Research On Gearbox Bearing Incipient Fault Diagnosis Based On Feature Enhancement

As an important transmission component in mechanical equipment,the operating condition of gearboxes is directly related to the safe operation of industrial production equipment.Due to the structural characteristics of the gearbox itself,as well as the variability of the operating conditions and complexity of the working environment in actual operation,the vibration signal generated is non-smooth and non-linear,and the impact components are easily masked by noise,which makes it difficult to extract effective fault characteristics directly,resulting in reduced fault identification accuracy.Because no matter how serious the fault is,it develops from an incipient fault with no obvious fault signs.This thesis focuses on the feature enhancement,feature extraction and fault identification of incipient faults in gearbox bearing,and the main research content includes the following parts:(1)To address the problems of small amplitude and weak impact components of gearbox bearing incipient fault signals,the research is based on Ensemble Empirical Mode Decomposition(EEMD)and Maximum Correlated Kurtosis Deconvolution(MCKD)based feature enhancement method.Firstly,EEMD is used to reduce the noise of the signal,eliminate the interference of noise to the signal and obtain the impact components in the signal;then the MCKD algorithm is used to enhance the signal,which can enhance the weak impact component;finally,the envelope analysis of the signal is performed to extract the weak fault feature information.The experimental results show that the method is effective in enhancing the weak shock components in the signal.(2)Given the low decomposition efficiency of EEMD,and the enhancement effect of MCKD method is easy to be affected by parameters,Particle Swarm Optimization(PSO)was introduced to solve the difficult problem of MCKD parameter selection.A Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)and parametric optimization MCKD for gearbox bearing incipient fault feature enhancement are proposed.The reconstruction error of CEEMDAN is small,and the decomposition efficiency is higher than that of EEMD,which can better remove the noise in the signal,and the combination of PSO-MCKD method can realize the feature enhancement of incipient faults.The simulation results show that the CEEMDAN-PSOMCKD method can effectively enhance the weak impact of gearbox bearing in both constant and variable condition incipient fault detection,so as to realize fault diagnosis.(3)A bearing fault diagnosis method based on Multiscale Entropy and Limit Learning Machine with parameter optimization is studied for the problem of feature extraction and fault identification of incipient faults in gearbox bearings.Based on CEEMDAN-PSO-MCKD incipient fault feature enhancement,combined with the advantages of ELM,the extracted Multiscale Entropy was input as feature vector into the improved ELM model for gearbox bearing fault identification.The simulation experimental results show that the average recognition accuracy of the method is99.125%.The incipient fault data of rolling bearings collected by the Key Laboratory of Modern Design and Rotor Bearing System of the Ministry of Education of Xi ’an Jiaotong University is taken as an example to verify the validity of the proposed method.