Research On Time-frequency Characteristics Of Acoustic Emission Signals In Rock Failure Stage

The acoustic emission signal contains a large number of useful information about the rock failure process,which is of great significance for monitoring and early warning of rock burst disasters.Focusing on the problem of automatic classification and identification of acoustic emission signals in the rock failure stage,this paper uses a variety of methods to analyze acoustic emission signals from multiple aspects such as time domain,frequency domain and time-frequency domain.Firstly,the whale algorithm is used to optimize variational mode decomposition to preprocess the noise reduction of the acoustic emission signal,and then the differential feature vectors of the time domain,frequency domain and time-frequency domain of the acoustic emission signal are extracted,and the feature fusion is fused,and finally the genetic algorithm is used to optimize the support vector machine to realize the classification and identification of the acoustic emission signal at different stages of rock failure.The main research results are as follows:(1)Aiming at the fact that the decomposition effect of VMD is easily affected by the decomposition layer k and the penalty factor α,a noise reduction method based on whale algorithm to optimize VMD parameters is proposed,and the whale optimization algorithm is used to use fuzzy entropy as the fitness function to find the optimal parameter combination of VMD(K,α).Then,this optimal parameter combination is used to set the VMD parameters to decompose the acoustic emission signal,and for the K IMF components obtained by decomposition,the noise component is removed by the correlation coefficient,and the acoustic emission signal after noise reduction is reconstructed with the remaining effective component.(2)Extract the characteristic parameters of the acoustic emission signal from different aspects such as time domain,frequency domain,and time-frequency domain,and use the main component analysis method to process multiple basic time domain characteristics,and finally get 3-dimensional time domain feature parameters.Use the fast Fourier transformation to obtain the spectrum of the sound transmission signal,and on the basis of this,the frequency domain feature parameters such as the center of gravity frequency and the average frequency are extracted.Analyze the frequency domain feature differences of the sound transmission signal at different stages,and find that there are relatively obvious differences in the two characteristic parameters of the center of gravity frequency and frequency standard difference between the four different stages of acoustic emission signal.n the end,the center of gravity frequency and frequency standard deviation is used as an important frequency domain feature of acoustic emission signal at different stages of the acoustic emission signal.Based on the WOA-VMD algorithm,the composite multiscale sample entropy of the acoustic emission signal is extracted as the frequency domain feature parameter.(3)The genetic algorithm is used to optimize SVM parameters,and an automatic classification recognition model based on GA-SVM is established.The characteristic parameters of the time,frequency and time-frequency domains of the acoustic emission signal are extracted to perform feature fusion to obtain multi-domain fusion feature vectors.Then normalize it,and then input the processed multi-domain fusion feature vector into the automatic classification model of GA-SVM to realize the automatic classification and recognition of acoustic emission signals at different stages of rock failure.The results show that compared with single feature recognition,the recognition accuracy of multi-domain fusion features is the highest,which can reach 87%.The recognition accuracy of GA-SVM classification model for acoustic emission signals is significantly higher than that of ELM and SVM.