Modeling and Optimization of Photovoltaic Installations at Urban Scale

Abstract

The building sector in developed countries consumes 20% to 40% of the global primary energy, contributing to 30% of the CO 2 emissions. The population growth, particularly in urban areas, is expected to exacerbate this trend, compromising the sustainability goals outlined in international agreements and accelerating climate change. However, the increasing interest and political support in mature renewable energy sources, particularly solar photovoltaic (PV), offers improvement opportunities. The deployment of PV systems in urban areas presents advantages regarding emissions, economic, environmental, and social benefits, enhancing grid efficiency, increasing energy independence, and promoting sustainability awareness and community involvement. To increase the adoption of rooftop PV self-consumption (PVSC) systems in urban environments, studies on the PV potential can help overcome social barriers and enable public administration, utility companies, and private corporations to optimize energy planning, promote PVSC, and attract private investment in clean energy. Although most of the PV potential studies in the literature rely on geospatial physical models, the opportunities provided by Machine Learning (ML) and agile statistical approaches to reduce the high computational cost of the previous models are not fully exploited. These opportunities, especially in the economic potential assessment, remain scarce.