Photovoltaic dust accumulation index: Monitoring PV dust with Sentinel-2 satellite imagery

Dust deposition negatively impacts the power generation efficiency and operational stability of photovoltaic (PV) systems. To effectively detect PV dust conditions, this study proposes a novel Photovoltaic Dust Accumulation Index (PDAI). Based on hyperspectral data collected from dust-contaminated PV modules, the spectral response characteristics of PV dust were analyzed. Relevant spectral bands were selected and optimized to construct the index for differentiating varying levels of dust accumulation. The performance of the proposed PDAI was validated using both ground-based measurements and satellite imagery. The results show that at the ground scale, PDAI achieved higher accuracy, with an R2 improvement of approximately 3 % compared to soil- and sand-based indices. At the satellite scale, PDAI also performed better, yielding an increase of about 10 % in overall accuracy (OA) over the traditional indices. Multi-temporal analysis of Sentinel-2 imagery further demonstrated that, compared to conventional sand- and soil-based indices, the PDAI more effectively captures the spatial distribution of PV dust across different installation surfaces. The findings provide technical support for rapid assessment of PV dust and assist in operational and maintenance decision-making for PV power stations.