A physical downscaling algorithm for the generation of high-resolution spatiotemporal solar irradiance data. (1st March 2021)

• Record Type: Journal Article
• Authors: Buster, Grant; Rossol, Michael; Maclaurin, Galen; Xie, Yu; Sengupta, Manajit
• Is Part Of: Solar energy. Volume 216(2021)
• Page Start: 508
• Page End: 517
• Publication Date: 2021-03-01
• Subjects: Solar radiation -- Solar resource data -- Spatiotemporal data -- Downscaling methodology; Solar energy -- Periodicals; Solar engines -- Periodicals; 621.47
• DOI: 10.1016/j.solener.2021.01.036
• Languages: English
• ISSNs: 0038-092X

Abstract

Highlights: A method for downscaling solar resource data is presented. The downscaled solar resource data is validated against ground measured irradiance. Downscaled photovoltaic performance is compared to historical power generation. A downscaled dataset is presented for Puerto Rico and the U.S. Virgin Islands. Abstract: The National Solar Radiation Database (NSRDB) provides more than 20 years of solar resource information covering North and South America at a 30-min, 4-km by 4-km resolution. However, high-resolution solar resource data is generally required for solar generation models such as the System Advisor Model (SAM), power system models such as the Integrated Grid Modeling System (IGMS), and renewable integration studies. The models and tools require data at a temporal resolution of minutes and spatial resolution of 1–2 km. This study developed a physics-based algorithm to downscale the NSRDB data to higher spatiotemporal resolution for use in various downstream models. The model uses the NSRDB radiative transfer models and is sufficiently generic to downscale solar resource data from the NSRDB without site-specific tuning. The downscaled solar resource data shows comparable accuracy to the native NSRDB when validated against high-temporal-resolution ground measurements. The downscaled solar resource data is also used to generate power production profiles that are compared to high-temporal-resolution photovoltaic plant power generation profiles with good qualitative agreement. Finally, an initial high-resolution publicly accessible data product developed using this method is presented for Puerto Rico and the U.S. Virgin Islands.