The study area is located in Gulou District, Nanjing, Jiangsu Province (Figure 2), which lies in a subtropical monsoon climate zone and receives a high annual total solar radiation. With a 100% urbanization rate and a high-density building layout, the district offers ideal conditions for rooftop and facade photovoltaic deployment (Mirkovic and Alawadi 2017). As a core mixed-use commercial and residential area, its continuously growing energy demand, coupled with policy incentives, drives the development of distributed photovoltaic systems (Bensehla, Lazri, and Brito 2021), making it advantageous both in terms of solar resource potential and practical application scenarios.

Meteorological data for the study area were obtained from the XiHe Energy Meteorological Big Data platform (www.xihe-energy.com), which provides hourly meteorological data from 1:00 to 23:00 on 1 August 2023, in Gulou District, Nanjing (specific location: 118.77°E, 32.07°N). The platform not only offers key meteorological variables for the study area but also supplies actual ground-level horizontal radiation data, forming a solid basis for validating the subsequent simulation results. To ensure the accuracy of radiation data, the platform employs the Direct Insolation Simulation Code (DISC) model to estimate direct normal irradiance. This model uses the clearness index in conjunction with solar position, extraterrestrial horizontal irradiance, and air mass parameters to convert global horizontal irradiance into direct normal irradiance. Additionally, the platform incorporates a Global Tilted Irradiance (GTI) model, which comprehensively accounts for direct, diffuse, and reflected radiation, enabling precise simulation of irradiance on photovoltaic array surfaces.

The experimental area (Figure 3) is a multifunctional site with an approximate area of 0.20 k m 2 , comprising multiple buildings and urban green spaces. The experimental data consist of LiDAR point cloud data for the area, containing approximately 4.2 million points, collected using the Lark system. The extraction of rooftop information from buildings involves three main stages: data preprocessing, rooftop surface extraction, and rooftop information retrieval (Figure 1).

The 3-D structure of buildings is characterized from two perspectives: urban block types and three-dimensional structural parameters of the building. Three common urban block types were selected, and seven urban block models were defined accordingly. Based on these models, five three-dimensional structural parameters were defined to refine building model characteristics, thus providing a comprehensive framework for analyzing the influence of building 3-D structural characteristics on photovoltaic potential. In this study, “urban block” is defined as the spatial unit within an urban street layout. The three typical urban block types analyzed are: vertical-type buildings, which consist of independent buildings with a square footprint; central-type buildings, which comprise low-rise buildings around the periphery with high-rise buildings in the center; hybrid-type buildings, which is a combination of buildings of varying heights without enclosed spaces.

These types are examined to assess the impact of building 3-D structure on photovoltaic potential. According to ENVI-met’s parameters, each urban block model is generated in a 10 m × 10 m × 30 m grid (with 10 m spacing representing typical street widths). The height of high-rise buildings is set at 40 m, and low-rise buildings at 20 m, with hybrid-type building heights as illustrated below. All other basic urban planning parameters remain identical to establish a foundation for comparing simulation results under the same planning conditions. The following building three-dimensional structural parameters are defined based on the common urban block types: building height; building spacing, which refers to the parallel distance between buildings; building coverage ratio, representing the percentage of the site area covered by a building; rooftop area ratio, which is the ratio of a building’s roof area to its total surface area; Sky View Factor (calculated from ENVI-met simulations), quantifies the degree of occlusion at a given point on the building surface, based on the ratio of visible sky measured in stereo angles within a specific physical environment.

This is used to characterize the spatial and geometric features of the building. The three-dimensional structural parameters were calculated for each model according to the definitions (Figure 4).

Photovoltaic potential refers to the capacity of a specific area or site to generate electricity using photovoltaic technology, typically expressed as the maximum power output achievable under given conditions. As different photovoltaic technologies vary in conversion efficiency, system design quality also influences the degree to which photovoltaic potential can be realized (Calcabrini et al., 2021; Ebhota and Tabakov 2022; Cristea et al., 2020). In this study, all urban block building models are assumed to use the same photovoltaic components and cover the same total area of photovoltaic modules.

According to the Design Specification for Photovoltaic Power Stations, photovoltaic power for a building can be calculated as follows (Equation 1): E p = Q s w × A × η (1)

Where E p is the photovoltaic power ( kW ); Q sw is the total irradiance on the horizontal plane ( W / m 2 ); A is the total area of photovoltaic modules ( m 2 ); η is the photovoltaic conversion efficiency. According to the Specification Conditions for the Photovoltaic Manufacturing Industry, the average photovoltaic conversion efficiency of polycrystalline silicon photovoltaic panels is no less than 18.4%.

The above formula demonstrates that photovoltaic power is linearly correlated with shortwave solar radiation reaching the ground, allowing the calculation of regional photovoltaic power to be transformed into calculations of shortwave solar radiation reaching the ground. Since the experimental assumptions are based on the same time frame, the linear relationship between photovoltaic power and photovoltaic potential remains consistent throughout the study. Irradiance, a key metric for quantifying radiation, represents the energy received per unit area over time and can assess the solar energy resource richness in a specific area, thus guiding the design and installation of solar energy facilities.

The total radiation received on the ground comprises direct, diffuse, and reflected radiation (Figure 5). Direct radiation is the sunlight reaching the Earth’s surface without atmospheric scattering or absorption, primarily influenced by the solar position and the slope and orientation of the receiving surface. When neighboring objects cast shadows, the target point cannot receive direct solar radiation. Diffuse radiation results from sunlight scattered or deflected by atmospheric particles such as gases, aerosols, or clouds. Different atmospheric conditions significantly affect diffuse radiation. Reflected radiation refers to sunlight reflected off the ground or other surfaces, influenced by the material and 3-D structure of the building.

This study uses ENVI-met, a microclimate simulation tool capable of modeling the exchange of shortwave and longwave radiation in complex urban environments. Its radiation transmission model utilizes attenuation coefficients between 0 and 1 to account for the obstruction of shortwave and longwave radiation by buildings. Shortwave radiation in space considers the effects of direct, diffuse, and reflected radiation. Irradiance is usually expressed in terms of solar radiant power received per unit area. Radiant intensity represents the density of radiant flux radiated outward by a light source. Both of the above variables are expressed in units of. Based on the built-in 3-D radiation transmission model in ENVI-met, the shortwave irradiance at any point in the study area can be expressed as follows (Equation 2): Q s w ⁢ z = σ s w , d i r ⁢ z ⁢ Q s w , d i r 0 + σ s w , d i r ⁢ z ⁢ σ s w , d i f z Q s w , d i f 0 + 1 − σ s v f ⁢ z ⁢ Q s w , d i r 0 · α ¯ (2)

Where Q s w , d i r 0 and Q s w , d i f 0 represent the direct and diffuse irradiance respectively, σ s w , d i r and σ s w , d i f represent their respective attenuation coefficients, α ¯ denotes the average reflectivity of all walls in the study area, the Sky View Factor σ s v f quantifies the visible portion of the sky at a given point, which in turn influences the irradiance received by the building surface. It can be calculated as follows: σ s v f = 1 360 ∑ π = 0 360 cos ⁡ ω π (3) ω represents the maximum cutoff angle in a given direction, which is the largest angle allowed for light to pass through certain media or systems.

This study employed ENVI-met to simulate building surface irradiance, involving four main steps (Figure 6).

Building Model Creation: Based on seven building models under three urban block types, the simulation domain is defined by setting the location, dimensions, building layout, height, spacing, and surface materials. Upon completion, 3D visualization adjustments are applied.

Meteorological Data Input: According to meteorological data provided by the XiHe platform (Table 1), atmospheric parameters, surface conditions, and urban parameters for the study area are configured, along with the simulation time and duration.

Climate Model Configuration: Boundary conditions are defined, and the radiation module (MRT) is selected. The six-direction weighted method is used to calculate mean radiant temperature, and the ACRT (Advanced Canopy Radiation Transfer) module is introduced to more accurately simulate the radiation transmission process. In the building module, the indoor temperature is set to 20°C to reflect the difference from the external surface temperature of 28.5°C under air-conditioning control.

Result Visualization: Simulation outputs are visualized and analyzed using Leonardo and Panoply software.

By comparing surface irradiance results across the full day, it is evident that central-type buildings demonstrate the highest photovoltaic potential, particularly during peak solar hours. The B1 model reaches a peak irradiance of 736.2–754.56 W·m⁻² between 12:00 and 13:00, with the B2 model also showing outstanding performance. These results confirm that although central-type buildings have lower building coverage and rooftop area ratios, their optimized height and spacing configurations reduce shading effects, allowing them to effectively receive radiation during times of maximum solar altitude. Notably, from 11:00 to 14:00, central-type buildings consistently outperform the other types, indicating that their spatial configuration is particularly well-suited to maximizing irradiance during peak sunlight hours.

Vertical-type buildings rank second overall but exhibit unique advantages during specific periods. The A3 model reaches the highest single-point irradiance of the day—805.71 W·m⁻² at 13:00—indicating that vertical layouts also offer high potential at certain times. The data support the idea that uniform building heights in vertical blocks help minimize mutual shading, especially in the early morning (7:00–9:00), when vertical-type buildings—particularly the A2 model—receive more radiation than other types.

Hybrid-type buildings show comparatively weaker performance across most periods of the day, confirming their lowest photovoltaic potential. Although they have the highest building coverage and rooftop area ratios, the peak total irradiance of both C1 and C2 is significantly lower than that of the other two types, especially around noon. This confirms that the staggered building heights typical of hybrid layouts create serious shading problems, particularly where taller buildings obstruct sunlight from reaching shorter ones. The differences in photovoltaic potential across urban block types indicate that each three-dimensional structural parameter influences photovoltaic performance to varying degrees.

Differences in photovoltaic potential across urban block types reveal that each 3-D structural parameter affects photovoltaic potential to varying extents.

Building height has a time-dependent impact on radiation. Under lower solar altitudes, lower building heights help reduce shading—for example, while central-type models perform poorly in the morning, the A3 model receives more radiation than A1 and A2. As solar altitude increases, the effect of building height diminishes, and at noon, the lower vertical-type A3 and moderately high central-type models perform best, reaching 695.48 W/m² and 751.16 W/m², respectively.

Building spacing has a more complex relationship with radiation. A2 outperforms A1 in the morning due to its smaller spacing, which enhances reflected radiation under low solar altitude. However, by 12:00, their irradiance levels are nearly identical, indicating that the impact of spacing weakens under high solar angles. Central-type buildings, with wider spacing, perform best during peak photovoltaic hours, confirming that appropriate spacing reduces shading and enhances solar gain.

Building coverage ratio should be evaluated in conjunction with other structural parameters. A moderately low coverage ratio appears to be optimal—for example, B2 reaches the highest irradiance of 751.16 W/m² at 12:00. In contrast, although hybrid-type buildings have the highest coverage ratios, they exhibit the weakest total irradiance due to excessive shading.

Rooftop area ratio contributes to total irradiance, but this effect is strongly modulated by building arrangement and height variation. At 12:00, A3 receives more radiation than A1 and A2, confirming that rooftop area plays a role in midday peak irradiance. The data suggest that an optimal rooftop area ratio must balance vertical surface area and shading effects. The moderate rooftop area ratios observed in central-type buildings are consistent with their superior performance.

Sky View Factor quantifies the visible portion of the sky at a given point (The formula is expressed as shown in Equation 3), which directly correlates with the extent to which buildings obstruct the sky. The Sky View Factor values for each three-dimensional building model are provided in Table 2, and these results were derived from the built-in model of ENVI-met. The Sky View Factor is highly correlated with direct radiation, and areas with high Sky View Factor exhibit strong performance in total radiation throughout the day. Sky View Factor plays a more critical role under low solar altitude conditions—for instance, vertical-type buildings (A) outperform other types in the morning and afternoon.

Figure 14 compares the simulated and actual total radiation data received by the three urban block building models. Both the models and the measured data exhibit similar diurnal variation patterns—gradually increasing in the morning, peaking at noon, and declining in the afternoon—indicating that the simulation results generally reflect actual radiation trends. At most time points, the simulated values are higher than the measured ones, with the largest discrepancies occurring during the morning hours (6:00–9:00), primarily due to the idealized conditions of the simulation environment. When the solar altitude is at its highest, the simulated and actual data align most closely, as direct radiation dominates during this period and the influence of diffuse radiation is minimal.