Influence of Wind Direction on Fire Spread on High-Rise Building Facades

Abstract

In order to study the influence of wind direction on fire spread on building facades, a three-dimensional model of a high-rise building was built with PyroSim software. Numerical simulations were conducted with five wind directions (0°, 45°, 90°, 135° and 180°) and a reference wind speed of 3 m/s. The results show that the fire spread on the building facade was most significant on the leeward side, followed by the upwind condition. The horizontal spread of fire was promoted by crosswinds at 45°, 90° and 135°, while its vertical spread was inhibited. Among these, the inhibitory effect of the 45° crosswind was the greatest. The fire spread area under the leeward condition was the greatest at approximately 341.3 m², accounting for 17.1% of the facade, and the vertical spread velocity under the condition of no wind was the fastest at 0.075 m/s. In this study, the motion characteristics of facade fire spread under different wind directions were determined, providing supporting evidence to enhance fire safety prevention and control measures in high-rise buildings.

1. Introduction

High-rise buildings are densely populated. A fire can cause heavy casualties and property damage, posing a significant threat to people’s safety [1]. For example, in June 2017 [2], a very large fire broke out in a 24-story residential building in London, UK, killing as many as 79 people. In August 2021 [3], the Triumph International Building in Dalian, Liaoning Province, caught fire due to an electrical fault and burned for 7 h. In February 2024, 15 people were killed and 44 injured when an electric bicycle caught fire on the ground floor of a six-story building in a high-rise residential community in Nanjing. In recent years, China’s fire department has enforced several special measures to improve high-rise building fire prevention and made significant progress. In June 2021 [4], the Ministry of Emergency Management of China issued “Regulations on Fire Safety Management of High-rise Civil Buildings”, and, in June 2022 [5], the National Security Committee of the State Council published a notice on carrying out the “Special Rectification of Major Fire Risks in High-rise Buildings”. However, as the scale of high-rise buildings is continuously expanded, their structure and surroundings are becoming increasingly complex, which increases the amount of combustible material used inside and outside these buildings, thus leading to an increase in fire accidents. There are many common causes of fire accidents in our daily lives, such as indoor electric vehicle charging, cluttered hallways or balconies, electrical circuit aging, non-certified insulation materials or inadequate fire protection facilities. Therefore, it is necessary to put into effect high-rise building fire research, so as to enhance people’s emergency response abilities.

Many scholars have carried out extensive research on the influence factors and spread characteristics of fires in high-rise buildings. Through the analysis of fire spread on an isolated building’s facade, Abu-Zidan et al. [6] found that the effect of building geometry on the near-wall flow field altered the path of fire spread. Jeffs et al. [7] presented that the combustion source, ventilation condition and building structure significantly influenced flame spread and that vertical and horizontal spread could be effectively prevented through modifications to the building’s geometric shape or the installation of fire partitions. Pitts [8] analyzed the interaction between downwind and fire in urban large-scale fires and provided a systematic summary of the processes of occurrence, development and spread. Zheng et al. [9] found that the horizontal flame spread rate increased and the vertical rate decreased when the angle between the wind direction and the building wall increased from 0° to 60°. Meroney [10] pointed out that outside winds can alter the infiltration of air through exterior doors and windows, thereby distorting the heat and smoke columns rising above the test flame in the atria, causing the plume to hit the atria wall. Fang et al. [11] investigated the coupling effect between ambient wind and two side walls of the concave shaft. They found that the flame height over the facade increased as the distance between the two side walls decreased, but the former decreased as the wind speed increased. Yuan et al. [12] pointed out that the higher the wind speed, the faster fire smoke spreads in stairwells. Hu et al. [13] found that the height of flames on the facade decreased monotonously with the increase in crosswind speed. Ren et al. [14] investigated the vertical temperature profile of a facade fire plume ejected from compartments with an opening subjected to external wind and found that the temperature at a given height decreased with the increase in wind speed. Li et al. [15] argued that wind could change the fire propagation direction and make the room above the fire more dangerous for people to be in. Bai et al. [16] pointed out that as the wind speed increased, the inclination angle of the flame also increased by using the reduced-scale test and numerical simulation.

In conclusion, some scholars have discussed the influence of wind speed and wind direction on the fire spread characteristics of building facades. However, studies on this subject are relatively sparse, with few systematic conclusions about spread distance, spread area, spread speed, etc. Therefore, in this study, we investigated facade fire spread characteristics using numerical simulation due to its low cost, visualization ability and wide applicability. This paper is organized as follows: Firstly, a model consistent with the field experiment was established using PyroSim software (version 6.7.1) and the experimental data were compared with the simulated results to verify the properties of the building facade materials. Secondly, a simplified rectangular building model was established and a variety of fire scenarios were described, including five wind direction angles (0°, 45°, 90°, 135° and 180°). Finally, the influence of wind direction on fire spread on the building’s facade was analyzed and discussed.

2. Methods

2.1. Mathematical Model

Fire Dynamics Simulator (FDS), a fluid dynamics software developed by the National Institute of Standards and Technology, uses large eddy simulation (LES) to show the process of flame propagation and the path of smoke diffusion in a fire scenario. PyroSim software is developed based on FDS and can simulate temperature, oxygen concentration, smoke concentration and other key parameters during the fire development process. Among many fire simulation software applications, PyroSim is widely used in fire simulation due to its fast model establishment and convenient parameter setting. Therefore, PyroSim was selected in this study to simulate fire in the model [17].

Based on the principle of conservation of mass, momentum and energy, the space geometry of an object is divided into a number of small rectangular control bodies or computation units. On this basis, the numerical solution of the control units is obtained, and the accuracy of the results is verified based on the experimental data [18].

Mass conservation equation:

$${\displaystyle \frac{\partial \rho }{\partial t}}+\nabla \cdot \rho \mathit{u}=0$$

(1)

Momentum conservation equation:

$$\rho \left({\displaystyle \frac{\partial \mathit{u}}{\partial t}}+(\mathit{u}·\nabla )\mathit{u}\right)+\nabla p=\rho g+\mathit{f}+\nabla \tau$$

(2)

Energy conservation equation:

$${\displaystyle \frac{\partial }{\partial t}}\left(\rho h\right)+\nabla \cdot \rho h\mathit{u}={\displaystyle \frac{Dp}{Dt}}+{\dot{q}}^{‴}-{{\dot{q}}^{‴}}_b-\nabla \cdot {\dot{\mathit{q}}}^{″}+\epsilon$$

(3)

where ρ is the density, kg/m³; ∇ is the Hamiltonian operator; u is the velocity vector, m/s; t is the time, s; p is the pressure, Pa; g is the gravitational acceleration, m/s²; f is the external force vector acting on the fluid, kg/(s²·m); τ is the viscous force tensor, kg/(s²·m); h is the specific enthalpy, J/kg; D is the diffusion coefficient, m²/s; ${\dot{q}}^{‴}$ is the heat release rate per unit volume, kW/m³; ${{\dot{q}}^{‴}}_b$ is the energy transmitted to the evaporating droplets, kJ; ${\dot{\mathit{q}}}^{″}$ is the conduction heat and radiant heat, kW/m²; ε is the dissipative function.

2.2. Numerical Model

To ensure the reliability of the numerical model as shown in Figure 1, we construct a numerical model in FDS that is consistent with the experimental model [19]. The material of expanded polystyrene (EPS), because of its insulation, soundproofing and lightweight properties, is widely used in applications for walls, roofs and floors, effectively reducing energy consumption and costs. Its physical parameters are shown in

Table 1. Physical parameters of EPS.

Density (kg·m−3)	Specific Heat Capacity (kJ·kg−1·K−1)	Thermal Conductivity Coefficient (W·m−1·K−1)	Combustion Heat (MJ·kg−1)	Ignition Point (°C)
15	1.34	0.04	46	360

. At the same time, concrete is used to replace polymer cement mortar (PCM) to cover the outer surface of the EPS. The thickness of the concrete is about 0.02 m, and the thickness of the EPS is about 0.18 m.

Numerical simulation involves calibrating by adjusting the combustion characteristics of the materials until the temperature distribution in the numerical model matches the corresponding experimental data. As shown in Figure 2, the test data show that the surface temperature histories of No. 4 vary with test time, and the total duration of the test is 1200 s [19]. It can be found that the overall temperature trends are similar for both methods. Therefore, according to the above facade material model, the main structure of the building was constructed, and tests were conducted to assess how fire spreads along the facades of high-rise buildings under different wind directions.

2.3. Physical Model

A high-rise office building in Zhengzhou, Henan Province, with dimensions of 84 m high, 40 m long and 20 m wide, was taken as an example, as shown in Figure 3a. However, due to limitations in computational capacity, the upper 50 m of the building was selected as the research object, and it was simplified into a rectangular body, as shown in Figure 3b. The fire source was located on the surface of the exterior wall, and its central location was 10 m away from the bottom and about 20 m from the side wall of the building. In addition, 1 temperature slice and 4 temperature measuring points were installed on the front of the EPS exterior wall. The ambient temperature was 20 °C, and the atmospheric pressure was 101.325 KPa.

2.4. Case Design

According to the “Technical Specification for Smoke Control and Exhaust System”, the maximum heat release rate was 6 MW [20], and the reaction was set to “POLYURETHANE” [21,22]. The source of the fire ignited the exterior insulation material, causing the fire to spread. In addition, the fire model of t² was used to describe the fire development process of the high-rise building [23].

$$Q_{max}=\alpha t^2$$

(4)

where Q_max is the maximum heat release rate, kW; α is the growth coefficient, kW/s²; and t is the time for the fire source to reach the maximum heat release rate, s.

The model of fire development includes four types: ultra-fast, fast, medium and slow, which are defined in the relevant literature [24]. In this study, we selected ultra-fast fire as the research object due to the feature of high-rise building fire, and the growth coefficient was set to 0.0469 kW/s². Therefore, the time for the source to reach the maximum heat release rate was 113 s.

2.5. Wind Profile

Generally, when studying high-rise buildings, it is necessary to pay attention to the vertical gradient change in environmental wind. The wind was modeled using the Monin–Obukhov similarity theory [5]. The wind speed was represented in terms of the logarithmic law of Equation (6). According to a survey, the perennial wind speed in the local area was 2 m/s~6 m/s [25]. Therefore, the reference wind speed in this study was set to 3 m/s. The position of the fire source was assumed to be the reference height (z_ref = 10 m). The wind speed was U_ref = 3 m/s at z_ref of 10 m. The open terrain roughness z₀ was selected to be 0.01 m and the von Kármán constant κ was equal to 0.41. The friction velocity u_∗ = 0.178 m/s was calculated. According to the above analysis, the variation rule of the wind speed in the vertical direction is shown in Figure 4.

$$U={\displaystyle \frac{{{\displaystyle \mathrm{u}}}_{\ast }}{\kappa }}\ln \left({\displaystyle \frac{\mathrm{z}}{{\mathrm{z}}_0}}\right)$$

(5)

where U is the velocity at height z m, m; u_∗ is the friction velocity, m; κ = 0.41 is the von Kármán constant; z is the height, m; and z₀ is the aerodynamic roughness, m.

2.6. Grid Setup

Appropriate grid division can save computation time and improve calculation accuracy. The user guide of FDS recommends that the fire characteristic diameter, D* (m), be used to estimate the grid size of the research object, where its reasonable range is 0.0625 D*~0.25 D* [26]. The characteristic diameter (D*) can be defined as [27]

$$D^{\ast }={\left({\displaystyle \frac{Q}{{\rho }_{\infty }{C}_p{T}_{\infty }\sqrt{g}}}\right)}^{{\displaystyle \raisebox{1ex}{$2$}\!\left/ \!\raisebox{-1ex}{$5$}\right.}}$$

(6)

where Q is the total heat release rate, kW; ρ_∞ is the air density, kg/m³; C_p is the specific heat capacity of air at constant pressure, J/(kg·K); and T_∞ is the initial temperature, °C, with a value of 20 °C at room temperature.

The maximum heat release rate was 6 MW and D* was 1.90 m [28]. Therefore, the source location grid size was set as 0.4 m, which was within the acceptable range. Considering the flow characteristics of the wind field, the size of the calculation domain is determined based on the height of the building [6,29]. As shown in Figure 5, the length and width are approximately equal to 4 H, and the height is equal to 2 H. Additionally, to reduce the total number of computational elements, selective refinements are specified near the building surface, while coarser elements are specified elsewhere in the domain.

2.7. Working Conditions

Based on the difference in relative location between the fire source and the incoming wind, the wind direction is classified into four categories: upwind, crosswind, vertical wind and leeward, as shown in Figure 6. In the case of the crosswind working condition, there are two wind direction angles: 45° and 135°. Therefore, five working conditions have been established to analyze the impact of wind direction on fire spread across the building facade. During the whole simulation process, the total simulation time was 720 s and the reference wind speed was 3 m/s, with other parameters unchanged.

As shown in Figure 7, the main parameters discussed include the following:

(1)

Average spread velocity of fire

The average spread speed includes vertical velocity and horizontal velocity and is defined as

$$v={\displaystyle \sum _{i=1}^N(}{\displaystyle \frac{z_i}{t_i}})/N$$

(7)

where v is the average spread velocity of the fire, m/s; z_i is the location of the fire spread at time t_i, m; t_i is the spread time, s; and N represents the number of time points.

(2)

Spread area of fire

This study primarily focused on the high-temperature zone of fire propagation, using the ignition temperature threshold of facade materials (360 °C) as the reference point.

(3)

Spread distance of fire

In the vertical direction, the distance from the fire source to the farthest site with a temperature greater than 360 °C was defined as the vertical spread distance. Similarly, in the horizontal direction, the corresponding distance was defined as the horizontal spread distance.

3. Results and Analysis

3.1. Fire Temperature Distribution

Figure 8 shows the temperature distribution of the facade fire at 180 s, 360 s, 540 s and 720 s. It can be seen that wind direction has a great influence on the spread of fire.

Under the conditions of no wind, upwind and leeward, fire spread was mainly confined to the central part of the building surface, and the affected area of the fire gradually expanded as time progressed. However, compared with the no wind condition, the upwind inhibited the fire’s spread, while the leeward promoted it. Due to the blocking effect of the building, wind speed and wind pressure were lower in the middle part of the building in the leeward position. This led to the accumulation of smoke and heat in the region, thereby intensifying the spread of the fire.

At wind directions of 45°, 90° and 135°, the presence of external wind changed the fire spread direction and reduced the intensity of the fire’s spread, which had a certain inhibitory effect on the development of the fire. On the one hand, the wind direction and the building changed the air pressure distribution around the building, causing the flame to tilt to the right at wind direction angles of 45° and 90° and to the left at a wind direction angle of 135°. On the other hand, the influence of wind flow prevented the accumulation of heat in the surrounding environment and slowed down the spread rate of the fire. Therefore, it can be seen that the inhibiting or promoting effects of wind flow on the development of fire were closely related to the wind direction.

However, Abu-Zidan [6] merely pointed out the feature that the presence of external wind reduced the intensity of facade fire without mentioning its promoting effect. Therefore, understanding this relationship helps us to guide fire prevention and control in high-rise buildings.

3.2. Fire Temperature Variation

Figure 9 presents the fire temperature variation of the measuring points at the heights of 14 m, 26 m, 38 m and 50 m. It can be seen that the temperature rise results were consistent with Figure 8.

Under the conditions of no wind, upwind and leeward, the temperature above the flame was significantly higher than the other three conditions. The main reason lay in the gradual accumulation of the temperature above the flame under the influence of thermal buoyancy, and when it reached the ignition point of the material, it caused the combustion of the material. In addition, the flame temperature at different measuring points finally stabilized at around 1000 °C, and the fire spread the fastest under the condition of no wind. Therefore, it can be said that the wind had a certain effect on the spread rate of the fire, but it did not reduce the temperature of the flame. At the same time, it can also be observed that the flame temperature fluctuated significantly at the height of 14 m in the leeward position. This dramatic fluctuation also played a role in promoting the rapid development of subsequent fires.

However, for the wind direction angles of 45°, 90° and 135°, the flame temperature remained almost constant after the height exceeded 26 m because the wind altered the path of the fire.

3.3. Fire Spread Area

Figure 10 illustrates the fire spread area on the building facade where the wall surface temperature exceeded 360 °C.

Before 180 s, there was no significant change in the fire spread area under each working condition. After 180 s, the fire spread area began to develop rapidly under the conditions of no wind, upwind and leeward. In other cases, however, the rate of fire spread increased slowly. With the continuous development of the fire, the spread area rapidly increased under the leeward condition, and it exceeded other working conditions after 300 s. Until the end of the simulation, its spread area reached about 341.3 m², accounting for 17.1% of the facade area. The fire spread areas under the conditions of no wind and upwind were 200.2 m² and 161.8 m², accounting for 10% and 8.1% of the facade area, respectively. These areas were second only to those under the leeward condition. In comparison, the fire spread areas of the other three conditions were small, with the area under the 45° crosswind condition being the smallest at about 11.5 m², accounting for 0.06% of the facade area. In addition, it can also be observed that the order of fire spread area growth rate was as follows: leeward, no wind, upwind, vertical wind, 135° crosswind and 45° crosswind.

3.4. Fire Spread Velocity

Figure 11 presents the average spread velocity of fire on the building facade. With variations in the wind angle, the fire spread velocity on the building facade presented different characteristics.

In the vertical direction, it first decreased and then increased as the wind angle increased, but the highest fire spread velocity was observed under no wind conditions, reaching 0.075 m/s. However, in the horizontal direction, it first increased and then decreased as the wind direction increased. The highest fire spread velocity was observed under the condition of a 45° crosswind, whereas the lowest velocity was recorded under the condition of an upwind. Moreover, it can also be seen that a 45° crosswind had the strongest inhibitory effect in the vertical direction, and the fire spread velocity in the horizontal direction was 0.024 m/s. However, under the conditions of vertical wind and a 135° crosswind, the horizontal and vertical spread speeds were basically the same. No such conclusion has been found in existing studies.

3.5. Fire Spread Distance

In this analysis, we examined the vertical and horizontal spread distances of the fire on the building facade under various wind angles, including 0°, 45°, 90°, 135° and 180°.

Figure 12 shows the vertical fire spread distance over time. It can be seen that the fire spread to the top of the building under the conditions of no wind, upwind and leeward.

The vertical spread speed under the leeward condition was the highest, with fire reaching the top of the building approximately 60 s earlier than under the upwind condition. On the one hand, the building’s obstruction resulted in a reduced impact on flame spread under the leeward condition, leading to air buoyancy causing the flames to spread rapidly upward along the building facade. On the other hand, the effect of the upwind condition offsets the vertical buoyancy generated by the fire plume, thus reducing the vertical spread speed of the fire and delaying the increase in vertical spread distance. In addition, the vertical spread distance under a 135° crosswind almost coincided with that under vertical wind. The fire spread distance under a 135° crosswind and the vertical wind was larger than that under a 45° crosswind before the fire spread to the top of the building. This indicated that ambient wind could cause a delay in the vertical spread of fire, the effect of which was related to the wind direction. It can be concluded that the impact of the 45° crosswind condition was the greatest, followed by the 135° crosswind and vertical wind conditions, while the smallest impact was that of the no wind, upwind and leeward conditions.

Similarly, Figure 13 shows the horizontal spread distance of fire over time under crosswind and vertical wind. Since the horizontal diffusion distance was not obvious in other working conditions, it was not considered here. Before 120 s, the fire was in an initial burning state with little apparent horizontal spread. After 120 s, the horizontal spread distance of the fire gradually increased and showed a linear growth trend. The horizontal spread speed of fire under 45° crosswind was greater than that under the vertical wind. Moreover, the horizontal spread distance was about 10.3 m in the case of a 45° crosswind by the end of the simulation.

4. Discussion

(1)

Flame propagation was more intense in the leeward condition than in other conditions. Therefore, fire isolation and fire spread control under the leeward condition should be strengthened. For example, an effective fire barrier can be designed on the exterior wall of the building to prevent the spread of fire to other parts of the building or adjacent buildings. In addition, fire-resistant coatings and materials can also be used to strengthen the fire resistance of the building facade, thereby reducing and controlling the spread of fire.

(2)

Under vertical and crosswind conditions, the fire was more likely to spread to the other side of a building’s facade. Therefore, if the wind direction can be detected early, the most suitable side of the building facade for escape can be found.

5. Conclusions

In this study, FDS software was used for numerical simulation of the building fire. The influence of wind direction on fire spread on the building facade was studied, and the main conclusions are as follows:

(1)

The fire spread and temperature rise rates were the most significant on the leeward side of the building and the fire spread area was approximately 341.3 m², which accounted for 17.1% of the facade area after 720 s of the fire developing. At the same time, the fire spread under no wind and upwind conditions were second only to that under leeward conditions.

(2)

The crosswinds at 45°, 90° and 135° inhibited vertical flame spread and promoted horizontal flame spread. In addition, the fire spread area of 45° crosswind is the smallest at about 11.5 m², accounting for 0.06% of the facade area after 720 s of the fire developing.

(3)

The wind direction had a great influence on the vertical and horizontal spread speed of a fire. In the vertical direction, the fire spread speed from fastest to slowest typically varied as follows: no wind, leeward, upwind, vertical wind, 135° crosswind and 45° crosswind. In the horizontal direction, the horizontal spread of fire was more significant under conditions of 5° crosswind, 135° crosswind and vertical wind compared to the other three cases.

In the future, the fire spread characteristics under the coupling effect of wind speed and wind direction will be further systematically studied.