Numerical Evaluation of a Novel Vertical Drop Airflow System to Mitigate Droplet Transmission in Trains

Abstract

Owing to the outbreak of COVID-19, researchers are exploring methods to prevent contact and non-contact infections that occur via multiple transmission routes. However, studies on preventing infections caused by droplet transmission in public transportation are insufficient. To prevent the spread of infectious diseases, a new ventilation system in railway vehicles must be developed. In this study, a novel vertical drop airflow (VDA) system is proposed to mitigate the effect of droplet transmission in a high-speed train cabin. The droplet transmission route and droplet fate are investigated using three-dimensional fluid dynamics simulations, performed employing the Eulerian–Lagrangian model. Additionally, a porous model is adopted to simulate the effect of close-fitting masks. The results indicate that 120 s after coughing, the decrease in the droplet number in the VDA system is 72.1% of that observed in the conventional system. Moreover, the VDA system effectively suppresses droplet transmission because the maximum droplet travel distances of the VDA systems are 49.9% to 67.0% of those of the conventional systems. Furthermore, the effect of reducing droplet transmission by wearing a close-fitting mask is confirmed in all systems. Thus, the decrease in both droplet number and droplet transmission area in train cabins validate that the proposed VDA system has an effective airflow design to prevent droplet infection.

1. Introduction

COVID-19 has infected millions of people worldwide since its inception in December 2019 [1]. This pandemic has resulted in public panic and economic losses on a global scale [2]. COVID-19 can be transmitted via contact and non-contact routes [3]. Non-contact infection occurs when virus-laden droplets from an infected person spread through the air [4]. Moreover, COVID-19 can survive in droplet form for 3 h in indoor environmental conditions [5]. Therefore, the analysis of transmission routes of droplets is essential to eliminate them in a closed space with numerous people.

Public transport is vulnerable to droplet infections because passengers stay in confined spaces for several hours. Furthermore, COVID-19 droplets can travel through the air for a distance of over 2 m as the air is circulated in the heating, ventilation, and air-conditioning system inside the cabin [6]. Since 2020, cases of infection have been reported in various public transportation systems. On a cruise ship, approximately 700 cases of COVID-19 were laboratory-confirmed among passengers and crew [7]. In an airplane, laboratory-confirmed COVID-19 was detected in 12 patients after a five-hour flight [8]. Based on epidemiological evidence obtained from outbreak investigations, one study reported the possibility of droplet transmission of COVID-19 in a bus [9]. The epidemiological investigations of COVID-19 cases and their close contacts determined that the infection risk is high when passengers are seated within a distance of three rows and five columns from the patient in a train [10]. Therefore, extensive studies on droplet transmission in public transportation are required to ensure passenger safety.

In train ventilation systems, various airflow designs such as the ceiling inlet/ceiling outlet, ceiling inlet/floor outlet, and sidewall inlet/side wall outlet systems have been developed for increasing thermal comfort and spatial homogeneity [11]. However, these designs, which are not based on droplet transmission analysis, possess the risk of the droplet spread in the cabin. Woodward et al. [12] investigated experimentally that air mixing occurred well in the height and width direction in the intercity train of the ceiling inlet/outlet system. Furthermore, Wang et al. [13] reported that the maximum infection probability in Chinese long-distance trains with the ceiling inlet/side wall outlet system was 10.3%. In airplane ventilation systems, air enters the cabin from the overhead inlet and flows downward toward floor outlets at the sidewall. This could reduce air mixing between the front and rear seats. However, Yan et al. [14] reported that the droplet transmission in the window-seat passenger spread widely, which could lead to high exposure risks to nearby passengers. Therefore, to prevent droplet transmission in public transportation, a novel ventilation system of the vertical drop airflow (VDA) design is proposed in this study. The proposed VDA system was designed to exhaust the droplets expelled from the infected person to the bottom of the cabin area through a strong downstream airflow. Furthermore, the inlets and outlets were located properly for each seat such that the droplet travel distance was reduced.

Computational fluid dynamics (CFD) is actively used in the analysis of droplet transmission routes so they can be visualized [15]. Several studies have analyzed droplet transmission routes in restaurants [16], classrooms [17], hospitals [18], skilled nursing facilities [19], and buses [20,21]. Through the experiment of particle image velocimetry measurement, the validity of a CFD simulation study on air flow was verified [22]. The simulation method for the droplet transmission route analysis is divided into tracer gas and droplet particle tracking methods, which employ the Eulerian and Eulerian–Lagrangian methods, respectively. The tracer gas method has the advantages of lower computation time and convenient simulation accessibility. However, this method exhibited lower accuracy than the droplet particle tracking method owing to the difficult implementation of various drag forces on the droplet [23]. Therefore, this study employed the droplet particle tracking method for analyzing the droplet transmission in the train cabin [24].

Various studies to prevent droplet transmission of the COVID-19 infection were conducted in 2021. Lee et al. [25] reported that wearing a face mask reduces the transmittance distance of droplets by approximately 90–95% depending on the mask type. Heo et al. [26] investigated the effect of air purifiers on the spread of COVID-19. Ren et al. [27] reported that 70-cm-high barriers in workstations could reduce the risk of infection by 72%. Sha et al. [28] reported that the optimal control of building ventilation systems could reduce both infection risk and energy consumption. These studies suggested masks, air purifiers, barriers, and ventilation controls to prevent droplet transmission, but did not perform airflow design studies.

The aforementioned review of existing studies reveals that investigations on droplet transmission and reduction of infection risk in high-speed train cabins are limited, as shown in

Table 1. Previous studies on droplet transmission and reduction of infection risk.

Target	Analysis Method	Reduction of Infection Risk	Reference
Building
Restaurant	Simulation	-	[16]
Classroom	Simulation	-	[17]
Hospital	Simulation	-	[18]
Skilled nursing facility	Simulation	-	[19]
Room	Experiment and simulation	-	[22]
	Simulation	Mask	[26]
Test chamber	Experiment	Air purifier	[27]
Office	Simulation	Barrier	[28]
	Simulation	Ventilation control	[29]
Transportation
Cruise ship	Laboratory-confirmed case	-	[7]
Airplane	Laboratory-confirmed case	-	[8]
	Simulation	-	[14]
Bus	Laboratory-confirmed case	-	[9]
	Simulation	-	[20]
	Simulation	-	[21]
Train	Laboratory-confirmed case	-	[10]
	Experiment	-	[12]
	Simulation	-	[13]
	Simulation	Novel airflow design	Present study

. Furthermore, research on mitigating droplet transmission through the novel design of airflow systems in public transportation is insufficient. Despite the mandated wearing of masks, studies on the effectiveness of masks in cabins are also insufficient.

The objective of this study is to evaluate the mitigating effect of droplet transmission in the novel VDA system of the high-speed train, comparing it with the conventional system. Three-dimensional (3D) fluid dynamics used the droplet particle tracking method with the Eulerian–Lagrangian model to analyze the droplet transmission. Droplet transmission in the conventional and VDA systems was compared for 120 s after an infected person coughed. The airflow streamlines, droplet route, droplet number, droplet fate, and droplet travel distance were investigated. Additionally, the effect of preventing droplet transmission while wearing a mask was analyzed.

2. Methodology

2.1. Simulation Domain

Figure 1 depicts the simulation domain and schematic structure of the high-speed train cabin. A five-column section of a realistic high-speed train cabin was adopted to reduce the calculation time. The cabin length, width, and height were 4650 mm, 2862 mm, and 2196 mm, respectively. The distance between the seats was 930 mm. In a conventional system, two inlets were located adjacent to the window in each row, and one outlet was located at the bottom sidewall of each row, as shown in Figure 1a. The conventional system was designed based on the Korea train express (KTX). The proposed VDA system was designed with twenty inlets and outlets for each seat on the ceiling and floor of the cabin to create vertical downward airflow, as shown in Figure 1b. The difference in the ventilation rates between the conventional and VDA systems was 1.1%. All seats were identified (Figure 1c), and the infected person was seated at 2A. Droplet transmission was investigated in the conventional and VDA systems for 120 s after an infected person coughed once. The height of the person was 1130 mm, and the height of the mouth from the floor was 1000 mm. As depicted in Figure 1d, a person wearing a 1-mm-thick close-fitting mask with an area of 7332 mm² was designed to simulate the case of wearing a mask.

Table 2. Specifications of the high-speed train cabin.

Parameters	Values
Cabin length	4650 mm
Cabin width	2862 mm
Cabin height	2196 mm
Seat length	550 mm
Seat width	535 mm
Seat height	1100 mm
Distance between seats	930 mm
Height of the seated person	1130 mm
Height of the person’s mouth from the floor	1000 mm
Mouth area	264 mm2
Mask thickness	1 mm
Mask area	7332 mm2

and

Table 3. Geometric parameters of the conventional and vertical drop airflow (VDA) systems.

Parameters	Values
Conventional system
Inlet area	47,520 mm2
Number of inlets	4
Outlet area	130,200 mm2
Number of outlets	2
Ventilation rate	27.0 h−1
VDA system
Inlet area	9613 mm2
Number of inlets	20
Outlet area	30,800 mm2
Number of outlets	20
Ventilation rate	27.3 h−1

summarize the specifications of the high-speed train cabin and geometric parameters of the conventional and VDA systems, respectively.

2.2. Simulation Model

Figure 2 illustrates the flowchart of the simulation. The first step involved designing high-speed cabins in conventional and VDA systems using the SpaceClaim 2021 R1. The second step generated a grid using ANSYS Mesh 2021 R1. In the third step, fluid dynamics simulations with the Eulerian model were performed until the streamline of the airflow reached a steady state using the commercial CFD software, namely CFX 2021 R1. In the fourth step, fluid dynamics simulations tracked the droplet particle using the Eulerian–Lagrangian model. Finally, post-processing of the simulation results was performed using the CFD-Post 2021-R1. The simulation model was developed based on the following assumptions [21]:

• The influence of droplets on airflow is negligible;
• Droplets are assumed to be ideally shaped spheres;
• Coagulation of droplets is negligible;
• The energy equation is negligible;
• The droplet expelled from the patient’s mouth does not evaporate.

2.2.1. Eulerian Model

The governing equations of continuity and momentum in the simulation domain are presented in Equations (1) and (2), respectively [29].

$$\frac{\partial \rho V_i}{\partial x_i}=0$$

(1)

$$\frac{\partial }{\partial x_j}\left(\rho V_i{V}_j\right)=-\frac{\partial p^{\prime }}{\partial x_i}+\frac{\partial }{\partial x_j}\left[\left({\mu }_e\right)\left(\frac{\partial V_i}{\partial x_j}+\frac{\partial V_j}{\partial x_i}\right)\right]$$

(2)

where V donates the velocity and ρ indicates the density. In addition, p′ donates the modified pressure, and μ_e indicates the effective viscosity.

The shear stress transport turbulence (SST) model was applied in the fluid domain. The SST model can accurately simulate under near-wall and wall-shear boundary conditions by combining the Κ-ε and Κ-ω models [30,31,32].

2.2.2. Lagrangian Model

Considering that the droplet travels in the fluid domain, the dominant forces on the particle motion are the drag force (F_D) and buoyancy force (F_B), as expressed in Equation (3) [33].

$$m_p\frac{d{V}_P}{dt}=F_D+F_B$$

(3)

The drag force can be calculated using Equation (4).

$$F_D=\frac{1}{2}{C}_D{\rho }_F{A}_F\left|V_F-V_P\right|\left(V_F-V_P\right)$$

(4)

where A_F denotes the effective particle cross-section, and C_D indicates the drag coefficient, which is derived using the Schiller–Naumann drag model [34], as follows:

$$C_D=\{\begin{array}{l}\frac{24}{R{e}_p}\left(1+0.15R{e}_p^{0.687}\right) if R{e}_p\le {10}^3\\ 0.44 if R{e}_p>{10}^3\end{array}$$

(5)

$${\mathrm{Re}}_p=\frac{\rho \left(u-u_p\right)d_p}{\mu }$$

(6)

where Re_p denotes the particle Reynolds number, and d_p indicates the particle diameter.

The buoyancy force can be expressed as shown in Equation (7).

$$F_B=\frac{\pi }{6}{d}_p^3\left({\rho }_p-\rho \right)g$$

(7)

2.2.3. Mask Model

The mask was assumed to be a porous domain to simulate the flow through it [35]. The porous model in the mask can be expressed using Equation (8).

$$\frac{\Delta p}{d_f}=-D\mu V-0.5I\rho V^2$$

(8)

where d_f denotes the mask thickness, D indicates the Darcy coefficient, and I represent the inertial coefficient. The Darcy coefficient can be determined as [36]:

$$D=\frac{64{\xi }^{1.5}\left(1+56{\xi }^3\right)}{d_{p,c}^2}$$

(9)

where ξ denotes the packing density of the fibrous porous material, and d_p,c indicates the critical diameter. According to a previous study [37], the inertial coefficient can be calculated using Equation (10) as:

$$I=\frac{1}{{1.28}^2}\frac{1}{0.5d_f}$$

(10)

where the value of the mask porosity is 0.88, and the filtration efficiency is obtained from another study [25].

2.3. Boundary Conditions

Figure 3a depicts the cough flow rate applied by a pulse air jet with a duration of 0.5 s on the mouth surface. The experimental data [38] of the cough flow rate were applied to the boundary condition as a user-defined function. The mouth was maintained open after coughing as it is difficult to simulate the movement of the mouth. The initial conditions of the mass and number fractions of the droplets were applied based on the experimental data [39], as depicted in Figure 3b. The total number of droplets in one cough was 1850 [40]. Droplets were composed of 90% liquid water and 10% sodium chloride [21]. In the cabin, the inlet velocity was set to 1 m s⁻¹, and the outlet pressure was set to 1 atm. The wall was set under the deposition condition.

2.4. Validation

Grids were dominantly generated on the mouth, inlet, outlet, and mask. As depicted in Figure 4, grid-dependency tests were performed for the conventional system with no mask [41,42]. During the 3 s, a significant difference was observed in the droplet mass depending on the number of grids because a turbulent cloud with numerous droplets was spread out of the mouth during coughing, as shown in Figure 4a [43]. As illustrated in Figure 5, a finer grid in the cough cloud area was generated with an increase in the grid number. A grid number of 3.4 million was selected owing to its trend of the mass droplet and velocity similar to that of 5.2 million grids. Moreover, from the grid number of 3.4 million, the average element quality was higher than 0.8. The time-dependency test was conducted with five timesteps of 0.01, 0.02, 0.04, 0.08, and 0.1 s. The results indicate that the variations in droplet masses were within 1.9% for the timesteps less than 0.02 s. During the simulation time of 120 s, physical computation times of 26 and 52 h were required for timesteps of 0.01 and 0.02 s, respectively. Therefore, the timestep of 0.02 s was set to reduce computation time.

3. Results and Discussion

3.1. Comparison of Conventional and VDA Systems

Figure 6 depicts the streamlines in the high-speed train cabins of the conventional and VDA systems. In the conventional system, airflow was exhausted through the inlet near the window and moved to the top of the cabin. A large vortex was generated around two seats as the airflow reaching the ceiling of the room converged in the corridor, with the airflow discharged from the opposite side. Subsequently, the flow reached the corridor floor and formed a small vortex. A certain portion of the airflow moved toward the outlet, and the remaining portion rose to the top of the seat. The streamline in the x–y direction exhibited the shape of a rotational flow surrounding the two seats. The streamline in the y–z direction exhibited airflow diffusion to the bottom and top of the seat. In the VDA system, most airflows moved from the inlet to the outlet rapidly owing to the downward airflow design. Furthermore, a small vortex was formed in certain sections at the top and bottom of the seat. Overall, the VDA system exhibited a lower airflow circulation in the cabin than the conventional system.

Figure 7 depicts the droplet distribution of the conventional and VDA systems between 0 and 10 s after coughing; a relative droplet size was observed. In the conventional system, droplets of 100 µm or more descended directly under the seat in the initial stage after coughing because of gravity. Droplets smaller than 100 µm began to deposit as they reached the back surface of the front seat. After 5 s of coughing, most of the non-deposited droplets descended; however, droplets smaller than 10 µm began to rise upward under the influence of airflow in the cabin. After 10 s of coughing, only small droplets remained and rose to the ceiling. Therefore, large droplets exhibited a high risk of contact infection on the back surface of the front seat because of their quick deposition, whereas small droplets showed a high risk of non-contact infection as they were floating in the cabin. In the VDA system, all droplets descended regardless of the diameter at the initial stage after coughing, owing to the influence of a strong downward airflow. After 3 s of coughing, most of the droplets moved to the outlet area at the bottom of the seat, and after 10 s, only a few droplets smaller than 10 µm appeared to spread under the seat. Overall, the VDA system mitigated droplet transmission in the cabin by blocking the rise of droplets to the upper part of the cabin.

Figure 8 illustrates the variation in the droplet mass of the conventional and VDA systems according to the time after coughing. In the initial stage after coughing, both conventional and VDA systems exhibited a similar tendency of decreasing the droplet mass owing to its deposition on the back surface of the front seat. However, after 1.2 s, the droplet mass in the VDA system decreased more rapidly than that in the conventional system. This was because the droplets were exhausted through the outlet on the floor with a short traveling path. In the conventional system, the droplet mass decreased relatively slowly until 8 s, because a long travel distance is required for the droplets to be exhausted through the outlet.

Figure 9 depicts the droplet deposition distribution of the conventional and VDA systems at the back surface of the front seat and on the floor under the seat. There was no significant difference in deposition between the conventional and VDA systems. However, the remaining droplets under 10 µm spread through various routes in the cabin, as shown in Figure 8. In both conventional and VDA systems, the droplet deposition was concentrated on the back surface of the front seat at a height similar to that of the mouth. In the VDA system, a few droplets were deposited in the direction of gravity owing to the strong downward airflow. The deposition in the conventional system occurred around the front of the seated person’s left foot on the floor, close to the outlet, owing to the airflow circulation. In the VDA system, the deposition occurred around the outlet because of downward airflow. Overall, frequent cleaning of the back surface of the front seat is required to prevent virus spread through surface contact because the droplet deposition is predominantly observed on the seat in both the conventional and VDA systems [35]. Furthermore, it is necessary to change the location to prevent contact infection because the space behind the seat often has a structure that can store things.

Figure 10 depicts the droplet distributions in the conventional and VDA systems between 30 and 120 s after coughing. The droplets exhibited a uniform size. In the conventional system, after 30 s of coughing, numerous droplets were distributed predominantly around the seat of the infected person and the adjacent seat because of the airflow circulation. Furthermore, several droplets spread along the corridor to the seats at the front of the cabin. After 60 s of coughing, droplets spread to the seats in the front column of the infected person and the seats in the row adjacent to the corridor. Additionally, numerous droplets continued to float in the upper part of the cabin. Although the number of droplets was reduced by deposition or exhaust after 120 s of coughing, the remaining droplets spread throughout the cabin. During 120 s, droplets moved to all seats except seat 5D, indicating a high infection risk. In the VDA system, droplets were distributed around the infected person and near the corridor ceiling after 30 s of coughing. After 60 s, the droplets spread around the seats behind, adjacent to, and in front of the infected person. However, owing to the strong downward airflow, the number of droplets and the spread area were lower than those in the conventional system. After 120 s of coughing, the droplets spread via various routes; however, seats 3C, 3D, 4A, 4B, 4C, 4D, 5A, 5B, 5C, and 5D remained unaffected by the droplets. Therefore, the mitigating effect of droplet transmission in the VDA system was confirmed even under the no-wearing mask case.

Figure 11 illustrates the variation in the number of droplets in the conventional and VDA systems according to the time after coughing. In the conventional system, the number of droplets decreased rapidly up to 10 s as they were predominantly deposited on the back surface of the front seat. Between 10 and 25 s, the decreasing rate of the droplet number stagnated because the droplet traveled a long distance to reach the outlet following the rotational airflow in the cabin. After 25 s, the droplet number gradually decreased because the droplets reached the cabin surface and were exhausted through various routes as they spread inside the cabin. In the VDA system, the droplet number decreased faster than in the conventional system up to 10 s, owing to the decrease in the exhaust distance. After 10 s, small droplets of less than 10 µm floated in the cabin and were exhausted slowly. At 120 s, the mitigating effect of droplet transmission was confirmed as the droplet number decreased by 72.1% from 136 in the conventional system to 38 in the VDA system.

3.2. Effect of Wearing Masks

The effect of wearing close-fitting masks was analyzed and compared with that of the no-mask case. Figure 12 depicts the droplet distributions in the conventional and VDA systems with a mask between 0 and 10 s after coughing; the droplets exhibited a uniform size. In the initial stage after coughing, only a few droplets were discharged into the cabin because most of the droplets did not pass through the mask. Additionally, the droplets did not reach the front seat in comparison with the no-mask case because the porous domain of the mask reduced the cough flow rate. In the conventional system with the mask, the maximum travel distance of the droplet in the linear direction was 0.36 m after 5 s of coughing. After 10 s, most of the droplets floated to the upper part of the cabin compared to the no-mask case because the droplets with reduced momentum were affected by the airflow circulation. In the VDA system with the mask, after 0.5 s of coughing, the droplets discharged through the mask exhibited a transmission trend similar to the conventional system with the mask. After 5 s of coughing, the droplets were distributed vertically owing to the influence of downward airflow. However, the maximum travel distance in the linear direction of the droplet was similar to that of the conventional system with the mask, owing to the momentum loss of the cough airflow from the mask. After 10 s of coughing, droplets in the VDA system with the mask were distributed at the bottom of the cabin and around the outlet in comparison with the conventional system with the mask, owing to the downward airflow.

Figure 13 depicts the variations in the velocity distribution and streamline of the VDA system when coughing with and without a mask. In the no-mask case, a strong straight coughing airflow with a velocity of more than 5 m s⁻¹ was expelled from the mouth at 0.1 s. During this period, turbulence predominantly affected the initial transport pattern of droplets [44]. After 0.3 s, the cough airflow encountered the downward airflow and exhibited a descending streamline. In the VDA system with the mask, a cough airflow with a velocity of less than 2 m s⁻¹ was expelled from the mouth at 0.1 s, owing to the resistance from the mask. After 0.3 s, although the cough airflow encountered the downward airflow, it did not penetrate the center of the downward airflow because of the low flow rate. Therefore, the droplets in the with-the-mask case descended slowly compared to the droplets in the no-mask case.

Figure 14 illustrates the droplet distributions in the conventional and VDA systems with a mask between 30 and 120 s after coughing. In the conventional system with the mask, a few droplets floated near the seat adjacent to the infected person after 30 s of coughing, whereas the remaining droplets spread along the aisle to the seat in front. After 60 s, most of the droplets were deposited or exhausted, and the remaining droplets floated around the infected person. After 120 s of coughing, droplets appeared on the seats in front of the infected person. During 120 s, seats 2B, 2C, 2D, 3A, 3B, 3C, 4A, 4B, and 4C were affected by the droplets. In the VDA system with the mask, most of the droplets were floating in the area between the downward airflow and infected person after 30 s of coughing. After 60 s, several droplets spread to the front seats, adjacent seats, and aisle. After 120 s of coughing, only a few droplets floated near the top of the cabin. During 120 s, seats 1A, 1B, 2B, 2C, 2D, 3A, and 3B were affected by the droplets. Therefore, the number of seats at risk of infection in the VDA system is lower than that in the conventional system.

Figure 15 depicts the variation in the number of droplets in the conventional and VDA systems with and without the mask. In both cases, the droplet number in the VDA system was smaller than that in the conventional system. In with-the-mask cases, the number of droplets reduced significantly compared to the no-mask case owing to the filtering of droplets. Therefore, wearing a close-fitting mask effectively reduced the number of droplets in the cabin. However, the mask does not adhere to the skin completely in actual conditions during coughing; therefore, droplets leak through an open gap in the nose and cheek regions [25]. Additional research that considers airflow leakage during coughing is required.

Figure 16 illustrates the variation in the minimum and maximum droplet travel distances in the conventional and VDA systems with and without masks. In both cases, the minimum and maximum droplet travel distances increased faster in the conventional system than in the VDA system. This was because the droplets were transmitted further owing to the effect of the circulating airflow in the cabin. After 120 s of coughing, the minimum droplet travel distances of the VDA systems with the no-mask and mask cases were 84.8% and 39.1% of those of the conventional systems, respectively. Furthermore, the maximum droplet travel distances of the VDA systems with the no-mask and mask cases were 67.0% and 49.9% of those of the conventional systems, respectively. Therefore, the VDA system effectively suppressed droplet transmission by significantly reducing the droplet travel distance compared to that of the conventional system.

Figure 17 depicts the proportion of droplets in the conventional and VDA systems with and without the mask after 120 s of coughing. In with-the-mask cases, the mass fraction was calculated for the droplets that passed through the mask. In comparison with the conventional system, the VDA system with and without the mask exhibited smaller remaining droplets in the cabin. Particularly, the VDA system with the mask exhibited the highest exhaust portion of 48% owing to the short distance from the inlet to the outlet of the airflow. Therefore, the VDA system effectively exhausted droplets and reduced droplet transmission compared to the conventional system. After 120 s of coughing, the mass fraction of the remaining droplets in all systems was less than approximately 0.2%. However, as droplets smaller than 10 µm continued to float in the cabin, the number of droplets was relatively large despite the small mass fraction of droplets. Therefore, further studies are required to effectively exhaust the small droplets.

4. Conclusions

This study proposed a novel VDA system to supply a strong downstream airflow to mitigate droplet transmission in a high-speed train cabin. A five-column section of the cabin and one infected person with a close-fitting mask and without a mask were adopted in the simulation domain. The 3D fluid dynamics simulations were conducted using the droplet particle tracking method to analyze the droplet transmission for no-mask and with-the-mask cases. Compared to the conventional system, the VDA system exhibited a lower airflow circulation in the cabin than the conventional system owing to the downward airflow design. After 5 s of coughing, the back surface of the front seat exhibited a high risk of contact infection owing to droplet deposition in both conventional and VDA systems. After 120 s, the droplet number in the VDA system decreased by 72.1% than that in the conventional system. Moreover, the number of seats at risk of infection in the VDA system was lower than that in the conventional system. The results of mask effects indicated that the number of droplets in the with-the-mask case was significantly lower than that in the no-mask case owing to the effect of droplet filtering. In the cases of both with and without the mask, the proposed VDA system effectively suppressed droplet transmission as the maximum droplet travel distances of the VDA systems were 49.9% to 67.0%, respectively, of those of conventional systems. Moreover, the VDA system exhibited small remaining droplets and high exhaust through the outlet. Overall, both droplet number and droplet transmission area can be reduced in the train cabin owing to the effective airflow design of the proposed VDA system, thus preventing droplet infection. In the future, various case studies will be performed using the developing models considering the droplet evaporation and heat transfer according to the temperature and humidity. In addition, mitigating effects of the droplet transmission in the VDA system will be verified experimentally in an actual train system.