Research On Gearbox Fault Diagnosis Method Based On SSD

As an important part of the rotating mechanical transmission system,the gearbox directly determines whether the entire mechanical system can operate normally,a research of fault diagnosis on gearbox has important practical significance.Therefore,this paper mainly takes the bearings and gears in the gear box as the research objects.Based on the study of singular spectrum decomposition,it is combined with Teager energy operator,singular value decomposition,composite multi-scale permutation entropy and other methods to realize the fault diagnosis of the gearbox.The main research contents of the paper are as follows:1.Research on rolling bearing fault diagnosis method based on SSD and Teager energy operatorThe performance of singular spectrum decomposition(SSD)in processing nonstationary signals is discussed.A fault diagnosis method based on singular spectrum decomposition and Teager energy operator is proposed.Firstly,SSD is used to decompose the fault signal into a set of singular spectrum components(SSC)and then the SSC with the biggest kurtosis is selected as the best component according to the kurtosis criterion.The Teager energy operator is used to calculate the component to obtain the Teager energy spectrum of the signal.Analysis of the simulated and measured signals shows that the proposed method can effectively demodulate fault feature information of the bearing and the diagnostic effect is better.2.Gear fault feature extraction based on ISSD and singular value decompositionAiming at the problem that the gear fault characteristics are weak and difficult to extract effectively under strong background noise,a gear fault feature extraction method based on improved singular spectrum decomposition(ISSD)and singular value decomposition(SVD)is studied.Aiming at the defect that the modal parameters need to be selected by experience in the singular spectrum decomposition algorithm,the singular spectrum decomposition is improved based on the dispersion entropy optimization algorithm.Then,the improved singular spectrum decomposition is combined with singular value decomposition,and the singular value energy standard spectrum is used to adaptively determine the signal reconstruction order number to restore the signal and improve the noise reduction effect.Finally,the fault features are extracted by using envelope spectrum analysis.The proposed method is applied in simulated signals and gear measured signals,and compared with EMD-SVD and other methods.The results showed that the proposed method has better effect of noise reduction and feature extraction,and can more effectively realize the identification of gear faults.3.Fault identification of wind turbine gearbox based on SSD and composite multi-scale permutation entropy.Aiming at the problems of severe noise interference of vibration signals of wind turbine gearboxes and the difficulty of extracting fault features,a method for fault identification of wind turbine gearbox based on SSD and composite multi-scale permutation entropy(CMPE)is proposed.Firstly,the vibration signal of fan gearbox is decomposed by SSD,the main SSC component is selected by using the SSD-TEO index,and then the composite multi-scale permutation entropy value of the main SSC component is calculated to initially identify the gear state.Finally,the feature vector composed of entropy value is input into PSO-SVM for precise identification of gear state.The verification of the fan gearbox example shows that the SSD-CMPE method proposed in this paper can accurately identify the fault state of the gear;compared with the SSD-MPE method,it has higher recognition accuracy.