Forecasting Short-Term Photovoltaic Energy Production to Optimize Self-Consumption in Home Systems Based on Real-World Meteorological Data and Machine Learning
This analysis demonstrates improved energy production forecasts in residential systems using meteorological data, suggesting optimized self-consumption strategies.
Key Points
XGBoost achieved the highest forecasting accuracy with a mean absolute error of 1.25 kWh, indicating significant improvements over other models.
The analysis utilized real-world data from a functioning photovoltaic installation and a local weather station to assess model performance.
Classical linear regression, Random Forest, and XGBoost were compared for forecasting short-term energy production in residential photovoltaic systems.
This approach aids energy management in homes without storage, promoting more efficient energy resource usage at a local level.