Scholarlink

The output of distributed photovoltaic (PV) systems is influenced by factors like weather and illumination, which show significant spatial and temporal correlations. Accurately forecasting PV output requires a comprehensive approach that considers both temporal dynamics and geographical factors. Traditional forecasting methods focus on temporal data in Euclidean spaces, often missing spatial characteristics in non-Euclidean graph structures. To address these challenges, we propose a novel forecasting methodology using a temporal graph convolutional network (TGCN) to analyze spatiotemporal correlations among distributed PV systems. This paper applies the method to photovoltaic data from Guangde City, China, and compares its performance with existing techniques to demonstrate improved accuracy.

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

Graph Convolutional Neural Networks for Forecasting Spatial Generation in Distributed Photovoltaic Systems with Temporal-Spatial Correlation

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Peng Fu,
  • Wei Jia,
  • Lei Ye,
  • Yuanyuan Chen

摘要

The output of distributed photovoltaic (PV) systems is influenced by factors like weather and illumination, which show significant spatial and temporal correlations. Accurately forecasting PV output requires a comprehensive approach that considers both temporal dynamics and geographical factors. Traditional forecasting methods focus on temporal data in Euclidean spaces, often missing spatial characteristics in non-Euclidean graph structures. To address these challenges, we propose a novel forecasting methodology using a temporal graph convolutional network (TGCN) to analyze spatiotemporal correlations among distributed PV systems. This paper applies the method to photovoltaic data from Guangde City, China, and compares its performance with existing techniques to demonstrate improved accuracy.