Mechanical Fault Diagnosis Of Typical Distribution Network Switches Based On Motor Current Detection

With the increase of power demand,people pay more and more attention to the safe and stable operation of transmission and distribution switchgear.At present,the maintenance of transmission and distribution switchgear has been changed from traditional periodic maintenance and accident maintenance to state maintenance,which realizes the fault prevention and online real-time monitoring of transmission and distribution switchgear.This paper takes the circuit breaker,electric chassis car and earthing switchgear in the typical distribution network switchgear of Xiamen Huadian Switchgear Company as the research objects,and studies the mechanical fault diagnosis technology of the distribution network switchgear based on the motor current.The main content and innovation of the paper The points are as follows:1.According to the structure mechanism of circuit breaker energy storage mechanism,chassis car and earthing switchgear in the distribution network switchgear and the action process of the respective mechanisms when the DC motor is used as the drive motor,the common mechanical faults in the switchgear and the current of the drive motor are analyzed Corresponding relationship.Through the analysis of the principle of DC motor and the experiment of the relationship between the output torque of DC motor and the armature current,the feasibility of mechanical faults such as high and low operating voltages,transmission jams,spring fatigue and other mechanical faults of the distribution network switching device based on the drive motor current detection is further verified.2.Design the fault simulation test and collect the drive motor current of the circuit breaker,chassis car and earthing switchgear in the distribution network switchgear under different working conditions.The collected signal is processed by two methods of wavelet filtering and wavelet combined with median filtering to reduce noise.The root mean square error RMES of the signal after wavelet and median filtering is smaller than the signal after wavelet filtering.The noise ratio SNR is larger than the signal after wavelet filtering,which proves that the signal processed by wavelet combined with median filtering algorithm is more stable.Combined with the operation process of the distribution network switch equipment,the current profile method is used to extract the current characteristic quantities of the energy storage motor,the chassis motor,and the earthing switchgear motor.The energy storage motor current and the earthing switchgear motor current are divided into three stages:motor starting,motor idling,and motor tension spring.The integral value of the motor current in the spring tension stage is extracted by the segmented integration method to solve the problem of difficulty in detecting spring fatigue;Using the characteristics of current signal fluctuation caused by transmission jamming,the complexity of the signal is increased,and the sample entropy is used to extract the sample entropy of the motor current when the motor is idling and the motor is stretched to solve the problem that the mechanism is difficult to detect.The chassis motor current is divided into three stages:motor start,chassis forward,and plum blossom contact bite,extract the integral value of the plum blossom contact bite stage,evaluate the dynamic and static contact bite of the circuit breaker,and prevent the circuit breaker plum blossom contact spring from loosening.The contact resistance between the dynamic and static contacts is too large;the sample entropy of the motor current during the advancement of the chassis and the bite of the plum blossom contacts is extracted to detect the jamming of the chassis mechanism.3.According to the extracted feature quantities,training samples and test samples are composed,and three diagnostic techniques of random forest,K-nearest neighbor,and Naive Bayes are applied to the circuit breaker energy storage mechanism,chassis car,and grounding switch in the distribution network switching equipment.Troubleshooting.Based on the comparison between the running time of the algorithm and the recognition rate,the optimal diagnosis algorithm of the circuit breaker energy storage mechanism is Naive Bayes,and the optimal diagnosis algorithm of the chassis car and the earthing switchgear is a random forest.