ESTIMATION OF GLOBAL SOLAR RADIATION AT CALABAR USING TWO MODELS

Abstract

In this study, the estimation of global solar radiation with Meteorological parameters at Calabar- Nigeria latitude 4⁰N and longitude 8⁰ E were carried out. The daily mean temperature and relative humidity for seventeen years (1991 to 2007) from Nigerian Meteorological Agency, Federal Ministry of Aviation, Oshodi, Lagos were used. The global solar radiation data were collected courtesy of Renewable Energy for Rural Industrialization and Development in Nigeria. Two models (multiple regression and artificial neural network) were used for the estimation. Comparing the graphs of correlation equation 4 and 5, and equation 6 and 7 of the first model, it is obvious that the first order correlation has better estimation power. Looking at the overview of all the Figures (1 -5A), is it is clear that the two models used in this study has estimation capacity, but Figure 5A shows better correlation with the measured values, which indicates that artificial neural network model is a better model for estimation. Therefore has been recommended for global solar radiation estimation at Calabar and its environs with similar weather condition. Alternatively, first order regression should be use for estimation in the absent of artificial neural network.

Keywords: artificial neural network, regression, model, global solar radiation