Numerical methods applied to global solar radiation modeling – comparison with measured data

Abstract

PurposeThe purpose of this paper is to predict the global solar radiation intensity in areas where meteorological stations do not exist and information on solar radiation cannot be obtained experimentally.Design/methodology/approachThe approach to achieve the objective of the paper is developing multiple regression relations between the global solar radiation intensity (the dependent variable) and geographical, geometrical, astronomical and meteorological parameters (the independent variables). The independent variables used for this purpose were selected based on their ease of measurability outside the meteorological station and without expensive equipment. The number of independent variables is arbitrarily chosen and directly affects the accuracy of predictions.FindingsLinear regression relations using one, two, three, four, five, six, and seven independent variables were developed to predict the global solar radiation intensity on horizontal surfaces. An advanced computer program based on least square analysis was used to obtain the regression coefficients. The relations having the highest correlation coefficients were selected. The study shows even when only one independent variable (declination angle) is used, the one variable regression relation predicts the global solar radiation with an accuracy that is satisfactory in most engineering applications.Originality/valueThe diversity of regression relations introduced in this paper gives the engineer such a broad freedom of choice, that knowing only an astronomical parameter of the site makes him capable of estimating the global solar radiation intensity within acceptable margins. The predicted values of global solar radiation intensity by this approach can be used for the design and performance estimation in solar applications. The statistical model developed in this research was validated when compared with the measured data in Yazd airport. The measured data used to analyze the model equations were collected in a 13‐year period. No investigation of this type exists having such degree of accuracy in geographical, geometrical, astronomical and meteorological parameters in Iran.