<![CDATA[Rapid growth in renewable energy adoption and increasing climate variability have created significant challenges for accurately forecasting solar energy generation, as fluctuations in environmental conditions directly influence photovoltaic system performance and energy stability. This study aims to develop a deep learning based forecasting framework capable of improving solar energy prediction accuracy under changing climatic conditions by effectively capturing complex nonlinear relationships among meteorological variables and energy production patterns. The proposed approach utilizes historical solar irradiance, temperature, humidity, wind speed, cloud cover, and atmospheric pressure data as predictive inputs and employs a Long Short-Term Memory (LSTM) deep learning architecture to model temporal dependencies and environmental variability affecting solar energy generation. Experimental evaluation was conducted using publicly available renewable energy datasets, with model performance assessed through Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination (R²). The results demonstrate that the proposed deep learning framework achieved an RMSE of 4.12, an MAE of 3.08, a MAPE of 5.43%, and an R² value of 0.952, indicating strong predictive capability and high agreement between predicted and actual solar energy outputs. These findings suggest that deep learning models can effectively support intelligent renewable energy management by improving forecasting reliability under climate uncertainty, thereby contributing to energy efficiency, grid stability, and sustainable green technology development. ]]>
Copyrights © 2026
