Micro-seismic signal classification and identification method based on deep learning

Abstract

The invention discloses a micro-seismic signal classification and identification method based on deep learning, and belongs to the field of signal analysis and identification. The method includes following steps: step 1, establishing a sample database of micro-seismic signals and blast signals; step 2, extracting characteristics of the dominant frequency, an after-peak attenuation coefficient, andan energy gravity center coefficient of sample signals, and forming a sample characteristic data training set and a test set; step 3, training a deep neural network classification and identificationmodel by employing the sample characteristic data training set, verifying a classification and identification effect of the signal classification and identification model by employing data of the testset, and continuously improving the classification precision through crossed training; and step 4, extracting a characteristic vector of a to-be-identified signal, inputting the signal into the signal classification model, and obtaining an identification result. The method has characteristics of simple algorithm, high adaptability and timeliness, and high identification accuracy, the coal mine micro-seismic signals and the blast signals can be effectively classified, and the technical value and the application prospect are very good.

Description

technical field [0001] The invention belongs to the field of signal analysis and identification, and in particular relates to a method for classifying and identifying microseismic signals based on deep learning. Background technique [0002] In addition to a large number of effective coal rock fracture microseismic signals, the signals collected by the coal mine microseismic monitoring system also include a large number of blasting vibration signals generated by coal mine blasting operations. The waveform of the coal rock fracture microseismic signal is very similar to the blasting vibration signal waveform. Rock fracture microseismic signals are identified from a large amount of monitoring data, and manual identification is used, which is difficult to identify and low in work efficiency. [0003] At present, domestic and foreign identification methods for coal mine microseismic signals and blasting signals mainly include Fourier transform, wavelet transform, wavelet packet ...

AI Technical Summary

Problems solved by technology

For example, the traditional Fourier transform is mainly used to analyze periodic stationary signals, and it is not effective in analyzing random and non-stationary microseismic signals containing spikes and mutations; the EMD method has boundary effects and modal aliasing phenomena, resulting in EMD The decomposed signal is unstable and non-unique, and these defects of EMD make it inevitable to have disadvantages in signal identification

When these methods are used for signal analysis, they solve the identification problem of the two types of vibration signals to a certain extent, but they ignore the application of the new generation of information technology such as the current coal mine monitoring big data environment and deep learning in the signal classification and identification technology. The accuracy of signal classification and identification has been further improved

Images