Estimation of global solar radiation on horizontal surface in Kano, Nigeria using air temperature amplitude

Abstract

The need for a renewable source of energy is inevitable due to globally uneven distribution of the fossil fuel as well as its associated greenhouse effect. For effective implementation of renewable energy resources in a particular location, its potential availability must be investigated. Solar energy studies of a given geographical location require inputs such as solar radiation profile for the performance estimation and development of various solar technologies. Advance solar radiation measuring equipment such as pyranometer and pyrheliometer are normally installed at the selected locations to collect, measure and output data. However, many developing nations could not afford the cost and maintenance of such equipment. The objective of this work is to calibrate, validate and evaluate and compare four empirical air temperature-based models and choose the most appropriate for estimating the global solar radiation at Kano Airport, Nigeria. The methodology used involved the application of four different empirical models namely the Hargreavees, Allen, Bristow-Campbell and the Samani models by adopting monthly average maximum and minimum air temperatures obtained from meteorological stations as the input parameters. The simulations and analysis were performed in R programing software applications. The analysis shows that average global solar radiation can be precisely estimated using air temperature as the only input parameter to the empirical models at locations where solar radiation data are not obtainable. Using statistical performance analysis to compare the estimated global solar radiation values with the measure solar radiation data. The Samani model with RMSE value of 3.5443, coefficient of determination (R2) value of 0.3508, mean bias error (MBE) of value -3.1506, coefficient residual mass (CRM) of value 0.1448 and mean percentage error (MPE) of value -14.0538 was chosen as the most accurate among the four models. The negative signs in both MBE and MPE indicate non-correlation between the models estimated and observed radiation data. From the simulation and analysis results, simple and non-complex empirical models can be utilized to predict solar radiation with easily obtainable air temperature where sophisticated solar radiation instrument are not readily available. Meanwhile, the accuracy of the estimated solar radiation values of the empirical models using the air temperature can be further improved by using auto- calibration of the models in future works.