Prediction of global solar radiation potential for sustainable and cleaner energy generation using improved Angstrom-Prescott and Gumbel probabilistic models

Abstract

Predicting and improving global solar radiation models is important in Nigeria because, most government meteorological stations are unable to continuously set-up or measure this radiometric parameter in most metropolitan cities and remote villages where there is a severe need for electricity. This is because most locations are not connected to the national grid due to high cost implications. Global solar radiation (H) prediction has a great deal of benefits in adapting and implementing clean and sustainable energy infrastructure, and much more in detecting and adapting climate mitigation measures in locations affected by climate change externalities. The new Gumbel (GP) probabilistic model applied in this paper is renowned for its high prediction power and low input requirements on the coefficients of the Angstrom-Prescott (AP) model used by researchers to predict global solar radiation since 1940. This study's foremost intention was to investigate ways of generating sustainable and cleaner energy by using coefficients of the AP model in order to accelerate the greener economy in Nigeria. As such, the accuracy and suitability of 36 existing and developed empirical AP models, as well as 11 fitted GP models, were evaluated to estimate H in different climatic regions of Nigeria. The results revealed that AP models fitted with generalized datasets outperformed the 10 location-based models considered in this work and the 7 models selected from literature, however, the GP model outperformed all empirically fitted models. The model established by GP (M13) was used to optimize AP models with poor performance and those selected from literature. This study suggests that it is more realistic to apply the generalized and GP model to predict global solar radiation and generate AP estimation coefficients across Nigeria than using empirical modeling and AP estimation coefficient. This is illustrated by the high prediction accuracy, little or no instrumentation network, and the computational effort for a GP model over an instrumentation network required to generate sunshine hour datasets required to implement AP model predictions. The proposed GP model is sufficient as a valid predictive model that will promote a holistic understanding of the available solar resources in Nigeria and disseminate the application of solar photovoltaic technologies within the country.