The Axle AE Signals Feature Extraction Based On Wavelet Packet Energy And LMD

With the rapid development of urban track traffic in our nation,the constant progress and innovations have made track vehicles smarter,more efficient and more complex.So this precaution is necessary to ensure track vehicles safety.Axle is an important part of track vehicles,it has the effect of loading,guiding and supporting.Without it,vehicles are unable to run normally.In the operation process,the failure of axle will lead to serious consequences,which results in enormous economic loss,serious threat to human security.This article analyses the fault mechanism of axle and processes the fault signals for features extraction.Application simulation and experiment for diagnosing faults in axles,in order to ensure the security and reliability of vehicle.More details is as follows:(1)This paper analysis the fault signals of locomotive axle with two different methods,and the two part are the extraction of the feature of acoustic emission signal based on resonance demodulation and the spectrum of wavelet packet analysis,the method of fault diagnosis which based on LMD algorithm.The crack signal of axle can be extracted effectively by different means,it may provide a new method to acoustic emission.(2)The measured crack signals can be divided into several sub-bands by wavelet packets based on the theory of wavelet and wavelet packet.Then the energy of each frequency band is calculated to draft the percentages of each band on total energy.After that,take it compare with the energy spectrum of normal acoustic emission signal and obtain brand which changes obviously.In the demodulation process,according to the range of frequencies,the fault frequency will be extracted by demodulation.(3)Simulations and experiments show that PF components divided from crack signals by using the algorithm of LMD.Then the energy of each PF components is calculated to draft the percentages of each components on total energy.Finally,based on the technology of resonance demodulation,PF component which occupies the most energy is drawn by time-frequency spectrum,which can properly indicate axle status.This paper uses two methods to research the fault signals.Both of them were verified their feasibility by experimentation,and then compares them.It opens up a new way for the research of AE signals and provides some condition for further research.