基于MODWT和ANN的微电网故障诊断

Fault Diagnosis in Microgrid Based on MODWT and ANN Network

近年来,微电网的故障诊断研究变得越来越重要,因为它可以确保微电网安全稳定运行。基于此背景提出了一种基于极大重叠离散小波变换(MODWT)和人工神经网络(ANN)的微电网故障诊断方法。首先利用MODWT对采集到的电流信号进行故障检测,判断是否发生故障;再利用MODWT对故障电流进行特征提取,得到相应的特征向量;随后使用ANN得到故障分类结果。结果表明,该方法可以准确的检测到故障发生时刻。此外,和其他几种智能故障分类方法相比,具有更好的故障分类精度。

In recent years,researches on fault diagnosis of microgrid are becoming more and more important,because it can ensure the stable operating states of microgrid.Under the above-mentioned background,an intelligent fault diag-nosis approach about microgrid is proposed based on maximum overlap discrete wavelet transform(MODWT)and arti-ficial neural network(ANN).Firstly,fault detection algorithm is carried out on collected current by using MODWT,de-termining whether a fault occurs.Then the features of fault current are extracted by MODWT,the features are put into ANN and finally the type is get fault.Result indicates this method can detect fault and its precise time of occurrence.In addition,this method has better fault classification accuracy compared with several intelligent diagnosis techniques.