Analysis Of Acoustic Emission Signal Of Fatigue Crack With VMD-CNN

In recent years,China's railway has developed rapidly,operating mileage has been increasing,and operating speed has been accelerating.Therefore,the safety problem in the operation of high-speed railway has become the focus of attention.Locomotive axle is an important part of locomotive and rolling stock running part.The main reason for its failure is fatigue crack.With the further expansion of crack,it will eventually lead to axle fracture,which will do great harm to running part and directly affect the safety of train operation and passengers.Therefore,it is of great significance to monitor the status of train axles and find out the faults in time.In this paper,the experimental data of fatigue crack acoustic emission(AE)signal of train axle is analyzed.In this paper,the axle fatigue crack signal is selected for analysis.In order to meet the actual running conditions of the train,the axle knock signal and the axle background noise signal are added to simulate the impact and noise produced in the actual operation process.In this paper,a feature extraction and recognition method of axles fatigue crack acoustic emission signal based on VMD,SVD and CNN is presented.The original acoustic emission signal is decomposed into several IMF components by VMD,the optimal IMF component is selected for SVD decomposition,and the singular value eigenvector obtained from SVD decomposition is used.As the input of CNN,training and testing are carried out to identify and analyze the acoustic emission signal of axle fatigue crack.The experimental results show that the VMD and SVD methods can effectively extract the characteristic information of axles crack acoustic emission signals,and can accurately identify the axles fatigue crack acoustic emission signals.At the same time,the feature extraction methods of EEMD and SVD are compared in this paper.Compared with EEMD and SVD,the feature extraction method of VMD and SVD is superior to that of EEMD and SVD,which makes the recognition rate of acoustic emission signal of axle fatigue crack improved obviously by CNN.In order to verify the effectiveness of the method used in this paper,the author selects another group of acoustic emission signals of fatigue crack of axle to verify.The results show that this method can extract the characteristic information of fatigue crack of axle effectively,The axles crack acoustic emission signals are effectively identified from the axle tapping signal and the background noise signal,which shows the validity and rationality of the proposed method in the experimental data analysis of the axle fatigue crack acoustic emission signals.