综合智慧能源2024，Vol.46Issue(6)：8-15,8.DOI:10.3969/j.issn.2097-0706.2024.06.002
基于并行融合深度残差收缩网络的有源配电网故障诊断
Research on fault diagnosis of active distribution network based on parallel fusion deep residual shrinkage network
摘要
Abstract
Since the faults of the distribution networks with distributed generators present diversity,and the fault diagnoses are vulnerable to nonlinear factors such as the type of distributed generators and their outputs,a fault diagnosis model based on a parallel fusion deep residual shrinkage network(P-FDRSN)is proposed.The P-FDRSN is constructed by two parallel networks,a fault identification branch and a fault location branch.The parallel structure introduces shrinkage mechanism into its residual module to reduce the influence of noise or redundant information on the network and to improve the robustness of the network against noise.After transforming fault recording signal waveforms into grayscale images and time-frequency images,the signals are fed into the DRSN for deep feature extraction,and then,the acquired features are fused in the convergence layer,so as to enhance the feature learning capability on the fault recording signals.Finally,the simulation results show that the fault location and identification accuracies of the proposed model for various types of distributed generators of different outputs can be maintained above 98.75%and 97.25%,respectively.Even under the interference of noise,the diagnosis accuracy of the model will be kept above 96.75%,showing a high accuracy and decent robustness.关键词
有源配电网/分布式电源/故障诊断/并行网络结构/并行融合深度残差收缩网络Key words
active distribution network/distributed generator/fault diagnosis/parallel network structure/fusion deep residual shrinkage network分类
信息技术与安全科学引用本文复制引用
冯骥,杨国华,史磊,潘欢,陆宇翔,张元曦,李祯..基于并行融合深度残差收缩网络的有源配电网故障诊断[J].综合智慧能源,2024,46(6):8-15,8.基金项目
国家自然科学基金项目(61763040) （61763040）
宁夏自然科学基金项目(2021AAC03062)National Natural Science Foundation of China(61763040) （2021AAC03062）
Ningxia Natural Science Foundation Project(2021AAC03062) （2021AAC03062）
