摘要
电能质量问题成为近年许多高等院校、科研院所的研究重点,电能扰动识别是电能质量研究的一个重要方面。为此,指出了电能扰动识别包括预处理、特征提取和模式识别等3个过程,研究了基于小波变换和人工神经网络的电能扰动模式识别方法。借助于Matlab软件生成120个电能扰动样本并使用小波变换提取特征后,采取反向传播神经网络和概率神经网络识别的正确率分别为87.5%和85%。仿真分析结果发现:使用小波变换提取特征向量并使用反向传播神经网络设计分类器所得到的识别系统的性能比较令人满意。
There are 3 procedures in power disturbance identification such as preprocessing, feature extraction and pattern identification. In this paper ,the basic knowledge of wavelet transform and artificial neural network are introduced and the application of wavelet transform to feature extraction and the application of artificial neural network to pattern identification of power disturbance are studied. Then, simulations of power disturbance identification are carried on via the Matlab software. 120 power disturbance samples are produced, feature extraction is carried on through wavelet transform, pattern identification is carried on through two kinds of neural network which are back propagation neural network (BPNN) and probabilistic Neural Network (PNN). When using the former neural network, the correct rate of identification is 87.5 %, when using the later neural network, the correct rate of identification is 85%. The results of simulation manifest that the ability to identify power disturbance of the identification system is satisfactory, in which feature of the power disturbance is extracted through wavelet transform and pattern identification is carried on through back propagation neural network.
出处
《高电压技术》
EI
CAS
CSCD
北大核心
2007年第7期151-153,181,共4页
High Voltage Engineering
基金
国家自然科学基金(50677044)~~
关键词
电能质量
电能扰动识别
特征提取
模式识别
小波变换
人工神经网络
power quality
power disturbance identification
feature extraction
pattern identification
wavelet transform
artificial neural network