Experimental Data Analysis Of Acoustic Emission Signal Of Axle Fatigue Crack Based On One-Dimensional CNN And LSTM

With the continuous development of high-speed railway in China,railway safety has become the top priority in people’s daily production and life.As an important part of rail vehicles,locomotive axle is prone to fatigue cracks due to its poor working environment.If not found in time,the cracks will continue to expand and eventually lead to axle fracture,which seriously threatens people’s life and property safety Therefore,the fault monitoring of locomotive axle is very important for the safe operation of the train.In this paper,a recognition method of acoustic emission signal of axle fatigue crack based on one-dimensional convolution neural network(1DCNN)is proposed on the basis of the perfect verification of the papers of our group.In order to simulate the impact and noise produced in the actual operation of the train,the acoustic emission signal of axle knock and axle background noise are added to the acoustic emission signal of axle fatigue No.to interfere.By selecting the best network parameters to train the network,the acoustic emission signal of axle crack can be identified from the interference signal,which can reduce the labor cost and time cost better.The experiment shows that the one-dimensional convolution neural network has a good recognition effect on the acoustic emission signal of axle fatigue crack.In order to compare the effectiveness and rationality of the methods used in this paper,two-dimensional convolutional neural network(2DCNN)and limit learning machine(ELM)and support vector machine(SVM)are used to carry out the comparative experiments.The results show that the one-dimensional convolutional neural network method can identify the acoustic emission signal of axle crack more effectively from the knock signal and background noise signal of axle number.After identifying the acoustic emission signal of axle fatigue crack,this paper also proposes a method based on long and short-term memory network(LSTM)to predict the waveform of the acoustic emission signal of axle fatigue crack.By training the network,the best network parameters are selected to make the network reach the best state and predict the waveform of the fatigue crack signal of axle.The experiment shows that the long and short-term memory network It has a good prediction effect on the acoustic emission signal waveform of axle fatigue crack.