Building integrated photovoltaics that move beyond rooftops

Abstract

Building-integrated photovoltaics (BIPVs) are essential for urban decarbonization, yet their potential beyond rooftops remains underexplored. Leveraging multi-source heterogeneous data, high-precision shading computation, and machine learning, this study presents a high-accuracy prediction model for multi-dimensional potential indicators. We identify 10.5 billion m 2 of building surfaces across 75 major cities in China, revealing that the rooftops' mean shading ratio is 2%–10%, while facades range from 15% to 50%. These surfaces can capture 5.73 trillion kWh/a of total solar energy, with rooftops and facades each contributing 50%. The BIPV's effective utilization potential reaches 278 billion kWh/a, 140 billion kWh/a more than rooftops alone, sufficient to meet the annual electricity demands of 334 million urban residents. Furthermore, via BIPV-based electricity supply-demand scenarios simulation and model interpretability analysis, we identify key factors for variation in improving energy-matching ratios in various regions, providing critical insights for national BIPV policy and urban energy planning.