Research On Bearing Fault Diagnosismethod Based On Parameter Optimization VMD And WOA-SVN

Rolling bearings are the most important parts of various mechanical structures and are also the most prone to failure and damage.Their damage may directly affect the normal operation of large production equipment,bringing huge economic losses and security risks.Therefore,to carry out the fault diagnosis research of rolling bearings,in order to ensure the safe and stable work of machines and equipment.In this paper,a diagnostic method combining parameter-optimized variational modal decomposition(VMD)and WOA-SVM is proposed and applied to the fault diagnosis of rolling bearings,seeking to overcome the difficulties in obtaining information on the weak characteristics of bearings and improving the accuracy of fault identification and analysis.Firstly,a genetic algorithm is used to optimise the variational mode decomposition,so as to determine the number of components and penalty factors of the VMD and to obtain the optimal decomposition results;secondly,the vibration characteristics of the bearing are analysed using the variational mode decomposition with the sample entropy as the fitness function and divided into several intrinsic mode components(IMF);finally,the energy entropy values of each intrinsic mode component are calculated and formed into a feature vector set,which Finally,the energy entropy values of each intrinsic mode component are calculated and formed into a feature vector set to complete the fault feature extraction.It is verified through experiments that the GA-VMD energy entropy feature extraction method proposed in this paper effectively suppresses the undesirable phenomena such as modal mixing and over-decomposition of EMD,and the effectiveness of the feature extraction method proposed in this paper is verified by comparing it with the GA-VMD sample entropy and GA-VMD alignment entropy feature extraction methods.To solve the problem of difficult parameter selection in the pole configuration of support vector machines,the structure of the support vector machine is optimised using an improved whale algorithm(WOA)as an adaptation function in terms of diagnostic accuracy,which in turn leads to more reasonable kernel width coefficients and penalty coefficients.Finally,the extracted features are input into the WOA-SVM diagnostic model for training and fault identification and diagnosis,and the method is experimentally compared with SVM,ELM and KELM to verify that it has higher diagnostic accuracy and achieves effective diagnosis of rolling bearing faults.