Study On Mechanical Fault Diagnosis And Degree Evaluation Method For Low Voltage Conventional Circuit Breaker

Low voltage conventional circuit breaker is widely used in the occasions of the total input-wire and important equipment for distributing system of low voltage.Its operation reliability is directly related to the safe operation of the low voltage distribution network.By making diagnosis of the circuit breaker,the hidden faults and fault trends can be found in time,and the predictive maintenance of the circuit breaker can also be realized.The existing fault diagnosis technology research is generally focused on fault location,but the fault degree is rarely involved.In fact,after mastering the damage degree of the mechanical parts,more effective equipment maintenance can be guided.Based on this,this paper mainly studies on mechanical fault diagnosis and degree evaluation of the circuit breaker contact system and operating accessories.In order to obtain the contact vibration signal and the operating accessories current signal during the operation of the low-voltage conventional circuit breaker,a fault diagnosis test platform is developed.Based on this platform,the circuit breaker can be electrically controlled,and the signals can be collected and stored.Firstly,aiming at the need of on-line fault diagnosis and degree evaluation,the switching vibration signal is measured,and an on-line diagnosis method of circuit breaker switching mechanical faults based on gray comprehensive correlation degree is proposed.For this method,by using the gray comprehensive correlation degree,the geometric similarity between the fault feature sequence curves is described to determine the fault attribution,so the changing trend of fault characteristics is fully combined and the fault state of circuit breaker is effectively identified.Secondly,in view of the problem that the traditional method can only make qualitative analysis of fault degree,a quantitative evaluation method for the fault degree of the switching mechanical structure is proposed.The partial mean of multi-scale permutation entropy of the vibration signal is obtained as the quantitative evaluation index of fault degree.Based on this,the fault degree characteristic curves of different fault states are obtained,and they have better monotonicity.The change trend of fault degree can be well reflected,and quantitative evaluation can be realized accurately.Thirdly,aiming at the problem that hidden faults can't be easily detected in routineinspection,a fault diagnosis method of conventional circuit breaker based on operating accessory current is proposed.In feature extraction,the local features and the global features of the current signal are combined to avoid deficiencies of the fault feature.The multi-kernel support vector machine based on genetic algorithm is used to diagnose the fault,so the kernel function selection and parameter determination problems are solved.This method has high fault diagnosis reliability and can detect hidden faults that can't be found by routine inspection.Finally,also for the quantitative evaluation,a degree quantitative evaluation method based on the relative entropy of switching coil loop fault is proposed.By calculating relative entropy of the ensemble empirical mode decomposition energy moment between the fault signal and the normal signal,the characteristic curve is established and then the quantitative evaluation of fault degree is completed.