Deep Learning-based Prediction of Renewable Energy Resources in Jebel Akhdar, Oman

The widespread utilization of fossil fuels and associated enormous ecological ramifications have directed attention towards alternate energy sources. Nevertheless, the general adoption of renewable energy resources presents difficulties due to the erratic character of environmental factors such as solar radiation and wind speed. Multiple studies indicate that gradient-boosting-based deep learning forecasting systems are highly regarded for accurate predictions of random variables. On the other hand, Oman presents a favorable environment for advancements in renewable energy, particularly in solar and wind power, owing to its unwavering commitment to reaching carbon neutrality by 2050. Therefore, the main goals of this paper are to improve the accuracy of forecasting climatic factors that impact renewable energy output in the Sultanate of Oman, specifically in Jebel Akhdar, where both wind and solar resources have significant potential. Moreover, the solar power potential in the Manah area and wind power potential in the Jabel Shams region were evaluated by projections of appropriate meteorological variables. This work utilizes the Convolutional Neural Network-based Categorical Boosting (CNN-CatBoost) algorithm to accurately forecast climatic factors for solar and wind energy sources. Results demonstrate the effectiveness of the CNN -CatBoost technique in predicting meteorological elements that influence renewable energy production.