Research On Gearbox Compound Fault Diagnosis Based On ITD And Blind Source Separation

The gearbox is one of the key transmission components in mechanical equipment,and its health and working life are important in the operation of the entire equipment.The key components in the gear box are rolling bearings and gears.In this paper,in order to solve the problem that the compound faults of rolling bearings and gears in the gear box are difficult to diagnose,the intrinsic time-scale decomposition(ITD)algorithm of the signal is studied,and the minimum entropy deconvolution Product(Minimum Entropy Deconvolution,MED)algorithm,and Fast Independent Component Analysis(Fast ICA)blind source separation algorithm,the three methods are combined,verified by simulation and actual measurement,which shows that this method is used in gearbox compound fault diagnosis effectiveness.Firstly,the paper introduces the types and vibration mechanisms of rolling bearings and gears in gear boxes when they fail.Based on the establishment of simulation signals,according to the different fault characteristics of gears and rolling bearings,the demodulation and extraction methods for different fault frequencies are studied.Bert envelope demodulation method is used to extract the characteristic frequency of the fault.Secondly,the thesis studies the preprocessing methods of fault signals.Aiming at the single-channel signal acquisition,a suitable decomposition method is found to pre-process the single-channel signal to extract the different fault characteristics therein.In this paper,the ITD decomposition method is combined with two kinds of evaluation methods,such as fast spectral kurtosis map and cross-correlation number,to screen out the true components of ITD decomposition and accurately obtain the information related to the fault signal.Considering the strong noise interference during actual testing,the algorithm adds a MED noise reduction method and optimizes its parameters to effectively filter out noise interference.Finally,the paper analyzes the fault characteristics of gearbox mixed fault signals.First analyze the fault signal directly by using the envelope and frequency spectrum;then use the MED method to perform noise reduction analysis on the fault signal to find the fault,then use ITD and Fastica to do the same processing to extract the fault characteristic frequency.The comparative analysis of the simulated and measured signals shows that the method used in this paper is effective for weak fault extraction.In summary,the ITD-MED-Fast ICA method proposed in this paper provides a new idea for gearbox composite fault diagnosis,and has certain application value for gearbox composite fault diagnosis in engineering practice.