A method for rapid multi-scenario prediction of urban-scale solar energy substitution rate: A case study in Nanjing

Predicting solar potential and energy consumption of urban plots in complex urban environments have significant value for the efficient and accurate use of solar energy at the early stages of urban design. This paper proposes a multi-scenario solar utilization potential prediction model at an urban scale. Firstly, based on urban morphology clustering analysis, seven categories with a total of 60 sample plots were selected, 3D models were established, and the morphological indicators of the plots were statistically analyzed. Secondly, using Ladybug and Honeybee, combined with Radiance and EnergyPlus, yearly insolation simulation and energy consumption simulation were carried out on the sample plots, and a database of morphological, radiation performance, and energy consumption characteristics was established. Subsequently, machine learning algorithms and data augmentation techniques were employed to train yearly insolation and energy consumption prediction models for the plots, which were then applied to large-scale urban environments to determine the distribution of urban-scale yearly insolation and energy consumption. The accuracy of the yearly insolation prediction model ranges from 0.66 to 0.86, while the accuracy of the energy consumption prediction model ranges between 0.78 and 0.85. Finally, multi-scenario solar utilization substitution rate distribution characteristics were formed based on different levels of solar energy application. In three scenarios—where photovoltaic panels are installed on the roof, on both the roof and façade, and on the roof, façade, and ground—the solar energy utilization substitution rates are achieved at 50 %, 58 %, and 71 %, respectively. The research results provide technical support for urban designers to understand and enhance the degree of solar energy utilization in cities.