基于KPCA与随机森林的牵引电机局部放电模式识别研究  

Research on Partial Discharge Pattern Recognition of Traction Motor Based on KPCA and Random Forest

作　　者：王卓 郑祥 王仁峰 杨景杰 许智海 WANG Zhuo;ZHENG Xiang;WANG Renfeng;YANG Jingjie;XU Zhihai(School of Automation and Electrical Engineering,Dalian Jiaotong University,Dalian 116028,China)

机构地区：[1]大连交通大学自动化与电气工程学院,辽宁大连116028

出　　处：《电力科学与工程》2022年第9期24-30,共7页Electric Power Science and Engineering

摘　　要：针对牵引电机定子局部放电类型识别过程中,传统局部放电信号特征提取方法存在维数过高、无效信息过多和表征不明显的问题,提出一种基于核主成分分析(kernelprincipal component analysis,KPCA)和随机森林的识别方法。首先,采用集合经验模态分解算法将局部放电信号分解为若干固有模态分量,计算各个分量的分形特征;然后,将传统特征与分形特征结合,采用KPCA算法进行降维,以克服高维特征识别速率慢的缺点;最后,将降维后的特征信息作为随机森林算法的输入,对牵引电机定子的放电类型进行识别。结果表明,该方法识别准确率均超过90%,且速率提升50%以上,具有良好的实际应用价值。Aiming at the problems of too high dimension,too much invalid information and unclear representation in the traditional partial discharge signal feature extraction method in the identification process of partial discharge type of traction motor stator,a new method based on KPCA and random forest algorithm was proposed.Firstly,the partial discharge signal is decomposed into several natural modal components by the ensemble empirical mode decomposition algorithm,and the fractal features of each component are calculated.Then,the traditional features and fractal features are combined,and the KPCA algorithm is used to reduce the dimension to overcome the disadvantage of the slow recognition rate of high-dimensional features.Finally,the feature information after dimensionality reduction is used as the input of the random forest algorithm to identify the discharge type of the traction motor stator.The results show that the recognition accuracy rate of this method exceeds 90%,and the speed is increased by more than 50%,which has good practical application value.

关 键 词：牵引电机 故障检测 局部放电 模式识别 核主成分分析 随机森林 

分 类 号：TM307.1[电气工程—电机] TM303.4