Deep learning-based solar power generation forecasting

• Main Author: Fei, Siqi
• Other Authors: Xu Yan
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering
• Online Access: https://hdl.handle.net/10356/175516
• Institution: Nanyang Technological University

This dissertation aims at forecasting solar power generation using historical time series data. The deep learning models were adopted for this purpose. The data samples obtained from the website Entso-e Transparency Platform were partitioned and normalized. Based on gradient descent and long short-term memory, the models were constructed and trained to learn the underlying patterns in the training datasets and then make predictions on the testing datasets. After that, evaluations were conducted on the models. To enhance the performance, an investigation into the impacts of hyperparameters was carried out. By adjusting parameters appropriately, both the gradient descent model and the long short-term memory model demonstrated satisfactory performance.