基于最优特征量选取的开关柜故障判别方法研究

Research on switchgear fault judgment method based on optimal feature quantity selection

为实现对开关柜运行状态进行快速准确评估,文章提出了基于多类型数据的最优特征量开关柜故障判别方法。基于实时监测的开关柜各类型电气量和非电气参数影响,采取最小冗余最大相关(Minimum Redundancy Maximun Relevance,MRMR)原则对采集的开关柜运行参数进行处理以获取特征样本,并对其进行优化获得最优特征子集;采用马氏距离法对实时监测的运行状态特征量与标准设定样本进行比较,从而判别出开关柜的故障状态。实际测试和算例分析表明,所提出的方法能够有效提高故障识别的精度和效率。

In order to improve the fast and accurate evaluation of the operation status of the switchgear,a switchgear fault identification method based on the optimal feature quantity of multiple types of data is proposed.Based on the real-time monitoring of various types of switchgear electrical quantities and non-electrical parameters,the principle of minimum redundancy maximum relevance(MRMR) is adopted to process the collected switchgear operating parameters to obtain feature samples,and optimize them to obtain the optimal feature subset.The adopted Mahalanobis distance method compares the real-time monitored operating status feature quantity with the standard setting sample to determine the fault state of the switchgear.Practical tests and analysis of calculation examples show that the proposed method can effectively improve the accuracy and efficiency of fault identification.