Research on gearbox composite fault diagnosis based on improved local mean decomposition

In order to solve the problem of modal aliasing in the processing of complex fault signal with background noise by local mean decomposition(LMD), and the difficulty of extracting and identifying the weak fault feature frequency in the complex fault signal, an improved LMD method is proposed. LMD can adaptively decompose a non-stationary, multi-component signal into a series of instantaneous frequency product functions with physical significance. However, under the influence of noise, the process of LMD will appear modal aliasing, resulting in missed diagnosis and misdiagnosis. In order to solve this problem, the maximum correlation kurtosis deconvolution is proposed to improve the LMD. Firstly, according to the multi-point kurtosis spectrum, the component of fault period contained in fault signal is obtained. According to the fault period, the maximum correlation kurtosis deconvolution is used to denoise the fault signal. Then, the denoised signal is decomposed into a series of components containing fault characteristics by means of LMD. Finally, the component with high correlation is selected for envelope spectrum analysis to extract the component. The characteristic frequencies of different faults and the types of faults are identified. The effectiveness of the proposed method is verified by analyzing the vibration signals of the broken gear-wear composite fault of the gears in the actual gearbox. Compared with the improved empirical mode decomposition method, the superiority of the proposed method is further highlighted.