Study On Fault Identification Of Train Axle Crack Acoustic Emission Signal Based On LMD Sample Entropy And Energy Ratio

With the rapid development of train vehicles,train vehicles have become China's main transportation tools.The train axle is an important part of the train body structure and the core part of the wheel pair.The fatigue crack of train axle will bring great danger.At present,the detection of axle fatigue crack is still off-line mode which exists missed inspection.In this paper,the on-line detection of fatigue crack of axle is studied by acoustic emission(AE)technology,and the characteristics of nonlinear and non-stationary acoustic emission signal are analyzed.We proposed a AE signal feature extraction method of axle fatigue cracks based on the combination of LMD-ELM sample entropy and energy ratio.The method uses LMD method to deal with the acoustic emission signal of the axle fatigue crack,and uses the sample entropy and energy ratio to quantify the PF component derived from the LMD decomposition,and obtains the complexity information on different PF components,which is used as the characteristic input vector of the classification and recognition of the axle fault signal type;Then,the sample entropy and energy ratio of the PF components are used as input vectors of the type classification of the AE signals to train and test in the limit learning machine.Based on the analysis of the feature extraction and recognition of the train axle acoustic emission signals obtained from the experiment,we can know that the method of the LMD sample entropy and energy ratio is effective for the feature extraction of the axle fatigue crack acoustic emission signal,which becomes the characteristic input vector for the classification recognition.According to the result of classification and recognition,The feature extraction method based on LMD-ELM sample entropy and energy ratio can not only extract the characteristics of the acoustic emission signal of axle fatigue crack,but also can quickly and effectively classify the fatigue crack signals of axle AE.