On the use of BRL model for daily and hourly solar radiation components assessment in a semiarid climate

Abstract

Solar radiation components assessment is a highly required activity for solar energy applications. Despite their importance, diffuse and direct solar radiation components are not available in many locations in the world. This is specifically the case in the studied region due to high fiscal demands and technical limitations. In the present work, our main objective is to estimate the two components using global solar radiation components as the only measured input parameter. Machine learning (ML) techniques seem to be a good solution of such an estimation problem but the main issue of ML techniques is the need of long historical data of both desired outputs, direct and diffuse, to build the optimum model in the training phase. However, in the present study, we adopt the use of the Boland–Ridley–Lauret (BRL) model to deal with the problem of estimation of the direct and diffuse components from the global radiation. The adjusted BRL model was applied on two time scales, daily and hourly components. It was found that the estimated direct and diffuse solar radiation values by the adjusted BRL model are in favorable agreement with the measured data.