Accelerating photovoltaic potential simulations for urban building energy modeling to inform policymakers

Abstract

Current solar analysis and simulation tools are either too optimistic because they ignore neighboring obstructions (e.g. NREL’s PVWatts) or too computationally intensive (e.g. EnergyPlus or other highly-detailed simulations) to provide fast, reliable, and actionable information to decisionmakers at the urban scale. In this paper we propose and test a novel rapid shading calculation algorithm dubbed centroid neighbor shading (CNS) that calculates the rooftop solar photovoltaic (PV) generation potential at city-scales using low-fidelity 2.5D GIS data. It uses state-of-the-art tools (namely PVWatts) and determines a shading mask for sub-hourly timesteps based on the neighboring buildings to correct and more accurately assess the PV generation potential from the rooftops of interest. By striking a balance between GPU-accelerated speed and accuracy, this simulation tool is over 1000 times faster than the current EnergyPlus baseline while maintaining similar accuracy levels. The CNS algorithm has been integrated into a freely-available urban-scale energy analysis website, UBEM.io, to enable government stakeholders and consultants to better assess their potential for carbon emissions reduction.