一种新的电能质量扰动特征提取与识别方法  

A new features extraction and recognition method for power quality disturbance signals

作　　者：熊建平[1] 陈克绪 马鲁娟[1] 肖露欣[3] 吴建华[3] XIONG Jianping;CHEN Kexu;MA Lujuan;XIAO Louxin;WU Jianhua(Shenzhen Polytechnic,Shenzhen 518055,China;Electric Power Research Institute,State Grid Jiangxi Electric Power Company,Nanchang 330096,China;Information Engineering School,Nanchang University,Nanchang 330031,China)

机构地区：[1]深圳职业技术学院,广东深圳518055 [2]国网江西电力公司电力科学研究院,江西南昌330096 [3]南昌大学信息工程学院,江西南昌330031

出　　处：《现代电子技术》2017年第14期174-177,182,共5页Modern Electronics Technique

基　　金：国家自然科学基金项目(61662047)

摘　　要：为了克服电能质量扰动识别时由于特征选择和提取不当造成最后识别精度低的缺点,提出一种基于数学统计的电能质量扰动幅值采样点数的特征提取方法和PSO-SVM电能质量扰动识别新方法。该方法根据10个周波信号的幅值差异,统计每段幅值范围内的采样点数,对其进行处理后作为各扰动信号的特征,然后采用PSO-SVM分类器对多种扰动信号进行分类识别。该方法特征提取的过程简单,减少了大量的计算处理时间。仿真实验结果表明,该方法能快速地识别出各种扰动信号,且识别精度高于传统方法并具有较好的抗噪声性能。To overcome the shortcomings of low recognition accuracy caused by improper feature selection and extraction inpower quality disturbance recognition,a new feature extraction and recognition method is proposed based on the sample pointsof power quality disturbance amplitude of mathematical statistics and PSO-SVM.According to the amplitude distribution difference over10cycles of signals,the number of samples in amplitude range of each section is calculated,and then used as features of different disturbances after preprocessing.PSO-SVM classifier is used for classification recognition of multiple disturbance signals.The proposed method is simple in the process of feature extraction and efficient in computation.The simulation results show that the proposed method is capable of classifying various disturbance signals at a high speed,and has a higher recognition accuracy and better anti-noise performance in comparison with the traditional methods.

关 键 词：电能质量 数学统计 特征提取 PSO-SVM 

分 类 号：TN911.254-34[电子电信—通信与信息系统] TM76[电子电信—信息与通信工程]