LSTM-Based Estimation of Solar Energy Production Using Meteorological and Environmental Data: Karabük Case Study.

Area of Science:

• Renewable Energy Systems
• Artificial Intelligence in Energy
• Environmental Science

Background:

• Accurate solar energy forecasting is crucial for grid stability and energy management.
• Traditional forecasting methods often struggle with the complex, variable nature of solar power generation.
• Integrating meteorological and environmental data can improve solar production predictions.

Purpose of the Study:

• To develop and evaluate a Long Short-Term Memory (LSTM) deep learning model for short-, medium-, and long-term solar energy production forecasting.
• To assess the impact of meteorological and environmental variables on forecasting accuracy.
• To validate the model's performance across different time scales and photovoltaic systems.

Main Methods:

• Utilized four years of hourly data from four photovoltaic power plants, including production, meteorological, and environmental variables.
• Implemented a multivariate Long Short-Term Memory (LSTM) deep learning model.
• Evaluated performance using Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and coefficient of determination (R²).

Main Results:

• The LSTM model demonstrated high prediction accuracy, with R² values exceeding 0.90 for seasonal forecasts.
• Incorporating meteorological and environmental data, especially air pollution parameters, significantly improved forecasting performance.
• The model effectively captured long-term production trends across all photovoltaic systems, though monthly forecasts showed more variability.

Conclusions:

• Long Short-Term Memory (LSTM)-based models are reliable and effective for solar energy forecasting.
• Multivariate forecasting incorporating environmental factors enhances prediction accuracy.
• The developed model shows strong potential for applications in energy planning and smart grid management.