Research On High-voltage Circuit Breaker Mechancal Vibration Signal Analysis And Fault Diagnosis Technology

High voltage circuit breaker condition monitoring provides a powerful guarantee for the safe operation of the whole power network.At the same time,it is of great practical significance to ensure the reliability of high voltage circuit breaker,reduce the waste of human and material resources.Aiming at the problem of the current circuit breaker on-line monitoring system that has the single function,lacks of effective detection means and so on,the research group and Hebei Electric Power Survey and Research Design Institute joint research in Fault Diagnosis System For High Voltage Circuit Breaker Information Fusion.According to the structure characteristics of 6SF circuit breaker and motion process,select the appropriate sensor type and installation method of the circuit breaker contact,circuit breaker vibration signal,circuit breaker action coil current and main circuit current monitoring.Through wavelet packet frequency band energy analysis,it is found that the distribution of the most sensitive frequency band is more diffuse than that in one or several frequency bands.In view of this situation,a method for judging the characteristic frequency band of circuit breaker vibration signal is presented.The energy state diagram is defined,and the characteristic frequency bands are extracted.The characteristic frequency band range is in agreement with theoretical analysis,at the same time,in the energy state of the state,the distribution of the multi-Region-like steady state has good repeatability.In the traditional method,the power spectrum estimation is used to analyze the vibration signals of the circuit breakers,but the analysis of the actual data shows that,both direct method,indirect method and the method of improvement do not have comprehensive analytical advantages.On the basis of the analysis of several power spectrum estimation models,the power spectrum estimation method is put forward,and the parameters are determined by using the theory of envelope variance regression.Based on support vector machine theory,the classification accuracy of multiple kernel functions is calculated by using the actual data,and the data of high voltage circuit breaker is analyzed by using the radial basis function support vector machine.On the basis of analyzing the traditional Method Hold-Out method,improved 10-CV cross validation method is proposed in this paper,which can overcome the shortcomings of the traditional method,which can overcome the classification accuracy of the verification set.At the same time,support vector machine parameter optimization is put forward,Starting with rough parameters optimization,the optimization of the parameters is gradually shifted to the optimal parameters.This method has the advantages of high efficiency and no cross thinking,and can greatly speed up the speed of parameter optimization.It is especially evident when dealing with high dimensional data.The analysis results show that the feature extraction method can effectively extract the fault signal characteristics.Compared with the traditional method,the modified 10-CV cross validation method based on radial basis kernel function is more rapid and accurate than the traditional method.To carry out verification test,the 6SF of LW25-252/ T4000-50 high voltage circuit breaker is used as the research object to test the fault diagnosis system of high voltage circuit breakers(including feature signal acquisition,vibration signal preprocessing,AR model power spectrum estimation and Labview data call processing system).Test results meet the design requirements of high voltage circuit breaker fault diagnosis,and the test results are in agreement with the theoretical analysis.