Global Solar Radiation of some Regions of Cameroon using the Linear Angstrom Model and Non-linear Polynomial Relations: Part 2, Sun-path Diagrams, Energy Potential Predictions and Statistical Validation

This paper aims to evaluate the global solar radiation of some representative localities of Cameroon using the linear Angstrom model and non-linear polynomial relations. The Angstrom correlation coefficients for the linear, quadratic and cubic polynomial models calculated using the least square method, for the different regions, would be used to predict the global solar energy of the chosen regions. The predictions would be evaluated by some chosen statistical methods. These include: the mean bias error (MBE), the mean relative error (MRE), the root mean square error (RMSE) and t-statistic (t-stat) error, which are the most widely used ones. The input data for the analyses are the measured global solar radiation and mean number of monthly sun shine hours. The data for the analyses is obtained partly from the archives of National Aeronautics and Space Administration (NASA) and partly from the archives of the Cameroon Department of Meteorology, Douala. The data used covers a period of 23 years. We demonstrate the validity of the developed models by comparing the evaluated values of global solar radiation with the measured ones. Hence the regression equations can be used to confidently predict the global solar radiation of the representative regions in the absence of experimental data. We equally implement a high-precision sun-tracking algorithm is necessary to follow the sun’s trajectory from dawn until dusk, in order to maintain high output power and stability of a solar power system.