The Research On Internal Defect Detection Method Of Brake Pads Based On Percussion Signal

Various defects are easy to occur in the production process of brake pads,especially foreign body,delamination,side crack,air bubble and other defects are the main ones.At present,the method of most domestic brake pad manufacturers eliminat the unqualified brake pads is still to let the technical staff with rich production experience judge by visual observation and artificial percussing and listening.This kind of detection method relies too much on experience,high labor cost,and has no good accuracy in time and effort.Therefore,it is necessary to study the new brake pads internal detection method to improve the detection efficiency and detection accuracy of the brake pads.In this context,a method for detecting internal defects of brake pads based on energy features of percussion signals is proposed.In order to solve the problem that it is difficult to detect internal defect of brake pads,firstly,the hammer and sound sensor were used to collect the tapping sound signals of brake pads with various defects,then we obtained valid data information and eliminated redundant extraneous signals by intercepting and preprocessing the collected signals.Time-frequency analysis of signals was perdormed using MATLAB software to compare waveforms and spectra of different defective brake pads.After all the signals were intercepted,the wavelet packet decomposition and variational mode decomposition(VMD)were used to process the signals and extract the defect features.db9 wavelet was used to do three-layer decomposition and eight bands of decomposition signals were obtained.The frequency order and energy distribution of the decomposed frequency band were analyzed,and the energy of each frequency band was extracted as the defect feature.Then we studied the theory of variational mode decomposition and selected the number K of modals according to the principle of similarity of frequency centers.The ? was determined by observing the influence of different penalty factors on the decomposition results and calculation time,the energy of each model was extracted as the defect feature.All energy features obtained by the two decomposition methods were integrated,and the dimensionality was reduced by mutual information method.The energy features obtained by wavelet packet decomposition and variational mode decomposition were used as classification features.We used K nearest neighbor algorithm(KNN)and support vector machine(SVM)to classify and identify defect features,compared KNN accuracy with different K values,the penalty factor c and kernel function parameters g of SVM were optimized by cross-checking.By comparing the classification results of KNN and SVM,it was proved that the possibility of determining the internal defects of the brake.pads by analyzing the energy features on percussion signals.Finally,in order to realize the automatic transportation,percussing detection and classification of the brake pads,we used SOLIDWORKS software to design the three-dimensional prototype model of the brake pads automatic percussing detection device,and introduced the working principle of the conveyor belt,percussion platform and the sorting and screening structure.