基于SVM算法的真空断路器永磁机构故障诊断方法

A Fault Diagnosis Method for Permanent Magnet Mechanism of Vacuum Circuit Breaker Based on SVM Algorithm

永磁操动机构作为新式真空断路器常用的操动机构,其运行状态决定真空断路器的性能,因此有必要对永磁操动机构进行故障诊断研究。本文以ZW45-12型永磁机构真空断路器为研究对象,对真空断路器永磁机构行程曲线进行特征参数分析,选择断路器启动时间、动作时间、刚分合速度、分合闸平均速度等四个参数作为诊断特征参数;基于断路器分合闸实验数据,对比常用的故障诊断算法,结果表明SVM算法性能最优;基于Spark平台搭建使用SVM算法的永磁机构故障诊断模型,并通过不断调整SVM算法的惩罚参数C和核函数kernel,对诊断模型进行优化。优化的SVM故障诊断模型对永磁机构回路电阻增大、机构卡涩及分闸弹簧单根脱落故障诊断精确率均在90%以上,分闸弹簧单根脱落故障诊断准确率可达96%,可以满足永磁机构故障诊断精度需求。研究结果为永磁机构的故障诊断提供参考。

The permanent magnetic actuator mechanism is widely used in vacuum circuit breakers and plays a crucial role in their performance.Therefore,it is essential to conduct fault diagnosis research on this mechanism.This research focuses on the ZW45-12 type permanent magnetic mechanism vacuum circuit breaker.Firstly,the stroke-time curve of the permanent magnetic mechanism of the vacuum circuit breaker is analyzed using characteristic parameters.Four specific parameters,namely starting time,action time,instantaneous open(closing)speed,and average speed of opening and closing,are selected as diagnostic characteristic parameters.Secondly,various commonly used fault diagnosis algorithms are compared using experimental data of circuit breaker opening and closing,and the SVM algorithm is found to have the best performance.Subsequently,a permanent magnet mechanism fault diagnosis model using the SVM algorithm is constructed based on the Spark platform.The diagnosis model is then optimized by continuously adjusting the penalty parameter C and kernel function of the SVM algorithm.The optimized SVM fault diagnosis model has a diagnostic accuracy of over 90%for the faults of increased circuit resistance of the permanent magnet mechanism,jams of the mechanism and single detachment of the detachment spring,and the accuracy of single detachment of the detachment spring can be up to 96%,which can meet the demand for the fault diagnosis accuracy of the permanent magnet mechanism.The result serves as a reference for the fault diagnosis of the permanent magnet mechanism.