Characterization of ground oscillations induced by underground mining

Abstract

We examine ground acceleration during M1.5 and M2.0 seismic events induced by underground mining at Upper Silesian coal basin and Legnica Glogow copper mine, respectively, using methods of nonlinear time series analysis, in order to confirm its stochastic nature. Recorded time series are firstly embedded into the adequate phase space using the mutual information and box-assisted methods. After this, we performed stationarity test, by which we confirmed that the examined ground acceleration belongs to a group of stationary processes. Surrogate data testing is applied then, which resulted in following: (1) horizontal ground acceleration at Legnica Glogow copper mine represents stationary linear stochastic processes with Gaussian inputs, (2) ground acceleration at Upper Silesian coal basin originates from a stationary Gaussian linear process that has been distorted by a monotonic, instantaneous, time-independent non-linear function, (3) vertical ground acceleration at Legnica Glogow copper mine could not be ascribed to any of the examined processes, probably due to high level of instrumental or background noise. Low values of determinism coefficient (c≤0.7), negative values of maximum Lyapunov exponent and quick saturation of neighboring points distance with the increase of embedding dimension indicate the absence of determinism in the observed ground acceleration time series.