Research on Fault Diagnosis of High Voltage Circuit Breaker Based on Spark Machine Learning

Abstract

High voltage circuit breaker (HVCB) is the most important control and protection equipment in the power system. The mechanism of HVCB is relatively complex, and the probability of failure is large. If the failure occurs, the safe and stable operation of the entire power system will not be guaranteed, which will cause huge economic losses. Due to the complexity of traditional monitoring and diagnosis methods, the information obtained is not comprehensive enough, which leads to insufficient precision of diagnosis and affects the diagnosis results. In order to determine fault characteristics and categories simply and accurately in HVCB, we propose a new fault diagnosis method based on Spark machine learning to predict fault types. Firstly, by preconditioning data and extracting characteristic values of the moving contact stroke signal data of HVCB, the contact stroke, contact opening and closing movement time, contact moving average speed and maximum speed are extracted as characteristic attribute signal input, and the fault state of circuit breaker is taken as experimental output. Secondly, the improved support vector machine (SVM) algorithm model in Spark machine learning is used for data analysis and processing, and the relevant performance indexes are obtained through model evaluation, and the classification diagnosis results are output by the optimal model. Finally, the improved SVM model of Spark machine learning is compared with the traditional linear regression and naive bayes classification model. The experimental result shows that the proposed fault diagnosis method based on Spark machine learning is faster and more accurate than other model diagnosis methods, and the classification effect is more obvious, which can meet the needs of rapid data processing and high precision. It is feasible to apply the improved SVM algorithm of Spark machine learning to the fault diagnosis of HVCB, which improves the fault diagnosis level of circuit breaker, and has a certain practical significance for the realization of predictive maintenance of circuit breaker.