Trend removing methods of vibration signals of deep hole bench blasting in near field

Abstract

Based on a large number of measured vibration signals of deep hole bench blasting in near field, this paper has contributed the trend mainly to the nonlinear distortion and the low frequency interference superposition with a large amplitude pulse input. On this basis, the effective monitoring range of test instruments has been chosen as criteria to identify the part of the trend. Using ensemble empirical mode decomposition (EEMD), the wavelet analysis, and other signal analysis methods, a trend elimination method is proposed here, which is based on the combination of the frequency band distribution of each intrinsic mode function component and artificial identification. In addition, a wavelet threshold denoising method is also proposed based on autocorrelation analysis to identify noise characteristics. Examples show that the methods are effective and can be realized by batch pretreatment of blasting signals.