Application Research Of Wavelet Bispectrum In Locomotive Fault Diagnosis

Due to the harsh working environment and frequent load changes,the locomotive gears and bearings are faulty.If they are not in operation,they need to be replaced in advance.Compared with the commonly used vibration signal based detection technology,the diagnosis method based on the stator current analysis of the traction motor has the advantages of economy and easy implementation.In order to accurately and reliably detect the locomotive fault,this paper introduces the wavelet dual-spectral vibration into the vibration and traction motor stator current signal analysis,and did the following work:On the basis of summarizing the quadratic phase coupling of locomotive vibration signal,the mechanism of quadratic phase coupling of traction motor stator current signal is discussed.The phase coupling mode is summarized.The relationship between quadratic phase coupling and fault is discussed.Bispectrum is an effective tool for identifying quadratic phase coupling.However,having a large estimation error requires averaging a large amount of data,which is not suitable for short data analysis.In this paper,wavelet bispectrum is introduced into the fault diagnosis of locomotive transmission system,and the problem of quadratic phase coupling extraction under variable speed conditions is preliminarily solved.Aiming at the problem that wavelet bispectrum estimation results are easily interfered by pseudo-coherence,the application of Biphase Random Wavelet Bispectrum and Instantaneous Wavelet Bispectrum is studied.The simulation signal analysis shows the excellent characteristics of wavelet bispectrum,which can be used to analyze the quadratic phase coupling of the fundamental frequency with time and effectively extract the fault characteristics.The experimental verification on the locomotive drive shaft test bench confirmed the superiority of the wavelet bispectrum method compared with the traditional bispectrum.Summarizing the above results,a fault diagnosis method for locomotive traction drive system using wavelet bispectrum detection signal quadratic phase coupling and fault feature extraction is proposed.The actual running test of the locomotive proves that the above method can accurately diagnose the faults of the locomotive gears and bearings.