Rooftop Photovoltaic Potential Assessment Using LiDAR Data: A Case Study of the Corticella District, Bologna

This thesis presents a LiDAR-based methodology for assessing the rooftop photovoltaic potential of the Corticella district in northern Bologna, Italy. The study evaluates 4,784 building footprint polygons using a six-stage pipeline that converts 33 airborne LiDAR tiles into a per-building database of installable roof area, annual irradiation, and PV viability, operating at 0.5 m resolution with a custom shadow-aware radiation model, a slope- and aspect-based suitability filter, and rooftop obstacle detection. The findings suggest that a fixed correction coefficient may significantly underestimate installable rooftop area: in Corticella, the LiDAR-derived estimate of 679,734 m² is approximately double the value obtained with the 0.30 coefficient adopted by Leonforte et al. for a comparable assessment in Milan, exceeding it by 99.4%. The building-specific coefficient averages 0.425 with a coefficient of variation of 51%, showing that no single fixed value can adequately represent a heterogeneous building stock. Among the 2,754 economically viable buildings, incident irradiation on installable rooftops is 925 GWh/year, corresponding to a generation potential of 148 GWh/year — a substantial share of the roughly 424 GWh produced by the Bologna metropolitan area from photovoltaics in 2022. The radiation model was compared against PVGIS-SARAH3 satellite estimates for twenty buildings spanning a wide range of roof conditions: nineteen of twenty agreed within ±5% and all twenty within ±6%; direct single-orientation comparisons agreed within ±3.6%, with the model marginally conservative. Supplementary analyses confirmed that the 0.5 m resolution is adequate for district-scale assessment. The results indicate that municipalities relying on fixed-coefficient cadastral methods may substantially underestimate their rooftop solar resource, with implications for municipal energy planning and rooftop PV prioritisation.