Solar Power Generation and Predictions During Rapidly Changing Weather Conditions and Rainfall

Date

2023

Authors

Marino, Daniele

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

This comprehensive study merges two crucial aspects of solar energy predictions and generation: the impact of weather data collection time intervals on solar power predictions and the effects of precipitations on solar power generation and predictions, using advanced machine learning (ML) models. The first part of the study focuses on the challenge of predicting solar energy generation during dynamic weather conditions, particularly at sunrise and sunset, periods with high gradients of weather parameters. By fine-tuning the temporal granularity of weather data (?T), we aim to enhance the forecast accuracy of solar power production, especially during these critical times. This is achieved using Random Forest regression models. The study demonstrates that an optimized ?T can significantly improve prediction accuracy, with up 50% improvements for predictions during sunrise and sunset. The second part of the research addresses the impact of precipitations on solar power production and the accuracy of ML models in predicting solar energy output during events with and without precipitations. Utilizing a comprehensive dataset from Texas, this segment investigates the correlation between solar power output and precipitations, employing statistical error metrics like MSE and MAE. Our findings reveal a noticeable decline in energy yield from solar systems during precipitations and underscore the importance of integrating rainfall data as a predictive feature in the models. By doing so, we demonstrate an improvement in the model's forecasting accuracy, emphasizing the need to consider varying weather conditions in predictive modeling for solar energy. Together, these insights contribute significantly to the field of renewable energy, offering strategies for more effective solar energy harnessing and forecasting.

Description

The full text of this item is not available at this time because the author has placed this item under an embargo until June 20, 2024.

Keywords

Machine Learning, Precipitation, Random Forest Regression, Solar Power generation predictions, Weather Impacts

Citation

Department

Mechanical Engineering