Scholarlink

With the rapid development of distributed photovoltaic (PV), distributed PV grid-connected operation safety issues including operation state estimation, safety warning, and fault diagnosis are becoming a hot topic at present, and data panoramic sensing as the research basis of the above content is particularly important. As the actual system may exist in the measurement device layout is not perfect, that is, there may be only part of the node deployment of sensors and other nodes are not deployed sensors, which leads to part of the data not being obtained, thus affecting the overall fault diagnosis and state assessment, as well as can not give timely safety warning. Therefore, this paper proposes a data panorama sensing method based on the TCN-LSTM-SE model, which uses the historical data generated by the simulation to train the neural network to predict the data of the unknown information nodes based on the known information nodes to complete the panorama sensing of the data. Simulation results show that the method has lower prediction error and higher data prediction accuracy than the mainstream network model.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Data Panoramic Perception Method for Distributed PV Systems Based on TCN-LSTM-SE Neural Network

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Wang Hongbiao,
  • Wang Chao,
  • Zeng Shuang,
  • Wang Yaqun,
  • Kang Taifeng,
  • Lan Tian,
  • Wang Jie,
  • Liu Congwei

摘要

With the rapid development of distributed photovoltaic (PV), distributed PV grid-connected operation safety issues including operation state estimation, safety warning, and fault diagnosis are becoming a hot topic at present, and data panoramic sensing as the research basis of the above content is particularly important. As the actual system may exist in the measurement device layout is not perfect, that is, there may be only part of the node deployment of sensors and other nodes are not deployed sensors, which leads to part of the data not being obtained, thus affecting the overall fault diagnosis and state assessment, as well as can not give timely safety warning. Therefore, this paper proposes a data panorama sensing method based on the TCN-LSTM-SE model, which uses the historical data generated by the simulation to train the neural network to predict the data of the unknown information nodes based on the known information nodes to complete the panorama sensing of the data. Simulation results show that the method has lower prediction error and higher data prediction accuracy than the mainstream network model.