基于能量差和支持向量机的电网扰动分类识别研究  

Research on power quality feature extraction method based on energy difference and SVM

作　　者：纪萍[1] 陈玲[1] 吴静妹 JI Ping;CHEN Ling;WU Jingmei(Hehai University Wentian College,Ma’anshan,Anhui 243002,China)

机构地区：[1]河海大学文天学院,安徽马鞍山243000

出　　处：《石河子大学学报（自然科学版）》2020年第5期548-553,共6页Journal of Shihezi University(Natural Science)

基　　金：安徽省自然科学基金重点项目(KJ2018A0618)。

摘　　要：针对电网中复合扰动信号难以分类识别的问题,本文提出一种基于能量差和支持向量机的分类识别新方法,利用小波变换和Pasval定理对采集的信号进行多分辨率分解,计算含有信号细节部分的信号能量差,获得新的特征向量,利用支持向量机进行样本训练、识别特征向量和分类识别。在无噪声和含噪声二种情况下对多重扰动进行验证实验,通过对比分析得知此方法的整体分类准确率高达98%,充分验证该方法具有可行性、鲁棒性较强和精度高,为电网质量智能化管理和提高电能质量提供新的理论基础。In this paper,a new classification and recognition method based on energy difference and support vector machine(SVM)is proposed to solve the problem that complex disturbance signals are difficult to be classified and recognized in power grid.Wavelet transform and Pasval theorem are used to decompose the collected signal,and the energy difference of the signal with the detail part of the signal is calculated,and a new eigenvector is obtained.Using wavelet transform and Pasval theorem,multi-resolution decomposition is carried out for the collected signals,and the signal energy difference containing the signal details is calculated to obtain the new feature vector.Support vector machine is used for sample training,feature vector recognition and classification recognition.In the experiment,the multi-disturbances were verified under the two conditions of noise-free and noise-containing.Through comparative analysis,the overall classification accuracy of this method was as high as 98%.he experiment fully proves the feasibility,robustness and high precision of this method,which provides a new theoretical basis for the intelligent management of power network quality and the improvement of power quality.

关 键 词：特征提取 多分辨率分解 Pasval定理 支持向量机 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]