| Summary: | The purpose of this study is to enhance the performance analysis of a large-scale solar farm in Eastern Malaysia by employing an Artificial Neural Network (ANN) technique. We created a reliable forecast model for energy output by including crucial environmental data such as temperature, irradiance, wind speed, and humidity. Our dataset, which ranged from December 2022 to December 2023 and included 46,682 five-minute interval records, encountered issues such as missing data owing to PV module failures. We assured data integrity and dependability by doing thorough data pre-processing, such as linear interpolation and outlier management. The ANN model, which used a 15-layer structure and the Levenberg-Marquardt method, obtained low Mean Squared Error (MSE) and high regression values (R = 0.996) during the training, validation, and testing stages, exhibiting predictive accuracy. The findings demonstrate the ANN's usefulness in capturing the underlying patterns of solar power generation, which aligns with Malaysia's renewable energy targets and contributes to global environmental initiatives. This study emphasizes the importance of sophisticated machine learning approaches in optimizing solar farm performance and proposes a framework for continuous monitoring and control systems to ensure operational efficiency. © 2024 IEEE.
|
|---|