Research on Tool Fault Diagnosis Based on GA-VMD Decomposition and Support Vector Machine
In order to improve the accuracy of tool fault identification,a variational modal decomposition(VMD)method optimized by genetic algorithm(GA)is proposed by studying the decomposition method of non-stationary vibration signals.The method takes sample entropy as the objective function value,and uses genetic algorithm to iteratively calculate the sample entropy to obtain the optimal decomposition level k and penalty coefficient for variational modal decomposition α.On this basis,the tool vibration signal was decomposed and tool fault features were extracted.Next,the fault features were screened using Nearest Neighbor Component Analysis(NCA)to obtain features with strong correlation with the tool fault status.Finally,the screened fault features were input into the PSO-SVM classification model for tool fault diagnosis.Conclusion:The optimized VMD decomposition method has significant noise reduction effect and can extract good tool fault features.The accuracy of tool fault recognition has been significantly improved,providing a theoretical basis for signal decomposition layers and tool fault diagnosis.
NETL
NSTL
万方数据
