摘要
近年来,针对分布式电网故障诊断的研究变得越来越重要,因为它能确保电网安全而稳定地运行。传统的故障诊断方法在故障分类的准确率上仍有待提高,对此,提出了一种基于极大重叠离散小波变换和长短期记忆网络的分布式电网故障诊断方法。首先采集故障数据,再进行特征提取和网络训练,最后得到故障分类的结果。结果表明,该方法不但能准确地识别出故障类型,且不受故障发生时刻和过渡电阻等影响。另外,与其他几种智能诊断方法相比,所提出的方法可以提供更好的故障分类精度。为了评价该方法的性能,以修正的IEEE13总线标准系统为例进行了验证。
In recent years,research on fault diagnosis of distributed power grids has become increasingly important,because it ensures the stable operation of power systems.Traditional fault diagnosis methods still need to improve the accuracy of fault classification.Thus,an intelligent approach for fault diagnosis of distributed power generation systems is proposed based on maximum overlap discrete wavelet transform and long-short term memory network.First,the fault data are collected,then the features are extracted and they are put into the long-short term memory network for training and testing,and finally the results of fault classification are obtained.The results show that the method can classify fault types accurately,and is not affected by the change of electrical parameters.In addition,compared with several existing intelligent diagnosis techniques,the proposed method can provide better fault classification accuracy.In order to evaluate the performance,the algorithm is verified by the case of the modified IEEE13 bus standard system.
出处
《科技创新与应用》
2021年第3期23-26,共4页
Technology Innovation and Application
