基于随机森林算法的断路器分合闸线圈故障电流曲线识别

To identify the fault types of high-voltage circuit breaker accurately and reliably, a fault diagnosis method for high voltage circuit breaker is formed. In this paper, the reclosing coil current of HV circuit breaker under different conditions is analyzed, in order to extract time and current as characteristic parameters to form training set. Combining random forest algorithm (RF) with training set, random forest classifier is established to distinguish fault types of high voltage circuit breakers efficiently. Experiments show that the algorithm can accurately distinguish the fault of high voltage circuit breaker, which greatly improves the utilization rate and efficiency of HV circuit breaker.