Scholarlink

In this paper, a novel method for forecasting solar photovoltaic (PV) power generation is presented which plays an essential role in grid stability of renewable energy rich power systems. This study utilizes some artificial intelligence (AI) and meta-heuristic approaches such as Random Forest Regressor (RFR), Logistic Regression and Neural Networks, along with Genetic Algorithm (GA). The framework uses the capability of AI models to learn complex relationships in the solar irradiance and weather data. Further, the parameters of these models are optimized by Genetic Algorithm which leads to more accuracy. The study uses various timescales and accuracy metrics such as Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), R-squared, etc. The results show the higher accuracy of hybrid methods than the standalone models of forecasting. This process increases both precision and accuracy. It also provides insights for future research such as investigating ensemble forecasting methodologies to improve solar PV power generation forecasts.

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

Solar PV Power Generation Forecasting Using Synergy of Artificial Intelligence and Metaheuristics Based Approaches

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Pankaj Yadav,
  • Gulshan Sihag,
  • Vikas Pratap Singh,
  • Vivek Vijay

摘要

In this paper, a novel method for forecasting solar photovoltaic (PV) power generation is presented which plays an essential role in grid stability of renewable energy rich power systems. This study utilizes some artificial intelligence (AI) and meta-heuristic approaches such as Random Forest Regressor (RFR), Logistic Regression and Neural Networks, along with Genetic Algorithm (GA). The framework uses the capability of AI models to learn complex relationships in the solar irradiance and weather data. Further, the parameters of these models are optimized by Genetic Algorithm which leads to more accuracy. The study uses various timescales and accuracy metrics such as Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), R-squared, etc. The results show the higher accuracy of hybrid methods than the standalone models of forecasting. This process increases both precision and accuracy. It also provides insights for future research such as investigating ensemble forecasting methodologies to improve solar PV power generation forecasts.