A Regression Model to Correct for Intra-Hourly Irradiance Variability Bias in Solar Energy Models
| Autor: | Richard Walker, Kristen Bradford, Mario Ibanez, Dennis A. Moon |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
business.industry
020209 energy Irradiance Regression analysis 02 engineering and technology 021001 nanoscience & nanotechnology Atmospheric sciences Solar energy Random forest Data modeling Solar Resource Assessment methods 0202 electrical engineering electronic engineering information engineering Environmental science 0210 nano-technology business Energy (signal processing) |
| Zdroj: | 2020 47th IEEE Photovoltaic Specialists Conference (PVSC). |
| DOI: | 10.1109/pvsc45281.2020.9300613 |
| Popis: | Industry-standard solar resource assessment methods assume hourly-resolution modeling, which typically overestimates generation due to irradiance variability within an hour. Depending on PV site location and configuration, the high bias introduced by hourly modeling methods is generally greater than 1.5% and can exceed 4% on the annual AC energy when compared to real-world operations. It is critical that bias corrections be applied to hourly solar energy simulations prior to making binding investment and financing decisions. This study presents a random forest regression model that accurately resolves the modeling bias attributed to intra-hour irradiance variability. The model considers site-specific meteorology and layout design parameters to resolve typical seasonal and diurnal variability patterns. It has been validated using minute-resolution observations from operational solar farms and pre-construction meteorological measurements, with model bias error shown to be −0.1% on annual energy. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|