基于小波熵和概率神经网络的配电网电压暂降源识别方法  

A Method to Identify Voltage Sag Sources in Distribution Network Based on Wavelet Entropy and Probability Neural Network

作　　者：贾勇[1] 何正友[1] 赵静[1] 

机构地区：[1]西南交通大学电气工程学院,四川省成都市610031

出　　处：《电网技术》2009年第16期63-69,共7页Power System Technology

摘　　要：分析了短路故障、感应电动机启动和变压器投运引起电压暂降的原理及各类电压暂降的特征,提出一种基于小波熵(wavelet entropy,WE)和概率神经网络(probability neural network,PNN)的电压暂降源识别方法。提取信号的小波能谱熵和小波系数熵特征向量,并将其输入概率神经网络,实现电压暂降源的自动识别。利用Matlab/Simulink建立简单配电网的仿真模型进行验证,结果表明,基于小波熵和概率神经网络的方法能很好地识别电压暂降源。On the basis of analyzing the principles and the features of various voltage sags due to power system short-circuit faults, startup of induction motors and energization of power transformers, a method to identify voltage sag sources based on wavelet entropy （WE） and probability neural network （PNN） is proposed. In the proposed method, the eigenvectors of wavelet energy spectrum entropy and wavelet coefficient entropy are extracted and input into PNN to implement automatic identification of voltage sag sources. By use of Matlab/Simulink, a simulation model of simple distribution network is built to verify the proposed method. Simulation results show that the proposed method can identify voltage sag sources well.

关 键 词：电压暂降源 小波熵 概率神经网络 配电网 

分 类 号：TM714[电气工程—电力系统及自动化]