<![CDATA[Abstract Improving the reliability of the power grid system and operational efficiency is essential to facing future energy challenges. This study aims to provide added value to the management of the power grid, especially solar photovoltaic power plants (PLTS), by developing a more accurate predictive model for estimating energy output. By utilizing two real-time data sets, namely weather data and PLTS data, as well as machine learning methods, this study compares the performance of the XGBoost and Linear Regression (LR) models. We built the model to overcome high variability in energy output and maintain the stability of the power grid. The results show that XGBoost has a better performance with an MAE value of 38.08 compared to linear regression, which has an MAE of 80.23, indicating the superiority of XGBoost in predicting PLTS energy output. This study also opens up opportunities for further research with a focus on the application of other models such as random forests and neural networks, as well as improving data quality and parameter optimization to further improve prediction reliability and operational efficiency. The best-performing XGBoost model enables more efficient energy utilization and enhances the operational efficiency of PV solar power plants.]]>
Copyrights © 2025
