辽宁工程技术大学学报（自然科学版）2025，Vol.44Issue(3)：358-364,7.DOI:10.11956/j.issn.1008-0562.20240346
基于CNN-LSTM混合模型的风电场集电线路接地故障定位
Grounding fault location of wind farm collector line based on CNN-LSTM hybrid model
摘要
Abstract
Based on the complex operating environment of wind farms and the single-phase grounding fault location requirements of multi-branch hybrid collector lines,a single-phase grounding fault location strategy based on convolutional neural network(CNN)and long short term memory networks(LSTM)hybrid model(CNN-LSTM)is proposed.The zero-sequence current is collected when the fault occurs,and the single-phase grounding fault data set of the wind farm is constructed.The CNN-LSTM hybrid model is improved into a prediction model suitable for fault location,and the model is applied to online fault location.The results show that compared with CNN and backpropagation neural network(BP),the CNN-LSTM hybrid model method has higher positioning accuracy and can be used in different fault distances and fault resistances.The research conclusions provide a reference for the grounding fault location of wind farm collector lines.关键词
故障定位/卷积神经网络/风电场/集电线路/零序电流/长短期记忆网络Key words
fault location/convolutional neural network/wind farms/collector line/zero-sequence current/long short term memory networks分类
信息技术与安全科学引用本文复制引用
刘宝良,张明,高庆忠,张晨,宋阳,程施霖,吴尚润..基于CNN-LSTM混合模型的风电场集电线路接地故障定位[J].辽宁工程技术大学学报（自然科学版）,2025,44(3):358-364,7.基金项目
辽宁省教育厅高校基本科研项目(LJ212411632027) （LJ212411632027）
