采用BP神经网络的基波高精度检测方法  

A high precision detecting method for fundamental using BP neural network

作　　者：付志红[1] 王好娜[1] 曹敏 王勇[1] 张淮清[1] 

机构地区：[1]重庆大学电气工程学院输配电装备及系统安全与新技术国家重点实验室,重庆400044 [2]云南电力试验研究院(集团)有限公司电力研究院,云南昆明650217

出　　处：《重庆大学学报（自然科学版）》2011年第12期61-66,共6页Journal of Chongqing University

基　　金：重庆市科技攻关计划项目(CSTC;2011AB3003);国家自然科学基金资助项目(40874094)

摘　　要：提出一种基于BP(Back Propagation)神经网络的电网基波频率和幅值的高精度检测方法。正弦信号过零点两侧对称两点连线与时间轴的交点和频率满足单调关系,但并非严格的线性关系,而且与幅值无关,据此用BP神经网络建立该交点与频率的映射关系,并提出了对称点优化选取方法。仿真表明,提出的算法对频率的检测精度达到10-4,幅值的检测精度高达10-5,远远高于普通傅里叶变换(Fast Fourier Transfromation,FFT)算法和FFT加Hanning窗算法;随机噪声和谐波对该方法测量精度的影响很小,具有较强的抗干扰能力。High precision detecting on grid fundamental wave is the basis of energy measurement,power quality assessment and power system automation.This paper presents a high precision detecting method for frequency and amplitude of grid fundamental based on BP neural network.The results show that the intersection of symmetry two points which are between the both sides of signal＇s zeros-crossing has monotone relation with frequency,but no linear relation.The intersection is independent with amplitude.Accordingly,we set the mapping of the intersection and the frequency with the BP neural network,and then propose the optimal selection method of symmetry points.These simulation results show that the precisions of frequency and amplitude are so high to 10-4 and 10-5,which are much higher than FFT and Hanning window algorithms.These influences of random noise and harmonic on this method are very small,so it has strong anti-interference.

关 键 词：BP神经网络 电力系统 基波检测 MATLAB仿真 

分 类 号：TM935[电气工程—电力电子与电力传动]