Calibration and validation of the Angstrom-Prescott model in solar radiation estimation using optimization algorithms

Area of Science:

• Meteorology
• Renewable Energy

Background:

• Accurate solar radiation estimation is crucial for renewable energy applications, especially in data-scarce arid and semi-arid regions.
• The Angstrom-Prescott (A-P) model is a common tool for estimating solar radiation (R_s), but its coefficients often require local calibration.

Purpose of the Study:

• To calibrate and validate the coefficients of the A-P model for solar radiation estimation in Iran's arid and semi-arid regions.
• To evaluate the performance of the A-P model and its improved versions incorporating air temperature and relative humidity.

Main Methods:

• The study involved calibrating and validating A-P model coefficients at six meteorological stations across Iran.
• Optimization algorithms, including Harmony Search (HS) and Shuffled Complex Evolution (SCE), were used for coefficient calibration.
• Performance was assessed using Root Mean Square Error (RMSE), Mean Bias Error, and coefficient of determination (R²).

Main Results:

• The original A-P model demonstrated higher precision and lower error compared to the improved models across all stations.
• The A-P model combined with the SCE algorithm yielded the best performance.
• The RMSE for the A-P model with SCE ranged from 0.82 to 2.67 MJ m^-2 day^-1 during calibration.

Conclusions:

• The standard Angstrom-Prescott model, when calibrated with the SCE algorithm, is superior for estimating solar radiation in arid Iranian regions.
• Optimization algorithms significantly impact the accuracy of solar radiation estimation models.