Unraveling the effects of lake-surface environment on floating photovoltaic electricity generation in Southeast Asia

ABSTRACT Floating Photovoltaic (FPV) systems are increasingly utilized for green electricity generation. While lake-surface environment influences FPV Surface Temperatures (FPVSTs) and FPV electricity generation, quantifying its effects on the electricity generation remains challenging, limiting widespread adoption of FPVs in large geographic scales. To address this issue, we propose an estimation framework that integrates four key modules. The first module segments FPV areas from high-resolution satellite imagery by building a U-Net model, enabling precise identification of FPV areas. The second module estimates FPVSTs based on lake environmental variables through machine learning models. The third module computes dynamic photovoltaic conversion efficiency by integrating FPVSTs and solar irradiance, allowing for a detailed assessment of environment-driven performance variations. The fourth module estimates monthly FPV electricity generation, accounting for seasonal changes in lake environments. We reveal considerable performance improvements driven by the water-cooling effects across 270 lakes in Southeast Asia, which can achieve a 2.5% higher annual generation compared to ground-mounted PVs under identical conditions. We identify that water-surface temperature contributes 90.56% to FPVST prediction, while solar irradiance and FPVST account for 44.69% and 39.00% of PVCE variability. This study demonstrates the significant influence of water-cooling effects on FPV electricity generation, significant for planning FPV deployment.