Scholarlink

<p>Accurate estimation of global solar radiation (GSR) is vital for integrating solar energy into power grids and optimizing photovoltaic performance. However, the availability of reliable solar radiation data is often limited in remote or rural areas due to the high cost and complexity of direct measurements. This study presents a comprehensive review of solar radiation prediction models, outlining empirical, statistical, and machine learning approaches such as artificial neural networks, fuzzy logic, and hybrid models. Their strengths and limitations in addressing atmospheric nonlinearity and data scarcity are discussed. Building upon the insights gained from the review, the random forest (RF) machine learning model is employed to predict GSR and assess the solar energy potential across 28 districts in Tamil Nadu, India. The RF model is developed using input parameters such as month number, latitude, longitude, and minimum and maximum temperature, while GSR serves as the output variable. RF model performance is validated using experimental India Meteorological Department data. The predicted and observed values show strong agreement, with a correlation coefficient of 0.9714 and an RMSE of 0.7464 for Chennai. The estimated annual GSR ranges from 17 to 21&#xa0;MJ/m<sup>2</sup>/day, highlighting the region’s significant potential for solar energy development.</p>

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

Advanced modeling techniques for solar radiation estimation: enhancing renewable energy integration in power grids

  • 期刊论文

  • 发表日期：

    2026年1月8日

查看全文
  • R. Meenal,
  • V. Vinothina,
  • B. Sangeetha,
  • K. Vinoth Kumar

摘要

Accurate estimation of global solar radiation (GSR) is vital for integrating solar energy into power grids and optimizing photovoltaic performance. However, the availability of reliable solar radiation data is often limited in remote or rural areas due to the high cost and complexity of direct measurements. This study presents a comprehensive review of solar radiation prediction models, outlining empirical, statistical, and machine learning approaches such as artificial neural networks, fuzzy logic, and hybrid models. Their strengths and limitations in addressing atmospheric nonlinearity and data scarcity are discussed. Building upon the insights gained from the review, the random forest (RF) machine learning model is employed to predict GSR and assess the solar energy potential across 28 districts in Tamil Nadu, India. The RF model is developed using input parameters such as month number, latitude, longitude, and minimum and maximum temperature, while GSR serves as the output variable. RF model performance is validated using experimental India Meteorological Department data. The predicted and observed values show strong agreement, with a correlation coefficient of 0.9714 and an RMSE of 0.7464 for Chennai. The estimated annual GSR ranges from 17 to 21 MJ/m2/day, highlighting the region’s significant potential for solar energy development.