基于概率模型辅助的高压断路器故障预测研究

Research on Fault Prediction of High Voltage Circuit Breakers Assisted by Probability Models

概率模型可用于辅助研究高压开关柜断路器的故障预测,并可基于历史故障数据和其他相关因素,通过统计分析和机器学习算法来预测高压开关柜断路器的故障概率。高压开关柜断路器(HVSCB)中大多数故障的根源是运行机制,而行程曲线(TC)可以有效地揭示操作机构的状态。本文介绍了HVSCB常见失效模式对TC剖面的影响,并提出了一种新的模型辅助方法,将基于模型的方法和基于规则的方法相结合来模拟操作机构的行为。HVSCB的工况根据其运行速度分为三种模式即正常(约3 m/s)、故障1(小于正常范围)和故障2(大于正常范围)。HVSCB的模式变化已通过交互式多模型(IMM)在每次操作中进行了估计,所提出预测故障和分析故障原因的方法已通过了实验验证。

The probability models can be used for assist research in fault prediction of high-voltage switchgear circuit breakers,and can predict the fault probability of high-voltage switchgear circuit breakers based on historical fault data and other relevant factors through statistical analysis and machine learning algorithms.The root cause of the most faults in high-voltage switchgear circuit breakers(HVSCB)is the operating mechanism,and the travel curve(TC)can effectively reveal the status of the operating mechanism.This article introduces the impact of common failure modes of HVSCB on the TC profile and proposes a new model assisted method that combines model based and rule based methods to simulate the behavior of operating mechanisms.The HVSCB operating conditions are divided into the three modes based on its operating speeds of the normal(approximately 3 m/s),the fault 1(less than normal range)and the fault 2(greater than normal range).The mode changes of HVSCB have been estimated through Interactive Multiple Models(IMM)in each operation,and the proposed method for predicting faults and causes before they occur has been experimentally validated.