Research On Mechanical Fault Diagnosis Of Circuit Breakers Based On Current Characteristics Of Switching Coil

High voltage circuit breaker is an important switching element in distribution network,which plays a very important role in people's production and life.Once the mechanical failure of high-voltage circuit breaker occurs,it will cause great losses to the people and the country.Therefore,monitoring and diagnosis of mechanical failure of high-voltage circuit breaker is of great value to ensure the safety of power system and national production and life.This paper is mainly based on the characteristics of the closing coil current to diagnose the mechanical faults of the high voltage circuit breakers.Firstly,the research status of mechanical fault diagnosis of high voltage circuit breaker is introduced,and the current signal of switching coil is determined as the research object,and the related theory of switching coil current and the corresponding fault types are analyzed in detail.Then,the pretreatment,feature extraction and pattern recognition of the current signal are carried out.The current curve pretreatment adopts three methods of interpolation,fitting and morphology to denoise,and the selected morphology method has better smoothness effect.The optimal approximate derivative method is used to extract the current and amplitude characteristics of the current,and four global features,namely mean,standard deviation,kurtosis and energy,are added.Dynamic time warping(DTW)and support vector machine(SVM)are applied to classify the current characteristics of split and close coils.Finally,on the basis of analyzing the current signal,the vibration signal of spring operating mechanism is collected,short-term energy method and double threshold method are used to extract the characteristics of vibration signals of high voltage circuit breakers,and the multi-source characteristics are composed of current and vibration signal characteristics.In view of the large number of features,principal component analysis(PCA)is used to reduce the dimension of features and support vector machine is used to recognize multi-source features.Five kinds of mechanical states were simulated in the experiment,including normal state,closing spring fatigue fault,transmission loosening fault,low control circuit voltage fault and buffer spring fatigue fault.The analysis shows that the method can extract the current signal of the switching coil quickly and effectively,and the fault identification method based on the current of the switching coil and vibration characteristics has a more comprehensive and accurate classification effect.