Rearrangement ST-based low-signal-noise-ratio micro-seismic event identification method

Abstract

The invention discloses a rearrangement ST-based low-signal-noise-ratio micro-seismic event identification method. The method comprises the steps of firstly, performing ST on micro-seismic data to obtain time frequency spectrums of signals; secondly, according to a Parserval theorem and rules on scale transformation and translation in Fourier transformation properties, performing partial differential operation on the time frequency spectrums to obtain instantaneous frequencies of the signals; thirdly, converting a value calculated at any point for a spectrogram to a gravity center of energy distribution, thereby obtaining a rearranged time frequency matrix of the time frequency spectrums of the signals in a frequency direction; and finally, building a multi-classification SVM for realizingclassification of micro-seismic signals, blast signals and mechanical noises. The problems of low identification rate, rough classification and low classification accuracy of low-signal-noise-ratio micro-seismic signals in the prior art are solved, so that the time frequency resolution of the low-signal-noise-ratio micro-seismic signals is remarkably improved, the classification is accurate and the classification of multiple weak signals is realized; and the method can be well applied to the mine safety production and coal mine illegal mining monitoring technologies.

Description

technical field [0001] The invention relates to the technical field of information processing, in particular to a method for distinguishing microseismic events with low signal-to-noise ratio based on rearranged ST. Background technique [0002] Microseismic monitoring is one of the important means to achieve safe production in mines and prevent illegal mining in coal mines. Microseismic technology provides technical support for coal mine monitoring. Seismic signals are typical non-stationary signals with low signal-to-noise ratio, including microseismic signals (referred to as rock mass rupture signals in this paper), blasting signals, mechanical equipment noise such as drilling rigs and mine cars, and mine-drawing noise. The monitoring of microseismic signals is an important guarantee for safe production in coal mines, and the monitoring of blasting signals is the key to preventing illegal mining in coal mines. Therefore, how to accurately identify microseismic signals and...

AI Technical Summary

Problems solved by technology

The length of the short-time Fourier transform (STFT) window function is manually selected based on experience, and the signal resolution cannot be adjusted adaptively

Continuous wavelet transform (CWT) can achieve multi-scale analysis of signals, but it is difficult to choose the wavelet mother function, and the transformation results obtained by different wavelet mother functions are also different; compared with CWT, the S-transform (ST) algorithm is basically The wavelet is a fixed function, and the window size is adaptively adjusted according to the signal frequency to ensure the adaptive adjustment of the resolution, and enhance the energy of the high-frequency and weak-amplitude signals, but due to the limitation of the Heisenberg uncertainty principle, its resolution is limited ;

Images