Abstract: In order to analyze and diagnose the status of high-voltage circuit breakers more accurately, a multi-strategy improved sand cat swarm optimization (ISCSO) algorithm is proposed to optimize the fault diagnosis method of support vector machine (SVM). Firstly, sand cat swarm optimization(SCSO) algorithm is improved through various improvement strategies to enhance the algorithm's global search ability, local search ability, and global balance ability. Two different types of test functions are used to test the performance of ISCSO, verifying its stronger convergence and optimization ability. Then, ISCSO is used to optimize SVM and establish a fault diagnosis model. Next,complete ensemble empirical mode decompasition with adaptive noise energy entropy is used to extract features from vibration signals and construct a feature sample set. Finally, the extracted feature sample set is input into the ISCSO-SVM model for fault diagnosis of high-voltage circuit breakers. The experimental results show that the diagnostic accuracy of this method reaches 96.29%. Compared with the other three models, it has been proven that this method has higher accuracy, and better stability.

Key words: high-voltage circuit breaker, fault diagnosis, CEEMDAN energy entropy, ISCSO, support vector machine