تحسين دقة توقع إنتاج الطاقة الشمسية في ليبيا باستخدام نماذج الانحدار الخطي المتقدمة

Abstract

Solar energy is one of the most important renewable energy sources, significantly contributing to reducing reliance on fossil fuels and minimizing harmful emissions. This research aims to develop an accurate predictive model for solar energy production in the Libyan city of Al-Al_jaylat using machine learning techniques, with a focus on the Linear Regression model. Historical time-series data on solar radiation, temperature, and humidity for the year 2022 were utilized, collected from the NASA POWER platform. After processing and scaling the data, the model was trained and evaluated using performance metrics such as MAE (Mean Absolute Error), MSE (Mean Squared Error), and RMSE (Root Mean Squared Error). The results showed an MAE of 20.44 kWh, indicating the model’s ability to track general production trends. However, significant discrepancies were observed on days with unstable weather conditions, suggesting the need to integrate more advanced techniques, such as neural networks, to improve accuracy. This study provides a practical tool for Libyan energy institutions to enhance solar grid planning and reduce costs associated with inaccurate predictions.