An improved surface solar radiation estimation model using integrated meteorological-air quality data

Abstract

This paper proposes an improved high-precision surface solar radiation estimation model using the integration of the local meteorological data and air quality index based linear regression analysis. The proposed model was evaluated and compared to 8 conventional models and one generated by the commonly used PVsyst simulation software. The actual solar radiation, meteorological data and air quality index collected over 10 years (during 2011-2021) from standard measuring stations located at the northern zone of Thailand were used for developing the models while the collected data year 2022 were used for validating the developed models compared to the conventional models. The statistical error estimations in terms of mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), and mean absolute percentage error (MAPE) were used for the precision evaluation. The study found that the proposed models achieved better prediction results and the highest precision for monthly estimating of solar radiation than the other models by having the highest estimation precision of 94.70-97.19% compared to 87.53-96.74% of the conventional models and 90.38-95.96% of the PVsyst program.