Solar PV power forecasting using AI tech

• Main Author: Niu, Zhengyi
• Other Authors: Xu Yan
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/167652
• Institution: Nanyang Technological University

With the development of solar technology, solar power forecasting is essential for optimizing the performance of solar energy systems. Many studies used machine learning methods to create models for predicting solar power. Among the algorithms available, XGBoost is popular and useful. In this project, an XGBoost model was developed by Python to forecast solar irradiance. I used weather variables such as temperature and humidity as input features and trained the model using historical solar irradiance data. I evaluated the model performance using several evaluation metrics, including root mean squared error (RMSE) and coefficient of determination (R^2). Results showed that the developed model can accurately forecast hourly global horizontal irradiance for next four days, with an RMSE of 58.572 W/m2, and R^2 of 0.946. Finally, the model was used to predict the GHI of both Shijiazhuang and New York and compared with the predicted values from the Solcast website.