reduction in frequency throughout the subsequent event. It is shown that most of the energy of the seismic event is contained in IMF modes 2, 3 and 4 of the SH wave; the high-frequency first mode (IMF1) contains little energy. Such modal analysis is not possible using standard Fourier methods.
Blair, D P, 2015. Dynamic displacements from a blast-triggered seismic event, in Proceedings 11th International Symposium on Rock Fragmentation by Blasting, pp 203–210 (The Australasian Institute of Mining and Metallurgy: Melbourne).