Ezzini et al. (2025) Optimization of Solar Irradiation Prediction and Recalibration of Weather Stations Using a Hybrid GRU-ANN Model

Identification

• Journal: Lecture notes in networks and systems
• Year: 2025
• Date: 2025-10-07
• Authors: Mustapha Ezzini, Raja Mouach, Mohammed Ennejjar, Abdelali El Gourari, Mohammed Boukendil, Mustapha Raoufi
• DOI: 10.1007/978-3-032-01536-5_72

Research Groups

• LMFE, Physics Department, FSSM, University Cadi Ayyad (UCA), Marrakesh, Morocco
• LAMIGEP, Moroccan School of Engineering (EMSI), Marrakesh, Morocco
• Instrumentation, Physical Signal and Systems, FSSM, University Cadi Ayyad (UCA), Marrakesh, Morocco

Short Summary

This study develops a hybrid GRU-ANN model for accurate solar irradiance prediction and proposes an innovative method for indirect recalibration of weather stations, demonstrating high reliability and potential for optimizing photovoltaic systems.

Objective

• To develop and validate a hybrid GRU-ANN model for accurate solar irradiance prediction.
• To enable the indirect recalibration of weather stations using the model's predictions as a reference, thereby reducing costs and effort associated with physical recalibration.

Study Configuration

• Spatial Scale: Local (implied, specific weather stations and photovoltaic systems)
• Temporal Scale: Historical (unspecified duration)

Methodology and Data

• Models used: Hybrid Gate Recurrent Unit (GRU) and Artificial Neural Network (ANN) model.
• Data sources: Historical photovoltaic (PV) production data, pyranometer voltage measurements.

Main Results

• The hybrid GRU-ANN model achieved high accuracy in predicting solar irradiance with a Root Mean Square Error (RMSE) of 29.5924, a Mean Absolute Percentage Error (MAPE) of 6.68%, and an R-squared (R²) value of 0.9831 during training.
• The model explains 98.31% of the variance in the data, indicating high reliability in solar irradiance prediction.
• Upon implementation, the model demonstrated an R² of 0.96 when compared with pyranometer voltage measurements, confirming its reliable performance in real-world application.
• The model successfully enables the indirect recalibration of weather stations, using its predictions as a reference, which offers a cost-effective and efficient alternative to physical instrument recalibration.

Contributions

• Introduction of an innovative hybrid GRU-ANN model for highly accurate solar irradiance prediction.
• Development of a novel methodology for indirect, cost-effective recalibration of weather stations, eliminating the need for physical instrument adjustments.
• Enhancement of photovoltaic system optimization and energy resource management through improved prediction and recalibration capabilities, supporting the transition to sustainable renewable energy.

Funding

• Not explicitly stated in the provided text.

Citation

@article{Ezzini2025Optimization,
  author = {Ezzini, Mustapha and Mouach, Raja and Ennejjar, Mohammed and Gourari, Abdelali El and Boukendil, Mohammed and Raoufi, Mustapha},
  title = {Optimization of Solar Irradiation Prediction and Recalibration of Weather Stations Using a Hybrid GRU-ANN Model},
  journal = {Lecture notes in networks and systems},
  year = {2025},
  doi = {10.1007/978-3-032-01536-5_72},
  url = {https://doi.org/10.1007/978-3-032-01536-5_72}
}