Fault Diagnosis Of Rolling Bearing Based On Adaptive Chirped Modal Decomposition

Rolling bearing is one of the used commonly components in mechanical equipment.It is used widely in intelligent manufacturing,aerospace and other industrial fields.The safe and stable operation of rolling bearings will affect the operation status of the entire equipment.Due to the influence of factors such as working environment and strength,the life of rolling bearing has great uncertainty.Once a failure occurs,it may cause accidental property losses and even casualties.Therefore,it is of great significance to carry out state monitoring and fault diagnosis research.Taking rolling bearings as the research object,a fault diagnosis method of rolling bearing based on Adaptive Chirped Modal Decomposition(ACMD)is proposed in this thesis.To study the signal decomposition effect of the ACMD algorithm,a simulation signal is constructed for analysis,and the decomposition results are compared with those of classical EMD algorithm.The results show that the ACMD algorithm has more advantages in signal decomposition accuracy,time-frequency resolution of components,and mode aliasing suppression.The proposed algorithm is used to analyze the fault vibration signals and extract the fault feature information.The signal analysis effect of ACMD algorithm and EMD algorithm is compared,and the validity of ACMD algorithm applied to the fault diagnosis of rolling bearing is verified.To reduce the influence of interference noise and enhance the fault feature information,a fault diagnosis method of rolling bearing based on ACMD and Multipoint Optimal Minimum Entropy Deconvolution Adjust(MOMEDA)is proposed.The sensitive components decomposed by ACMD algorithm are screened according to the weighted kurtosis criterion,then the MOMEDA algorithm is used to enhance the feature information of the reconstructed signal.Finally,the envelope spectrum demodulation analysis is performed.Compared with the feature extraction results of EMD-MOMEDA,it reflects the advantages of extracting more fault information and more accurately by the method proposed in this thesis.To optimize the fault diagnosis process and improve the problem that traditional fault methods rely too much on prior knowledge,according to the characteristics of rolling bearings signals,a rolling bearing fault diagnosis method based on ACMD Refined Composite Multi-scale Symbolic Dynamic Entropy(RCMSDE)and GG clustering is proposed.Firstly,the ACMD algorithm is used to decompose the fault signal,and then the refined composite multi-scale symbolic dynamic entropy is introduced.The feature vector is constructed by using the RCMSDE value of the characteristic signal component,and the feature vector is imported into GG clustering,which successfully realized the recognition of different fault types of rolling bearings.The clustering effect is evaluated by correlation clustering indicators,and compared with ACMD-K mean clustering and ACMD-GK clustering methods.The results show that the method proposed in this thesis has a better clustering effect and can identify different fault types of rolling bearings with higher identification accuracy.To further improve the fault process,improve the effect of fault diagnosis,and make fault diagnosis more intelligent and convenient,this thesis combines ACMD algorithm and CNN-Bi GRU-Attention neural network model to propose a fault diagnosis method for rolling bearings based on ACMD-CNN-Bi GRU-Attention.Firstly,the rolling bearing fault vibration signal is decomposed by ACMD algorithm,and then the characteristic components are selected according to the weighted kurtosis criterion for signal reconstruction.Finally,the CNN-Bi GRU-Attention model is used for learning and intelligent classification.The proposed method is applied to the vibration signal data processing of 10 different types of rolling bearings,and the fault diagnosis results are evaluated through classification accuracy,precision and other indicators.The results show that the proposed method can accurately diagnose rolling bearing faults.Compared with the result of CNN-Bi GRU-Attention and ACMD-Bi GRU-Atten tion methods,the method in this thesis has higher fault accuracy,higher precision,and better diagnosis effect.