Solar irradiance components estimation based on a low-cost sky-imager

Abstract The design and location of solar energy technologies are based chiefly on the availability of solar resource. Specialized measurement equipment and solar resource assessment techniques are mandatory in order to identify appropriate locations for deploying solar systems while minimizing risk and increasing the return rate of the site. These needs have given the scientific community a motivation to design highly reliable and cost-competitive sensors. This work presents a novel methodology to estimate solar irradiance components using a low-cost sky-imager system. Captured images are processed to correct the spherical aberration of the dome and to achieve a pixel-radiance model, allowing each pixel of the image to be considered as an individual source of irradiation. The performance of the proposed model is analyzed and compared in four different weather scenarios against a secondary standard solarimetric station, obtaining a normalized mean bias error for direct normal irradiance, global horizontal irradiance and diffuse horizontal irradiance of 4.53%, −6.55% and 19.62% for a sunny day, 0.79%, −6.13% and 7.00% for a partially sunny day, 1.96%, −6.58% and −0.73% for a partially cloudy day, 100%, −10.48% and −0.65% for a cloudy day, respectively. According to the experimental results, the system stands out for its performance, easy setup, low-cost components and a robust estimation in all possible weather scenarios.