Intelligent Energy Systems: AI-driven Forecasting and Optimization in Hybrid Renewable Energy Systems
Markelov, Daniil (2025)
2025
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2025052616433
https://urn.fi/URN:NBN:fi:amk-2025052616433
Tiivistelmä
This degree thesis investigated the application of predictive Artificial Intelligence models and optimization algorithms to enhance the efficiency, reliability, and sustainability of hybrid renewable energy systems in off-grid locations. Focusing on a real-world hybrid renewable energy system, the research aimed to compare energy production and consumption forecasting approaches and develop a solution for energy system optimization to reduce diesel fuel usage and improve overall system performance.
The study utilized data from the Meteoria visitor center as a specific case study. Machine Learning models, including Random Forest Regressor, Gradient Boosting Regressor, Support Vector Regression, and Neural Networks, were developed and evaluated for forecasting solar energy production, wind energy production, and energy consumption based on historical system data, weather forecasts, and visit schedules. A mixed integer linear programming model was developed using the PuLP Python library to optimize the operation of the energy system.
Results indicated that Random Forest and Gradient Boosting models provided the most accurate forecasts. The integrated forecasting and optimization solution demonstrated significant potential for reducing diesel generator usage and energy waste based on simulations using historical data, highlighting the practical benefits of this approach for off-grid energy system management and contributing to reduced operational costs and environmental impact.
The study utilized data from the Meteoria visitor center as a specific case study. Machine Learning models, including Random Forest Regressor, Gradient Boosting Regressor, Support Vector Regression, and Neural Networks, were developed and evaluated for forecasting solar energy production, wind energy production, and energy consumption based on historical system data, weather forecasts, and visit schedules. A mixed integer linear programming model was developed using the PuLP Python library to optimize the operation of the energy system.
Results indicated that Random Forest and Gradient Boosting models provided the most accurate forecasts. The integrated forecasting and optimization solution demonstrated significant potential for reducing diesel generator usage and energy waste based on simulations using historical data, highlighting the practical benefits of this approach for off-grid energy system management and contributing to reduced operational costs and environmental impact.