Fault Diagnosis Of High Voltage Circuit Breaker Based On Improved SVM Algorithm

High voltage circuit breaker is an important protection and control equipment in power system.Its stable operation is very important to the normal operation of power system.The opening and closing action of high voltage circuit breaker is completed by connecting the control circuit with the operating mechanism.Due to the close and complex layout of circuit breaker components,various electrical and mechanical faults are prone to occur.The breaking operation of high-voltage circuit breaker is accompanied by the changes of various kinds of signals,that is,all kinds of accompanying signals can reflect the operation state of circuit breaker.Therefore,it is the key to realize the condition based maintenance and ensure the safe operation of power grid to identify,investigate and diagnose the main faults of circuit breaker in time by monitoring and processing all kinds of accompanying signals and establishing fault diagnosis model.Under this background,this thesis starts from the collection and feature extraction of the current signal of the opening and closing coil and the displacement signal of the moving contact,and combines with the improved SVM algorithm to diagnose the electrical and mechanical faults of the high-voltage circuit breaker.The detailed research contents are as follows.(1)In this thesis,the current signal of the opening and closing coil and the displacement signal of the moving contact are taken as the research objects,and the relevant theories of the two signals and the fault types that can be reflected are analyzed.Then,the 7-dimensional characteristics of the peak and valley current value,peak current time and valley current time in the opening and closing current signal and the 3-dimensional characteristics of the moving contact displacement signal are extracted as the important criteria for fault diagnosis,and fused into 10 dimensional multi-source features.(2)According to the feature quantity of the extracted high-dimensional data,this thesis selects PCA algorithm(Principal Component Analysis)to reduce the dimension of the 10 dimensional multi-source features,takes the cumulative contribution rate K(m)as the evaluation index,and takes the feature quantity when K(m)≥95% as the original data to determine the final feature set.(3)According to the data characteristics of such small fault samples as high voltage circuit breaker,this thesis selects support vector machine(SVM)as the basic algorithm of diagnosis.In order to further improve the accuracy and efficiency of diagnosis,this thesis proposes a fault diagnosis model of high voltage circuit breaker based on apso-pca-svm algorithm,Griewank evaluation function is used to test and verify its advantages compared with PSO-SVM and GA-SVM.Finally,APSO(Adaptive Particle Swarm Optimization)algorithm is used to optimize the SVM penalty factor and kernel parameters,and the APSO-PCA-SVM fault diagnosis model is established.(4)Taking ZW10-40.5kv VD4 Vacuum circuit breaker as the research object,this thesis simulates four common electrical and mechanical fault states except the normal state: low voltage of control circuit,loose transmission mechanism,large empty stroke of iron core and jammed electromagnet.By selecting the adaptive sensor to collect the current signal of opening and closing and the displacement signal of moving contact under five states as the data support of fault diagnosis,the example analysis results show that this method can remove the redundant information to the greatest extent,simplify the diagnosis model,and improve the diagnosis accuracy and efficiency,with the diagnosis accuracy of 96.67%,When there are few fault samples,using limited feature quantity can achieve more comprehensive and accurate fault diagnosis of small sample equipment such as high voltage circuit breaker.