基于熵和BP神经网络的单一电能质量扰动分类研究

Study on Classification of Single Power Quality Disturbance Based on Entropy and BP Neural Network

电能扰动信号识别是电能质量研究的重要方向,文中针对8类单一扰动信号,运用Matlab软件生成每类信号200个电能扰动样本(其中100个用于训练,100个作为测试集),提取样本的香农熵和Kolmogorov熵作为特征向量,采用多层前馈神经网络(EBP)作为识别机。仿真分析结果说明:运用人工神经网络作为识别机结合熵特征参数的识别系统对于电能质量扰动信号的识别具有较好的效果。

Identification of power disturbance signal is an important direction of power quality research. In this paper, for 8 kinds of single disturbance signals, 200 power disturbance samples of each type of signal are generated by using Matlab software （100 are used for training, 100 are used as test sets）. Shannon entropy and Kolmogorov entropy are extracted as feature vectors, and multi-layer feedforward neural network （EBP） is used as recognition machine. The simulation results show that using artificial neural network as recognition machine combined with the recognition system of entropy characteristic parameters has a good effect on power quality disturbance signal recognition.