Research On Incipient Fault Diagnosis Of Limit Switch

As a key component in mechanical equipment,limit switch plays an important role in working reliability and safety.The incipient fault signal of limit switch relatively weak,difficult to be detected,but the failure tends to show gradual development trend,small damage may gradually evolve into a fatal fault.Therefore,it is of great significance to research on the incipient fault diagnosis of limit switch common fault reasons to discover the hidden troubles of faults and to take passive maintenance as active prevention.First,based on the analysis the status of fault diagnosis technology at home and abroad,and also analyzing the limit switch structure and fault characteristics in the dissertation.By setting up the experimental platform to simulating the working process of the limit switch.The feasibility of researching the incipient fault diagnosis method of the limit switch is verified through collecting the voltage signal of the limit switch in the normal state,fastening bolts looseness fault and the limited stretch fault of spring leaf.Second,based on the research of Wavelet-Support Vector Machine(WT-SVM)and EMD-PCA-ELM,the incipient fault diagnosis of the limited switch fastening bolts looseness as the research object of the paper.Based on WT-SVM incipient fault diagnosis method use the voltage signal which had denoised,and feature extraction method is carried out by using wavelet analysis.The SVM is selected for fault classification,and the parameters of SVM optimization is performed using SA-PSO.Based on the EMD-PCA-ELM incipient fault diagnosis method,the emphasis is on feature extraction.Firstly,EMD method is used to extract the time-frequency feature of the voltage signal and calculate the kurtosis value of the IMF component.Meanwhile,extracting mean value,kurtosis,root mean square and so on.Besides,the principal component analysis method is used to reduce the dimensionality of the high-dimensional feature set,which combines the time domain and the frequency domain.Finally,the results of dimension reduction are input to the extreme learning machine for fault classification.Through the analysis of the results,it is shown that the classification accuracy of the research method based on WT-SVM is higher than stretch of spring leaf,the ARMA(p,q)time series model is built,the model parameters are estimated by maximum likelihood estimation,the model parameter values are taken as eigenvalues,and use the K-Nearest-Neighbor classification algorithm to identify the fault information.At last,through to the research of the limit switch fastening bolts looseness fault and limited stretch of spring leaf fault,which can effectively realize the early fault diagnosis of limit switch.