Scholarlink

Accurate solar energy forecasting plays a very important role in incorporating renewable energy into the power grid. This study compares the performance of two Long Short-Term Memory approaches: a single-view model that relies solely on historical energy production and weather data for predictions and a dual-view model that incorporates both historical data and future weather forecasts. Although it is widely acknowledged that the model using future weather data yields superior results compared to the single-view model, we assess the comparative performance of the dual-view model with recorded vs forecasted weather data. Although the results demonstrate a slight degradation in accuracy when using forecasted compared to using actual weather data, the dual-view model still outperformed the model that looks only into the past. This study highlights the importance of incorporating future weather forecasts in solar energy forecasting, even in the presence of forecast inaccuracies.

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

Evaluating LSTM Model Performance for Solar Energy Prediction Using Real vs. Forecasted Exogenous Weather Data

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Alexandru-Adrian Ciobanu,
  • Nebojsa Bacanin,
  • Catalin Stoean

摘要

Accurate solar energy forecasting plays a very important role in incorporating renewable energy into the power grid. This study compares the performance of two Long Short-Term Memory approaches: a single-view model that relies solely on historical energy production and weather data for predictions and a dual-view model that incorporates both historical data and future weather forecasts. Although it is widely acknowledged that the model using future weather data yields superior results compared to the single-view model, we assess the comparative performance of the dual-view model with recorded vs forecasted weather data. Although the results demonstrate a slight degradation in accuracy when using forecasted compared to using actual weather data, the dual-view model still outperformed the model that looks only into the past. This study highlights the importance of incorporating future weather forecasts in solar energy forecasting, even in the presence of forecast inaccuracies.