Influence of Building Height Variation on Air Pollution Dispersion in Different Wind Directions: A Numerical Simulation Study

Abstract

Due to the rapid advancement of urbanization, traffic–related pollutants in street canyons have emerged as the primary source of PM_2.5, adversely impacting residents’ health. Therefore, it is necessary to reduce PM_2.5 concentrations. In this study, a three–dimensional steady–state simulation was conducted using Computational Fluid Dynamics (CFD). Three representative wind directions (θ = 0°, 45°, and 90°, corresponding to parallel, oblique, and perpendicular winds) and five different building height ratios (BHR = 0.25, 0.5, 1, 2, and 4) were used to explore the effect of building height variations on PM_2.5 dispersion within street canyons. The results indicated that wind direction significantly influenced PM_2.5 dispersion (p < 0.001). As θ increased (θ = 0°, 45°, and 90°), PM_2.5 concentration in the canyon increased, reaching the most severe pollution under perpendicular wind. Building height variations had a minor impact compared to wind direction, but differences in PM_2.5 concentration were still observed among various BHRs. Specifically, under parallel wind, the influence of BHR on PM_2.5 dispersion was relatively small as compared to oblique and perpendicular winds. For oblique wind, PM_2.5 concentrations varied based on BHR. Street canyons composed of low–rise or multi–story buildings (BHR = 0.25 or 4) slightly increased PM_2.5 concentrations within the canyon, while the lowest PM_2.5 concentration was observed at a BHR of 0.5. Under perpendicular wind, symmetrical (BHR = 1) and step–down canyons (BHR = 2 and 4) exhibited comparable peak concentrations of PM_2.5, whereas step–up canyons (BHR = 0.25 and 0.5) showed relatively lower concentrations.

1. Introduction

In 1950, only 30% of the global population lived in urban areas. By 2018, this proportion had surged to 55%. Projections indicate that this trend of urbanization will continue, with urban populations reaching an estimated 66% of the global population by the year 2050 [1]. China has notably exemplified this trend, undergoing a transformative surge in urbanization from 17.92% in 1978 to 52.27% in 2012, primarily attributed to the transformative policies of Reform and Opening [2]. In urban environments, the street canyon is a common feature defined by buildings flanking both sides of the road [3]. Wind fields induce the formation of vortices in street canyons, leading to a diminished air exchange rate and the entrapment of traffic–related pollutants. This occurrence not only markedly impacts the air quality but also exacerbates environmental concerns, placing significant emphasis on the critical issue of PM_2.5 pollution. Consequently, the associated risks pose potential threats to public health and safety [4,5,6,7]. Research has demonstrated a significant correlation between PM_2.5 and asthma, as well as increased mortality rates [8,9]. The Global Air Quality Guidelines (2021) released by the World Health Organization (WHO) indicated a reduction in the annual target value of PM_2.5 from 10 to 5 μg/m³ [10]. As a result, it is crucial to manage PM_2.5 distribution and decrease concentrations in street canyons.

In May 2021, a joint initiative involving 15 departments, including the Chinese Ministry of Housing and Urban–Rural Development, released a pivotal document titled “Opinions on Strengthening Green and Low–Carbon Construction in County Towns”. The main focus of the document was to strictly regulate the vertical growth of residential structures. It underscored the importance of residential buildings and explicitly mandated that at least 70% of such structures should not exceed six floors. This implies that the majority of buildings in these areas should be low–rise or multi–story buildings. However, whether street canyons composed of low–rise or multi–story buildings can to some extent reduce the concentration of PM_2.5 remains a question.

Generally, it is assumed that variations in building height can either improve or reduce ventilation in street canyons [11]. In most studies, the representation of building height variance commonly involves the ratio between the height of the upwind building (H_u) and the height of the downwind building (H_d), referred to as the building height ratio (BHR, [12]). Street canyons are categorized as step–down street canyons when the building height ratio (BHR) is greater than 1, signifying that the upwind building is higher than the downwind building. Conversely, they are termed step–up canyons when BHR is less than 1, indicating that the upwind building is lower than the downwind building. In most studies on the impact of building height variations on the dispersion of pollutants in street canyons [13,14,15,16,17,18], the variability of building height is limited (0.33 < BHR < 2). In addition, given the background of regulating the vertical growth of residential structures, it is worth studying whether street canyons composed of low–rise or multi–story buildings can facilitate the dispersion of pollutants in the street canyon.

This study aims to enhance this understanding by examining how different building height variations, such as symmetrical, step–up, and step–down street canyons, affect PM_2.5 dispersion in different wind directions (i.e., θ = 0°, 45°, and 90°, corresponding to parallel, oblique, and perpendicular wind directions with respect to the street axis). This research aligns with the scientific initiative launched in May 2021 and seeks to provide valuable insights into mitigating the adverse effects of rapid urbanization on air quality. These insights can potentially inform strategies to optimize air quality within urban environments, addressing a critical aspect of the environmental challenges associated with urbanization.

2. Materials and Methods

2.1. Scenarios

Figure 1 illustrates the reference canyon configuration used in this numerical simulation. The canyon configuration consists of two identical buildings arranged parallel to each other. The street width is twice the building height, and the street length is ten times the building height.

To assess the impact of various building configurations on PM_2.5 dispersion, alterations were made solely to the height of the buildings. This resulted in the creation of two step–up and two step–down street canyons, as outlined in

Table 1. Parameter settings of physical model.

Scenario Types	Street Canyon Types	Model Diagram	BHR (H1/m, H2/m)
Reference Scenario	Symmetrical street canyon		1 (H1 = 18, H2 = 18)
Research scenario	Step–up street canyons		0.25 (H1 = 4.5, H2 = 18)
			0.5 (H1 = 18, H2 = 36)
	Step–down street canyons		2 (H1 = 36, H2 = 18)
			4 (H1 = 18, H2 = 4.5)

. Notably, among these configurations, the cases where the BHR is 0.25 and 4 represent the combination of low–rise and multi–story buildings.

The angle between the direction of incoming wind and street axis (i.e., y–coordinate axis) was defined as θ herein (Figure 1). To examine the effects of different wind directions on PM_2.5 dispersion, three typical wind directions (i.e., θ = 0°, 45°, and 90°, corresponding to a parallel, oblique, and perpendicular wind) were used.

2.2. CFD Model

2.2.1. Computational Domain and Boundary Conditions

The computational domain and boundary conditions were set according to previous studies [19,20,21]. The computational domain adopted a symmetric domain, and the positional relationship between buildings and the computational domain under different wind directions is shown in Figure 2. For the ground boundary condition, a full rough surface avoidance function with a surface roughness of 0.0275 H was employed [22]. The inlet boundary conditions were determined based on the average wind speed profile function $u(z)$ [23]:

$$u(z)=u_H{(\frac{z}{H})}^{\alpha }$$

(1)

where $u(z)$ is the wind speed at height $z$, $u_H=4.7 \mathrm{m}/\mathrm{s}$ is the mean velocity at height H, and the empirical constant $\alpha =0.30$ from the wind tunnel experiment served as a reference [24,25,26].

The turbulent kinetic energy $k$ (m² s⁻²) and turbulent dissipation rate $\epsilon$ (m² s⁻³) at the inlet for initial conditions can be defined as follows:

$$k=\frac{u_{*}{}^2}{\sqrt{C_{\mu }}}(1-\frac{z}{\delta })$$

(2)

$$\epsilon =\frac{u_{*}{}^3}{\kappa z}(1-\frac{z}{\delta })$$

(3)

where $u_{*}=0.52 \mathrm{m}/\mathrm{s}$ is friction wind speed, $\delta$ is the boundary layer thickness (m), in this paper, $\delta =144$, $\kappa$ is the von Karman constant, representing the empirical coefficient introduced by assuming the relationship between mixing length and velocity profile, and $C_{\mu }=0.09$ is the coefficient that determines the turbulence viscosity.

2.2.2. Turbulence Model

The standard $k-\epsilon$ model was applied as a turbulence model. The transmission equation adopted was as follows:

$$\frac{𝜕k}{𝜕t}+\overline{u_j}\frac{𝜕k}{𝜕x_j}=\frac{𝜕}{𝜕x_j}(\frac{v_t}{{\sigma }_k}\frac{𝜕k}{𝜕x_j})+v_t(\frac{𝜕\overline{u_i}}{𝜕x_j}+\frac{𝜕\overline{u_j}}{𝜕x_i})\frac{𝜕\overline{u_i}}{𝜕x_j}-\epsilon$$

(4)

$$\frac{𝜕\epsilon }{𝜕x_j}=\frac{𝜕}{𝜕x_j}(\frac{v_t}{{\sigma }_{\epsilon }}\frac{𝜕\epsilon }{𝜕x_j})+\frac{\epsilon }{k}\left[C_{\epsilon 1}{v}_t(\frac{𝜕\overline{u_i}}{𝜕x_j}+\frac{𝜕\overline{u_j}}{𝜕x_i})\frac{𝜕\overline{u_i}}{𝜕x_j}-C_{\epsilon 2}\epsilon \right]$$

(5)

where $v_t=C_{\mu }\frac{k^2}{\epsilon }$, $C_{\mu }=0.09$, $C_{\epsilon 1}=1.44$, $C_{\epsilon 2}=1.92$, ${\sigma }_k=1.0$, and ${\sigma }_{\epsilon }=1.3$. The simple algorithm was used for pressure–velocity coupling, and the discrete format was set to second–order upwind to minimize numerical diffusion and improve accuracy [27,28,29]. The convergence criteria for all residuals were set to 10⁻⁶. Scalar diffusion was performed using the same number of iterations for all simulations, ensuring consistent conditions throughout the diffusion process [30].

2.2.3. Simulation of Traffic Emissions

Four parallel volumes at street ground–level in the geometry were separated, which were defined as line sources, and were used for simulating traffic emissions on a four–lane road [31], as shown in Figure 1. The four volumes were defined as separate fluid zones; each line source exceeded the street canyon by 0.92 H on each side to simulate the traffic emission of the intersections [32]. The emission rate (Q) for each line source was consistently set at 10 g/s for particulate matter (PM_2.5) in the research cases and for sulfur hexafluoride (SF6) used in wind tunnel experiments for validation cases.

2.2.4. Pollutant Dispersion Model

To calculate pollutant concentration, the advection–diffusion (AD) module was used. In turbulence simulation, the calculation formula of mass diffusion by ANSYS Fluent 19.2 (ANSYS, Inc., Canonsburg, PA, USA) (https://www.ansys.com (accessed on 30 September 2021)) is:

$$J=-({\rho }_{m}D+\frac{{\mu }_t}{S{c}_t})\nabla Y$$

(6)

where ${\rho }_m$ is average density, $D$ is the molecular diffusion coefficient of PM_2.5 in the air, ${\mu }_t$ is turbulent viscosity, $S{c}_t$ is the turbulent Schmidt number, and $Y$ is the mass fraction of PM_2.5. $S{c}_t$ is an important parameter in dispersion simulation and should be optimized based on experimental data. In this study, based on the comparison with the wind tunnel database, the utilized values of $S{c}_t$ were 0.2 for parallel wind, 0.7 for oblique wind, and 0.5 for perpendicular wind.

2.2.5. Standard Simulated Concentrations

To facilitate verification and comparison, all concentrations were expressed in non–dimensional form. The simulated concentration data $c_m$ were normalized as according to the previous study [33]:

$$C^{+}=\frac{c_m{u}_{H}H}{Q_T/L}$$

(7)

where $C^{+}$ is the normalized concentration (–), $c_m$ is the simulated concentration (–), $H$ is the building height (m), and $Q_T/L$ is the release rate per unit length of line sources (kg m⁻¹ s⁻¹).

2.3. Grid–Insensitive Solutions and Model Validation

2.3.1. Grid–Insensitive Solutions

Mesh resolution is critical for accurate computational fluid dynamics simulations. A grid sensitivity study showed that a grid count of 20 cells per building height and 12 cells per street width was sufficient for solutions with acceptable low grid sensitivity [34]. In this study, the grid resolution met these requirements, and the computational domain was built using hexahedral grid cells (Figure 3). Approximately two million unstructured grid cells were used for each case.

2.3.2. Model Validation

Considering the employment of a simplified ideal street canyon model in this study, the validation method encompassed the comparison of simulated data with wind tunnel experiments, following a methodology derived from established studies [7,16,19,20,21,22,23,24,26,31,35,36,37]. The wind tunnel experiment data were obtained from CODASC (http://www.codasc.de (accessed on 23 June 2023)).

The experimental data for WT cases with a street canyon aspect ratio (W/H) of 2 and length–to–height (L/H) of 10, conducted under θ = 0°, 45°, and 90°, were obtained to validate the present numerical model. During the WT experiments, sulfur hexafluoride (SF6) gas was continuously released from line sources. The mean concentrations of SF6 were measured on the canyon walls and presented in terms of the dimensionless pollutant concentration, denoted as $C^{+}$. In Figure 4, the dimensionless pollutant concentration on both the leeward wall (LW) and the windward wall (WW) is depicted. Notably, these walls represent the investigated area where outdoor vehicle particles can easily penetrate indoors through doors and windows [38]. The presentation covers parallel, oblique, and perpendicular wind directions. Given the symmetry of airflow and canyon geometry in parallel wind direction, the profiles are only presented on one wall.

Statistical analysis was conducted to evaluate the overall model performance according to the instruction [39]. Several metrics were adopted, including the fractional deviation (FB), the normalized mean square error (NMSE), the fraction of predictions within a factor of two of observations (FAC2), and the correlation coefficient (R). The requirements for valid data were as follows: −0.3 < FB < 0.3, NMSE < 1.5, FAC2 > 0.5, and R > 0.8 [40]. In

Table 2. Statistical analysis metrics for overall performance.

θ	Wall	Mean Normalized Concentration			Statistical Analysis Metrics
		Numerical Simulation	Wind Tunnel	Relative Difference (%)	FB	NMSE	FAC2	R
0°	LW and WW	1.458	1.5547	6.63	0.064	0.19	0.739	0.874
45°	LW	8.2035	9.8403	−16.67	0.181	0.142	0.817	0.811
	WW	1.1442	0.8734	31	0.268	0.377	0.609	0.801
90°	LW	12.0438	14.9632	−19.51	0.216	0.072	0.853	0.918
	WW	6.2765	5.1357	22.21	0.199	0.104	0.997	0.966

, all of the above metrics were within acceptable criteria. Although when θ = 45°, the relative difference of WW was relatively large, all metrics were within an acceptable range, so this set of data could be accepted. The reason for the significant relative difference may be due to unstable monitoring of the wind tunnel experiment, resulting in some negative values in the wind tunnel database.

Overall, the numerical simulations successfully replicated the distribution of pollutants observed in wind tunnel experiments. Furthermore, the results from the four statistical analysis metrics employed to assess the model’s reliability consistently fell within an acceptable range. As a result, the model demonstrated its suitability for predicting air flow and PM_2.5 dispersion within street canyons.

3. Results

3.1. Airflow Structures in the Canyon

In this study, the y = −18 m, y = 0 m, and y = 18 m planes were defined as sections 1, 2, and 3, respectively, to study the distribution patterns of the flow field. The position of sections is shown in Figure 5.

3.1.1. Parallel Wind

When θ = 0°, in the symmetrical street canyon, the airflow structure in each section was nearly symmetrical as depicted in Figure 6a–c. In asymmetrical street canyons, irrespective of changes in building height, it was observed that the streamlines near higher buildings were denser (Figure 6d–o). Moreover, there was clear evidence of a decrease in wind velocity as it approached the ground, a phenomenon observed in both symmetrical and asymmetrical street canyons.

3.1.2. Oblique Wind

With the exception of BHR = 0.25, the airflow in each section exhibited a clockwise movement, as illustrated in Figure 7a–l. It is noteworthy that at BHR = 2, the size of the clockwise vortex expanded, and its center moved upward (Figure 7d–f). In the case of BHR = 0.25, owing to the low height of the windward building, no clockwise vortex was generated inside the canyon. Instead, two small eddies emerged near the WW and LW (Figure 7m–o). The wind velocity distribution in each section mirrored that of parallel wind.

3.1.3. Perpendicular Wind

Within the symmetrical street canyon, the airflow in each section moved in a clockwise direction (Figure 8a–c). In step–down street canyons, reverse flow occurred (Figure 8d–i), leading to the generation of a counterclockwise vortex inside the canyon. This phenomenon was particularly pronounced in the scenario of BHR = 2 (Figure 8d–f). As the heights of buildings A and B decreased simultaneously, the vortex gradually weakened and disappeared. The step–up street canyons resembled the symmetrical street canyons, with the occurrence of clockwise vortices, and the vortices were larger in the scenario of BHR = 0.5 (Figure 8j–l). When the BHR was 0.25, the vortex inside the canyon dispersed into two small eddies near the WW and LW (Figure 8m–o).

In the perpendicular wind direction, the velocity distribution in each section in symmetrical and step–up street canyons aligned with that of parallel and oblique wind directions, demonstrating a decrease in wind velocity closer to the ground. In step–down street canyons, the velocity distribution in each section did not exhibit a gradual decreasing trend. Due to the influence of reverse flow, the wind speed in the area above building B slightly increased. Additionally, under the influence of perpendicular wind, the wind velocity inside the canyon consistently remained lower than that under parallel and oblique winds.

3.2. PM_2.5 Concentration Distribution

In each scenario, the concentration distribution of PM_2.5 at the WW and LW exhibited an increasing trend as they approached the ground. However, for different building height ratios, there was a significant difference in concentration at the level of pedestrian breathing height (PBH) (Figure 9).

3.2.1. Parallel Wind

With parallel wind influence, PM_2.5 concentrations at the LW, WW, and PBH levels gradually increased from upstream to downstream along the street. Within the symmetrical street canyon, PM_2.5 concentrations at the WW, LW, and PBH levels were symmetrically distributed along the street axis (Figure 9c). In step–down street canyons, with BHR = 2, the normalized mean concentration of PM_2.5 at the PBH level was higher at 7.17, and pedestrians were more likely to be exposed to higher concentrations from the street axis to WW (Figure 9d). When the heights of buildings A and B decreased simultaneously to BHR = 4, the normalized mean concentration of PM_2.5 at the PBH level decreased to 6.77 (Figure 9e). In step–up street canyons, with BHR at 0.5, the concentration distribution on the PBH level was opposite to BHR = 2; that is, the concentration from the street axis to LW was higher than on the other side (Figure 9b). In the case of BHR = 0.25, the normalized mean concentration on the PBH level decreased to 6.77 (Figure 9a).

3.2.2. Oblique Wind

The PM_2.5 concentration at the LW and PBH levels exhibited a gradual increase from upstream to downstream of the street at θ = 45°. In symmetrical and step–down street canyons, the PM_2.5 concentration at the WW was higher near the ground upstream of the street. Specifically, in the symmetrical street canyon, the PM_2.5 concentration increased from the center of the street to the downstream (Figure 9h). In contrast, in step–up street canyons, the PM_2.5 concentration showed an increasing trend from the upstream to the downstream of the street (Figure 9f,g). Concerning PBH levels, pedestrians were more likely to suffer higher concentrations of PM_2.5 in the area between the street axis and the LW in all cases.

3.2.3. Perpendicular Wind

Under the influence of a perpendicular wind, particle concentrations at the WW, LW, and PBH levels of all cases was symmetrically distributed along Section 2. In the symmetrical street canyon, PM_2.5 concentration gradually decreased from the center of the street to both ends (Figure 9m). In step–down street canyons, PM_2.5 concentration at the WW and LW exhibited an opposite trend to the symmetrical canyon, gradually decreasing from both ends to the center of the street. Regarding the PBH level, concentrations were significantly higher near buildings with an H = 18 m (reference building height) (Figure 9n,o), indicating that pedestrians near six–story buildings (18 m) were more likely to be exposed to high PM_2.5 concentrations. In step–up street canyons, PM_2.5 distribution was relatively uniform at the WW, LW, and PBH levels (Figure 9k,l). In terms of PBH level, when BHR was 0.5, PM_2.5 concentration from the street axis to LW was higher than on the other side (Figure 9l). When the heights of buildings A and B simultaneously decreased, pedestrians near the street axis were more susceptible to higher PM_2.5 concentrations (Figure 9k).

4. Discussion

4.1. Flow Regimes and PM_2.5 Distribution

The angle at which wind enters the street canyon is of great importance in determining the distribution of pollutants [41]. Different wind directions result in different flow regimes within the canyon, such as the formation of vortices, eddies, or recirculation zones. These flow patterns significantly influence how pollutants disperse and distribute within the canyon.

4.1.1. Parallel Wind

In the symmetrical street canyon, the predominant flow regime is characterized by a parallel flow along the street axis. This flow pattern leads to the accumulation of traffic pollutants, resulting in a consistent increase in PM_2.5 concentration along the street axis, with the highest concentration occurring at the pedestrian level [42].

In asymmetric street canyons, due to the height difference between buildings, the airflow accelerates near higher buildings when passing through the canyon. In the parallel wind direction specifically, the streamlines near the taller buildings on the vertical central plane were denser, thereby promoting the dispersion of particulate matter. The movement pattern of this airflow velocity results in relatively lower concentrations of PM_2.5 near taller buildings in asymmetric street canyons.

4.1.2. Oblique Wind

In the symmetrical street canyon, corkscrew flow is generated by the superposition of components parallel and perpendicular to the street axis [43]. In this flow state, particulate matter is influenced by parallel wind direction components, being transported and accumulated along the street axis; simultaneously, particulate matter is transported and accumulated at the LW under the influence of perpendicular wind direction components. Consequently, the PM_2.5 concentration downstream of the street is higher than upstream, and the concentration at the LW is higher than at the WW [44].

In asymmetrical street canyons, the concentration of particulate matter at the LW is higher than at the WW. This phenomenon is mainly due to the presence of clockwise spiral vortices in most canyons (except for at a BHR of 0.25), which transport particles near the ground to the WW. In the canyon with a height ratio of 0.25, the small eddy near the LW transports particulate matter to the LW, resulting in a higher concentration of PM_2.5 at the LW.

4.1.3. Perpendicular Wind

In the symmetrical street canyon, buildings on both sides impede the airflow, creating unfavorable conditions for PM_2.5 dispersion and escape [45,46]. The flow regime involves a horizontal canyon vortex in the central area of the street and corner eddies at the street ends. The canyon vortex forces air downward from the roof of the building towards the WW, diluting the concentration of particulate matter at the WW, while also transporting particulate matter to the LW [47]. The corner eddies push air inward from both ends of the street, diluting the PM_2.5 concentration on both sides of the street. Therefore, the PM_2.5 concentration at the WW and LW shows a gradually decreasing trend from the center of the street to both ends of the street. The canyon vortex acts as the main source of air exchange in the central part of the canyon, while the combination of the canyon vortex and corner eddies enhances ventilation at the street ends.

In step–down street canyons, when BHR = 2, the vortex regime inside the canyon aligns with previous research [37], exhibiting a clockwise vortex and causing particles to re–enter the street and increase the concentration at the WW. When the heights of buildings A and B decrease simultaneously and BHR increases to 4, the vortex inside the canyon disappears, and the reverse flow directly transports PM_2.5 to the LW. In the step–up canyon with a BHR of 0.5, a single clockwise vortex appears, transporting PM_2.5 towards the LW. In the case of BHR = 0.25, the small eddy near the LW transports PM_2.5 to the LW, resulting in a higher concentration of PM_2.5 at the LW than at the WW.

4.2. The Synergy between Wind Direction and Building Height Variation

Overall, irrespective of the BHR, the total mean concentration of PM_2.5 within the canyon increased as θ increased (θ = 0°, 45°, and 90°), indicating that the wind direction played a more significant role than variations in building height (Figure 10). Additionally, the Two–Way ANOVA analysis of wind direction and BHR on PM_2.5 concentration revealed a significant positive correlation (p < 0.001) between wind direction and PM_2.5 concentration, whereas the correlation between BHR and PM_2.5 concentration was not very significant (p = 0.348).

Figure 10 illustrates the synergistic effect of wind direction and BHRs on PM_2.5 dispersion. Under parallel wind directions, the total mean concentration of PM_2.5 within the canyon remains consistently low in all cases, slightly exceeding 3. Compared to oblique and perpendicular winds, building height variations under parallel wind directions have a lower impact on the dispersion of PM_2.5. Under oblique winds, the BHR affects the total mean concentration of PM_2.5. For street canyons composed of low–rise or multi–story buildings (BHR = 0.25 or 4), the total mean concentration of PM_2.5 within the canyon slightly increases compared to the symmetrical canyon. With BHR values of 2 or 0.5, there is a decrease in total mean concentration of PM_2.5, with BHR = 0.5 exhibiting the lowest concentration in the canyon. This suggests that when the BHR is 2 or 0.5, the atmosphere is more easily mixed in the canyon, promoting the dispersion of PM_2.5. Under perpendicular wind direction, regardless of variations in building height, the total mean concentration of PM_2.5 within the canyon is higher than that of parallel and oblique winds (Figure 10), indicating that perpendicular wind direction is the least favorable for PM_2.5 dispersion, consistent with the research results of Di Sabatino et al. (2008). Specifically, the study found that the total mean concentration of PM_2.5 in symmetrical and step–down canyons (BHR = 2 and 4) is similar, with both forming concentration peaks within the canyon, while the total mean concentration of PM_2.5 in step–up street canyons (BHR = 0.25 and 0.5) is significantly lower. This indicates that both symmetrical and step–down canyons are not conducive to PM_2.5 dispersion in the perpendicular wind direction.

The research findings underscore the significance of wind direction and building height variations as crucial reference factors in urban planning. When considering wind direction as an influencing factor, the observed lowest PM_2.5 concentration in the canyon occurred under parallel wind directions, an observation that holds valuable insights for urban strategies; that is, in the process of urban planning, we can try to align the orientation of street canyons parallel to the dominant wind direction as much as possible. This approach can enhance ventilation and subsequently improve air quality within the canyon. On the other hand, when considering the impact of building height variations, in cities where wind flows parallel to the street axis, the influence of BHRs on PM_2.5 dispersion is comparatively small compared with oblique and perpendicular wind directions, affording urban planners greater flexibility in determining building height. Under oblique wind directions, canyons with low–rise or multi–story buildings may result in an increase of PM_2.5 concentration compared to symmetrical street canyons. Conversely, street canyons with a BHR of 0.5 help optimize airflow and promote the dispersion of PM_2.5 In perpendicular winds, step–up street canyons demonstrate superior dispersion of PM_2.5 compared to both symmetrical and step–down canyons. The impact of street canyons composed of low–rise or multi–story buildings (BHR = 0.25 or 4) on the total concentrations of PM_2.5 within the canyons depends on the height of the windward building (i.e., Building A). Notably, when the windward building’s height is small (BHR = 0.25), the concentration within the canyon is the lowest.

This article primarily explored the influence of variations in building height on the dispersion of PM_2.5 in ideal street canyons under different wind directions. We employed a simplified ideal street canyon model and affirmed its reliability through the comparison of simulated data with wind tunnel experimental data obtained from CODASC. Recognizing the irregular shapes of actual buildings and the intricate interplay of various factors influencing PM_2.5 dispersion in field measurements, future research aims to establish numerical models based on on–site measurements. Other factors, such as the aspect ratio of the street canyon, vegetation, and roof shapes, could also be taken into account simultaneously in further studies. Real street canyons exhibit a diverse array of features, and incorporating these nuances into simulation will provide a more realistic representation of pollutant dispersion in urban environments, allowing for a comprehensive assessment of the model’s performance. Additionally, as transient analysis can offer more detailed temporal evolution information, using transient analysis methods to explore the dispersion patterns of pollutants in street canyons is also an important aspect of future work.

5. Conclusions

This study used Computational Fluid Dynamics (CFD) for three–dimensional steady–state numerical simulation in order to investigate the impact of building height variation on air pollution dispersion under different wind directions.

The influence of building height variations on PM_2.5 dispersion varies under different wind directions. Specifically, in perpendicular winds, compared to symmetrical and step–down canyons, step–up canyons exhibit better performance in promoting PM_2.5 dispersion. For canyons composed of low–rise or multi–story buildings (BHR = 0.25 or 4), the total concentrations of PM_2.5 depends on the height of windward building, and when the windward building’s height is small (BHR = 0.25), the concentration within the canyon is the lowest. In oblique wind directions, street canyons with low–rise or multi–story buildings (BHR = 0.25 or 4) slightly increase total mean concentration, and the lowest concentration of PM_2.5 was observed at BHR = 0.5. In parallel wind directions, the impact of BHR on PM_2.5 dispersion is relatively small compared to oblique and perpendicular winds.