An optimizing microseismic method for rock burst early warning based on mining production process

Abstract Microseismic (MS) events have been reported in nearly every coal mining country, which could well lead to rock burst in underground coal mines. There has been substantial research progress in statistical methods for characterizing MS data and early warning of rock burst. However, accurately predicting rock burst in coal mines remains challenging, due to which the impact of the mining production process has always been ignored. This paper aims to propose an MS data analysis method combined with the mining production process. In this method, the cumulative mean fluctuation and normality tests were utilized to determine the total length of the time series ( T ), and the autocorrelation function (ACF) was adopted to determine the time window length ( T w ). To be specific, the time series of MS count and energy in the form of cumulative average fluctuation were used to ensure the stability of the background value. As a prerequisite, such indexes of the time series were guaranteed to follow the normal distribution by the kurtosis and skewness tests. Relying on the principles of ACF, the maximum correlation periods among these indexes can be identified as T w . On this basis, MS indicators were calculated in an hour interval, and the associated graded warnings were provided by quantifying their deviation from the average. The application of the method has been successfully demonstrated using actual MS data from a rock burst accident. Compared to traditional methods, the proposed method could recognize early warning precursors, especially in maintenance shifts, and thereby provide more timely alerts.