多标签分类法在电能质量复合扰动分类中的应用  

Application of Multi-label Classification Method to Catagorization of Multiple Power Quality Disturbances

作　　者：周雒维[1] 管春[1] 卢伟国[1] 

机构地区：[1]输配电装备及系统安全与新技术国家重点实验室(重庆大学),重庆市沙坪坝区400044

出　　处：《中国电机工程学报》2011年第4期45-50,共6页Proceedings of the CSEE

基　　金：国家自然科学基金项目(50807058)~~

摘　　要：提出一种基于多标签分类的电能质量复合扰动分类新方法。在k–近邻(k-nearest neighbor,KNN)和贝叶斯准则(Bayesian rule)的基础上,提出多标签分类排位分类算法k–近邻贝叶斯多标签分类法(k-nearest neighbor Bayesian rule,KNN-Bayesian)。首先对常见的电能质量扰动及其组合而成的复合扰动进行离散小波分解,提取各层分解系数的规范能量熵作为特征向量;然后,利用KNN-Bayesian进行分类识别。仿真实验结果表明,在不同的噪声条件下KNN-Bayesian可有效分类识别电压暂降、电压暂升、电压短时中断、脉冲暂态、谐波和闪变等电能质量扰动及其组合而成的复合扰动。A new method of identifying the catagory of multiple power quality disturbances based on multi-label classification was presented. A multi-label ranking learning method named k-nearest neighbor Bayesian rule （KNN- Bayesian） was designed based on k-nearest neighbor and Bayesian methods. Firstly, several common power quality disturbances and their compound ones were decomposed by discrete wavelet transform, and the norm energy entropy of the wavelet coefficients of each level were extracted as eigenvectors. And then, the disturbances were classified using KNN-Bayesian. The simulation results show that KNN-Bayesian can recognize the multiple power quality disturbances including voltage sag, voltage swell, interruption, impulsive transient, harmonics, flicker and their compound ones effectively under different disturbance conditions.

关 键 词：电能质量复合扰动 多标签分类 K-近邻 小波变换 贝叶斯准则 

分 类 号：TM72[电气工程—电力系统及自动化]