Estimating Global Horizontal Irradiance Using Tilted Off-the-Shelf Light Sensors via Reverse Transposition

While sensor networks capable of capturing cloud dynamics have been a solar forecasting research topic for more than a decade, the methods only seem applicable to photovoltaic (PV) panels today. However, there is another product already deployed in commercial products with the potential to form networks: visible light sensors. In this paper, we present initial results from an ongoing study evaluating how well measurements from four differently oriented light sensors align with pyranometer data over a 26-day period. After cosine correction and reverse transposition, all sensors achieve relative RMSEs below 10%, with a close visual correspondence of their curves with the pyranometer measurements. These findings suggest that ubiquitous low-cost sensors may augment solar forecasting systems without additional hardware deployment