基于SVM的电能质量扰动信号分类方法  

POWER QUALITY DISTURBANCE SIGNAL CLASSIFICATION METHOD BASED ON SVM

作　　者：郭云峰 杨晓梅[1] Guo Yunfeng;Yang Xiaomei(College of Electrical Engineering,Sichuan University,Chengdu 610065,Sichuan,China)

机构地区：[1]四川大学电气工程学院,四川成都610065

出　　处：《计算机应用与软件》2022年第7期95-100,120,共7页Computer Applications and Software

摘　　要：针对电能质量扰动信号分类中存在波形相似、准确率低的问题,提出一种双层支持向量机的分类方法。利用可调Q因子小波变换(tunable Q-factor wavelet transform, TQWT)对信号进行5层分解并提取特征,使用经粒子群算法优化后的支持向量机对扰动信号进行第一次分类;对第一次分类中错误样本集中的类别,结合小波去噪算法和TQWT提取特征;使用优化后的支持向量机对扰动信号进行第二次分类,以提高信号的分类准确率。仿真数据实验表明,所提出的分类方法能够有效识别14类扰动,与其他分类方法相比分类准确率更高,抗噪性更强,具有一定的应用价值。In order to solve the problem of waveform similarity and low accuracy in power quality disturbance signal classification, we propose a two-layer support vector machine(SVM). We used tunable Q-factor wavelet transform(TQWT) to decompose the signal into five layers and extract the features. We adopted SVM which was optimized by particle swarm optimization algorithm to classify the disturbance signal for the first time. For the categories in the error sample set in the first classification, we combined wavelet denoising algorithm with TQWT to extract features, and used the optimized support vector machine to classify the disturbance signal for the second time, so as to improve the classification accuracy of the signal. Simulation data show that the proposed method can effectively identify 14 types of disturbances, and compared with other classification methods, it has higher classification accuracy, stronger noise resistance and has certain application value.

关 键 词：支持向量机 电能质量 可调Q因子小波变换 特征提取 扰动分类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]