Research On Fault Diagnosis Methods Of Variable Speed Planetary Gearbox Based On Multiple Information Monitoring

Studying the fault diagnosis technology of gearbox is of great significance for the smooth operation of mechanical equipment.The signals of planetary gearbox under variable speed conditions have non-stationary and nonlinear characteristics.With the development of technology,the problem of noise reduction and classification of multiple signals for variable speed planetary gearbox is a difficult problem in the field of mechanical fault diagnosis today.The planetary gearbox is taken as the research object,and the noise reduction algorithms such as Chirplet Path Pursuit(CPP),Computed Order Tracking(COT),Distributed Compressed Sensing(DCS),Random Forest(RF),and Seagull Optimization Algorithms(SOA),as well as their applications in multi-signal processing of variable speed planetary gearbox are studied by this article.The main content of this article is as follows:In order to solve the problem that non-stationary signals are difficult to be processed by traditional signal analysis methods,the combination of CPP and COT is used.The CPP is used by this method to extract rotational speed from the original signal as the reference axis pulse to calculate the fitting of COT.By comparing it with three commonly used speed extraction methods through simulation and experimental signals,the results show that the algorithm has higher accuracy and noise resistance.The applicability of COT for nonstationary signals is verified by the analysis results of simulation signal,which can accurately retain effective order components.Experimental results show that the COT method based on CPP can obtain stable signals,retain effective information of signals and avoid order aliasing.In order to reduce the noise of multiple signals,DCS is used to reduce the noise of multiple signals of planetary gearbox.Aiming at the problem that DCS is difficult to effectively reduce the noise of non-stationary multiple signals,a COT-DCS method combining COT and DCS is proposed.The COT based on CPP is used to remove the signal non-stationary,and then DCS is used to reduce the noise of multiple signals.By comparing the results of this method with the result of SVD-GLCT and EMD-MPE through experimental signals,the COT-DCS method combines the advantages of the algorithm and is superior to the two comparison algorithms in suppressing non-stationary noise,denoising effectiveness,and preserving information components.Aiming at the low classification accuracy of RF algorithm,SOA is used to optimize its parameters.In order to overcome the shortcomings of the SOA,which is prone to falling into local optimization and has low solving accuracy,an Improved Seagull optimization algorithm(ISOA)with four strategies is proposed.The improvement effect is verified through basic testing functions,and the results show that the ISOA is superior to the basic algorithm in convergence analysis,optimal value search,and other aspects.The ISOA is used to optimize the parameters of the RF,and the features can be more effectively classified by the optimized RF.