Scholarlink

Predicting photovoltaic power plays a key role in the effective integration of PV systems into the power grid and in improving overall energy production performance. In this study, we assess machine learning (ML) models used for forecasting photovoltaic production. Eleven algorithms were chosen based on their ability to accurately predict the production power. The obtained results reveal a high level of predictive accuracy (R2 > 0.97), confirming the relevance of integrating artificial intelligence (AI) to forecasting productivity and ensuring the reliability of photovoltaic systems. The Linear Regression (LR) model, which is a member of the family of the linear dependence models, had best overall performance (R2 = 0.9922, RMSE = 234.42 W). The boosting-based ensembles, CatBoost (R2 = 0.9913) and XGBoost (R2 = 0.9910), were also great models, which demonstrated ability to capture non-linear dependencies as well. The results of this study support the use of ML methods for the prediction of the PV power generation and provide a basis for future studies.

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

Comparative Evaluation of Machine-Learning Models for Photovoltaic Power Forecasting

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Wissal Sfar,
  • Lahcen Amhaimar,
  • Abderrahim Khalidi

摘要

Predicting photovoltaic power plays a key role in the effective integration of PV systems into the power grid and in improving overall energy production performance. In this study, we assess machine learning (ML) models used for forecasting photovoltaic production. Eleven algorithms were chosen based on their ability to accurately predict the production power. The obtained results reveal a high level of predictive accuracy (R2 > 0.97), confirming the relevance of integrating artificial intelligence (AI) to forecasting productivity and ensuring the reliability of photovoltaic systems. The Linear Regression (LR) model, which is a member of the family of the linear dependence models, had best overall performance (R2 = 0.9922, RMSE = 234.42 W). The boosting-based ensembles, CatBoost (R2 = 0.9913) and XGBoost (R2 = 0.9910), were also great models, which demonstrated ability to capture non-linear dependencies as well. The results of this study support the use of ML methods for the prediction of the PV power generation and provide a basis for future studies.