Vision Transformers estimate solar irradiance in data-scarce conditions

Researchers have developed a novel dual-framework approach using Vision Transformers (ViTs) to estimate diffuse horizontal irradiance (DHI) from sky images, particularly in data-scarce environments. The method integrates calculated atmospheric parameters and uses a sequential pipeline to predict global horizontal irradiance (GHI) as a feature for refining DHI estimation. Empirical validation using data from the Chilbolton Observatory demonstrated that this approach can achieve root mean square error (RMSE) values comparable to state-of-the-art ViTs trained on complete datasets, even when trained on only a single month of data. AI

IMPACT This research offers a cost-effective solution for solar resource assessment in data-constrained environments, potentially improving photovoltaic system performance and energy forecasting.

RANK_REASON This is a research paper detailing a novel methodology for estimating solar irradiance using AI. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Chilbolton Observatory
• diffuse horizontal irradiance
• global horizontal irradiance
• Vision Transformers

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →