Personalized Ventilation as a Possible Strategy for Reducing Airborne Infectious Disease Transmission on Commercial Aircraft

Featured Application

Personalized ventilation systems for improving air quality around passengers in confined vehicles, such as airplanes.

Abstract

In the last decade, there has been an increase in ease and affordability of air travel in terms of mobility for people all around the world. Airplane passengers may experience different risks of contracting airborne infectious diseases onboard aircraft, such as influenza or severe acute respiratory syndrome (SARS-CoV-1 and SARS-CoV-2), due to nonuniform airflow patterns inside the airplane cabin or proximity to an infected person. In this paper, a novel approach for reducing the risk of contracting airborne infectious diseases is presented that uses a low-momentum personalized ventilation system with a protective role against airborne pathogens. Numerical simulations, supported by nonintrusive experimental measurements for validation purposes, were used to demonstrate the effectiveness of the proposed system. Simulation and experimental results of the low-momentum personalized ventilation system showed the formation of a microclimate around each passenger with cleaner and fresher air than produced by the general mixing ventilation systems.

1. Introduction

Last spring, the entire world anxiously watched as COVID-19, a new disease caused by the SARS-CoV-2 virus, was spreading from country to country by way of modern travel routes. The virus was first reported to the Chinese WHO office on 31 December 2019 [1]. It inevitably crossed borders, triggering a global pandemic [2]. Countries enacted emergency lockdown response procedures, and travel bans were put in place [3], thus cancelling almost all flights in order to control the spread of the virus. Worldwide, in 2019, airlines carried 4.5 billion passengers on scheduled services, an increase of 5% over 2018 [4,5], and prior to the COVID-19 era, that number was predicted to double until 2037 [6]. During the first week of June 2020, the number of scheduled flights worldwide decreased by 65.1% in comparison with the first week of June 2019 [5]. In June 2020, as the outbreak exceeded 7.8 million infected people around the globe, the International Air Transport Association (IATA) updated its estimated loss of annual passenger revenue to $300 billion globally for 2020, in addition to many other indirect social and economic losses [7]. In April 2021, IATA expected net airline industry losses of $47.7 billion in 2021, an improvement compared to the net industry loss of $126.4 billion in 2020 [8].

Current risk-assessment methods used by public authorities for airborne infections in airplanes are based on proximity to the index passenger, sitting within two rows of the index passenger, and a correlation with duration of exposure, depending on the virus type, typically with eight hours of exposure time [9,10].

Since SARS-CoV-2 spreads through the air [11,12], in the context of commercial aircraft, long-term passenger-protection solutions should be correlated with the conception of local ventilation devices centered on passengers. Previous severe epidemics were found to be very closely related to transmission of infectious particles between persons in indoor environments [13], which highlights the importance of indoor environmental control [14], further solidifying the necessity of correlating protection measures with aircraft ventilation systems.

Thus, current risk assessment might be fundamentally flawed because it does not take airplane ventilation into consideration, which is a key component of airborne infections [15]. In the largest in-flight SARS outbreak (Air China flight 112), a total of 22 persons seated within seven rows of the index passenger were infected during a flight of just three hours duration [9] (Figure 1a). Until this event, it was thought that there was only a significant risk of infection in flights of more than eight hours duration and in just the two adjacent seating rows [16].

In an attempt to offer to airlines new possibilities of social distancing inside the cabin, some aircraft interior design companies came immediately with temporary or permanent ideas regarding the separation of the passengers’ seats in economy cabins once widespread commercial flights resumed. A British company [18] proposed in the late spring 2020 a kit to allow airlines to offer increased protection for passengers by protective screens. Another Italian company (Aviointeriors) is envisioning the way forward by creating social several distancing-friendly options for economy class seating. Known for its standing seat product [17], Aviointeriors is proposing to separate passengers while maintaining revenue-friendly capacity levels proposing a reversed center seat together with protective screens (Figure 1b), or the use of plastic hood protections like small individual capsules around each seat. As the pandemic evolved and the society gained insight into the trans with the beginning of vaccination campaigns and the possibilities of extensive testing, the airline companies could restart their activity, with numerous airport restrictions and limited capacity. Freedom to travel is important and connecting the world by air will provide vital stability for tens of millions of people whose jobs have been lost or remain at risk from the pandemic, highlighting the need to achieve a healthier micro-environment for the passengers.

Air travel has long been correlated with the spread of viruses through infected passengers and potentially through in-flight transmission [19]. This is a problem of biology and physics, with airflow, dirty surfaces, and close contact with other travelers. Immediately after droplets are expired, the liquid content starts to evaporate, and some droplets become so small (nuclei) whose transport is influenced by air currents a more than by gravity, carrying their viral content over distances of meters or tens of meters from the point of exhalation [20]. By breathing, talking, coughing, or sneezing, everyone releases very small water droplets (aerosols, 0.3–20 μm) composed of around 97% water and 3% solutes (salts, proteins, and other substances) [20]. It appears that the relative humidity (RH) of the air plays an important role in the transmission of airborne diseases [21,22].

In low-RH environments (less than 40%) the water from the droplets evaporates faster, leading to a reduction in the mass of the droplets, which remain floating for a longer period and with increased distance travelled [23]. This causes the levels of salt in the ever-diminishing droplets to rise, altering the chemical composition of the droplets to the point in which they crystallize in a solid state. This leads, in some cases, to the creation of a floating microcapsule in which viruses and bacteria are preserved for longer and have a greater potential to infect other people. On the other hand, the cabin environment has very low RH during cruise conditions, and air distribution favors the transport and mixing of potential virus-laden microdroplets.

Most existing aircraft on the market and their ECSs (environmental control systems) were designed more than 30 years ago with a focus on space optimization and safety. The ECS maintains pressure inside the airplane cabin around 0.8 bar and a temperature of approximately 24 °C. The supplied air temperature is about 18 °C, with a relative humidity of around 10–20% or even less [24]. However, at the end of long-distance flights, humidity can be as low as 2–3% [25].

The air distribution system commonly used in aircraft cabins consists of the air supply at the top and exhaust air at the bottom of the cabin, with mixing air within the cabin [26] (Figure 2b). The mixing system forms recirculation zones (large vortex flows) in the cabin, resulting in local air stagnation. Mixing ventilation is not conducive to the exclusion of pollutants, and the dissemination of mutual interference between passengers seriously affects the cabin air quality. A literature review indicated that current air distribution in airliner cabins cannot effectively control the transport of airborne infectious disease viruses [27]. Airplanes also feature personalized ventilation systems composed of a system of gaspers located above passenger seats, which are adjustable for flow rate and direction.

Mixing between the supplied air and the air present in an enclosure is the main principle of mixing ventilation (MV) that is currently applied in aircraft, as well as in buildings. In the case of hazardous pollutants, such as virus-laden aerosols, MV is inefficient because clean air is supplied from diffusers placed far from occupants, and it has time to mix with the polluted ambient air before it is inhaled. Personalized ventilation (PV) is a new development in the field of air distribution for buildings but is widely used in the automotive field [28,29,30]. The main idea of PV is to provide clean air close to the face of each occupant and to give some control to the user to improve comfort in their microenvironment. However, PV systems might be used as protective shields against cross-infection in crowded environments, such as an aircraft.

In this case, the concept of protective personalized ventilation (PPV) is to supply clean air directly to the breathing zone of a seated person and to allow some control of the flow rate and flow direction. ASHRAE and REHVA position documents on airborne infectious diseases [31,32] recommend PV and other ventilation strategies (such as dilution ventilation, local exhaust, and source-control ventilation) for use as effective measures to control and prevent SARS-CoV-2 transmission. This proposed strategy is in keeping with evidence that PV, in conjunction MV, provides better protection against airborne infection than systems relying solely on MV.

Pantelic et al. [33] evaluated the performance of a desk-mounted PV diffuser and found that the PV could reduce both the peak aerosol concentration and the exposure time of droplets that were released during coughing. Displacement ventilation (DV) systems have also been widely used in buildings and have been shown to be more effective than MV systems in removing contaminants [30]. There have been some attempts to propose new strategies based on the displacement concept in order to reduce the transmission of airborne infectious disease viruses and/or to improve cabin air quality. Schmidt [34] and Müller [35] compared MV and DV in a section of an A320 cabin mockup. Schmidt [34] found that an MV system had higher draft risk, while a DV system could result in an unpleasant sensation of overheating. In [36], Zhang proposed an under-aisle air distribution system with the purpose of improving relative humidity from 10% to 20%. Zhang [37] also proposed a PV system with air terminals embedded in chair armrests on commercial airplanes.

Cleaner air in the breathing zone demands a certain degree of control over the flow features of the breathing zone, allowing for the interaction of PV flow with the free convection flow of the human boundary layer. Any attempt to supply a fresh flow of air to the BZ via PV needs to penetrate this boundary layer (i.e., velocities over 0.3 m/s). Previous studies showed that the PV flow strongly interacts with the convective boundary layer of the human body [18].

PV designs, such as furniture-incorporated air diffusers (in tables or chairs) [19] or air-terminal devices (ATDs) [20], are usually situated at a certain distance away from occupants’ breathing zone. Attempts have been made to dissipate the boundary layer through forced ventilation [21] or to diminish the strength of its flow by cooling the desk where the occupant is seated [22] in the hopes of improving PV efficiency. Results indicated improved levels of clean air for the occupant using reduced PV airflow (25% less), i.e., weakening of the convective boundary layer reduced required PV airflow from 8 L/s to 6 L/s.

Another method to improve PV performance is to reduce the distance between the occupant and the PV diffuser [23,24]. This can lower the PV flow rates to as little as 0.5 L/s, resulting in most of the air inhaled (>90%) being clean air. Other factors that influence this approach are the initial velocity and equivalent diameter of the diffuser, as well as its geometry [25].

Xu et al. [38] conducted a detailed experimental study to explore the flow interactions between personalized air and the thermal boundary layer of an occupant and indicated that concerns about PV jet penetration of the thermal plume are not as challenging as expected. They concluded that lower penetration velocity of clean air may be used to achieve lower energy consumption with less outdoor clean air. However, they also suggested that airflow disturbances in the breathing zone due to the thermal plume still need careful consideration, depending on the positioning and direction of the nonuniform flow from the nozzle, especially at a lower supply-air velocity.

Lobed orifices introduced in the design of mixing ventilation by a perforated panel-ceiling diffuser were found to have better initial spread due to increased induction, without reducing the jet-throw length [39,40]. The lobed edge of the nozzle generated large stream-wise structures known to be responsible for the jet induction phenomenon [41,42]. The possibility of controlling the characteristics of the inserted jet by introducing a lobed nozzle design was explored in a study by Bolashikov et al. [43]. The inhaled air quality was assessed by the personal exposure effectiveness index (PEE) introduced by Melikov et al. [44]. It has been reported that lobed-geometry nozzles have wider initial spread but increased mixing compared to free circular jets or plane jets [42,45,46,47,48,49]. For close distances between the lobed nozzles and the face corresponding to the situation considered in this article, initial side spread was considered beneficial because the inserted jet would better cover the mouth and the surroundings of the inhalation zone, resulting in higher PEE. The results, however, did not fully validate this expectation. The lowest PEE obtained was with the four-leafed lobed nozzle. The six-lobed jet performed equally well compared to the circular and elliptical nozzles for distances of 0.04 m and 0.06 m from the mouth and a velocity of 0.4 m/s.

If air is diffused in proximity of an occupant’s face, the inserted flow should not affect the body’s thermal sensation [23]. Local discomfort is likely if the initial velocity is high. Experiments evaluating the performance of a rectangular diffuser incorporated in a headset [32] with an initial velocity of around 1.7 m/s were met with reports of unpleasant sensations from the human test subjects. Velocities had to be diminishes by around three times (~0.6 m/s) to reduce the unpleasant sensations of draft [33].

PV airflow on the face can substantially reduce exposure, regardless of the location of the pollution source. Local exhaust ventilation integrated into seats may reduce exposure. Yang et al. [50] investigated how the performance of a desk-mounted PV system may be affected by the presence of two personalized exhaust (PE) systems above shoulder level. Their study found that such an integration introduced more PV air into the breathing zone and exhausted part of the exhaled air with the free convective flow.

PV diffusers in proximity of the human face also reduce exposure of the eyes due to the protective effect of the flow of fresh air if the PV can target the occupant’s breathing zone, which is rarely the case when the distance between the head and the PV diffuser is significant, in which case an increased frequency of eye blinking has been reported [34].

Considering the previous results, we propose to install a six-lobed PV diffuser situated in proximity to the occupant at a distance of 0.06 m (equivalent to 2De, based on previous studies [31]). This PV diffuser geometry and distance enable the PV jet to supply fresh air to a wider region around the face, compensating for small changes in head position.

The aim of the present paper is to study the possibility of using a low-momentum PV system that can penetrate a passenger’s convective boundary layer, improving air quality in their breathing zone. A successful implementation of such a PV solution is the first step towards increasing passenger protection against airborne pathogens.

The evaluation of the proposed PV solution is analyzed using numerical simulations, supported by experimental measurements for validation purposes, allowing us to make preliminary conclusions about the efficiency of such systems.

The numerical approach proposes comparison of two models reproducing a small region inside an aircraft cabin containing four seats and virtual passengers. The experimental validation was decoupled, given the complexity of the dynamics involved in velocity and temperature distributions. Therefore, we chose to compare velocity and temperature distributions inside the thermal plumes of passengers, both in coronal and sagittal planes. Another step of validation was based on the velocity distributions generated by the PV diffuser.

Reliable numerical models are the first step towards a more complete evaluation of a passenger’s microenvironment aboard aircraft, since a multitude of variables need to be accounted for (exhaled pollutants, pathogen transport, human impact on relative humidity, etc.), which are very difficult to accurately measure with experimental data. The authors consider the numerical models developed in the present study an essential steppingstone towards the future development of protective personalized ventilation solutions for commercial aircraft.

2. Materials and Methods

2.1. Experimental Setup

The experimental campaigns were carried out in an experimental chamber with dimensions of 3.8 m × 2.6 m × 2.5 m (L × W × H), where we reproduced a configuration of a small part of an aircraft cabin, including 2 rows, each with 2 passenger seats and a general ventilation configuration similar to that in Figure 2. One of our advanced prototype thermal manikins [51] was placed on one seat. Three thermal dummies with advanced control system were placed on the three other seats, simulating a total of 4 passengers (Figure 3). The purpose of this experimental setup was to perform rigorous measurements of the interactions between the thermal plumes of the four passengers, as well as the general ventilation system.

Thermal plumes were characterized with and without the functioning of the general ventilation system using infrared (IR) thermography, arrays of thermocouples (Figure 4), and particle image velocimetry (PIV) measurements. The thermocouples were calibrated prior to the experimental campaign using a thermostatic bath (Lauda Eco Silver). For data logging from the thermocouple array, we used an ALMEMO 710 acquisition system.

During thermal plume measurement, the air temperature at several points of the test chamber, as well as the wall temperature, was recorded with PT100 temperature sensors (see Figure 5) connected to a data-acquisition device (ALMEMO 710). Some data-acquisition devices were placed in the center of each wall (t₁–t₄), at a height of approximately 1.5 m, including floor (t₆) and ceiling (t₅), while others were used to measure the ambient air temperature and the air in the diffuser (t_a1 at the bottom of the room, 1 m from the floor, and t_a2 at 2.1 m).

A Flir E6 IR camera was used to periodically check the wall temperature of the experimental chamber and the temperature of the segments of the manikin. The Flir E6 IR camera has an accuracy of ±2% and a thermal sensitivity of <0.06 °C. Infrared thermography (IR) measurements were also taken with the Flir E6 IR camera to determine the gradient temperature of the thermal plume. A thin, black cardboard sheet was fitted on top of the manikin’s heads (Figure 4). A grid of measurement points was placed above the experimental manikin’s heads, with 50 mm distance between each of the grid’s measurement points (marked using black metallic paint as seen in Figure 4). The emissivity (ε) for the IR camera, while reading on the black duct tape, was set to 0.98, and 0.75 while reading on the aluminum foil covering the whole manikin. The black cardboard with the grid provided a scale for measuring distances.

Particle image velocimetry (PIV) measurements were taken to determine the velocity profiles of the thermal plumes. The PIV system is composed of one high-sensitivity Flow Sense EO camera with 4 × 10⁶ pixel resolution and a Dual Power 200 mJ laser with a wavelength of 532 nm. The acquisition frequency of the PIV system, between two pairs of images, was 7.5 Hz. The image calibration gave a spatial resolution of 182 µm per pixel, which corresponds to a 365 × 365 mm² field of view.

In order to see the airflow, the entire room was seeded with a fog generator that used olive oil. An air compressor was used to supply the fog generator with compressed air. During PIV measurement, the fog generator and the ventilation system were turned off, and the exhaust grilles were sealed. The laser was placed on a rack system, perpendicular to each manikin’s head, while the camera was placed on the left side of the laser, rotated at an angle of 90°, to observe the laser plane. The manikins were placed approximately 1 m from the laser. The smoke generator and compressor were placed on the floor of the test chamber (Figure 3).

To avoid the presence of reflections when measuring with the PIV system, black duct tape was placed on the surfaces of the manikins that intersected laser plane.

The images were processed through an adaptive multigrid correlation algorithm handling the window distortion and the subpixel window displacement. The time interval between two laser beams was between 1000 ÷ 2000 µs, capturing 500 images for each measurement. The relative error did not exceed 3% for the V component. For the U component, given its low value in the considered planes, outside the flow, it might attain up to 25%. Smoke generation and the time interval between the two lasers were adjusted according to the flow.

In order to validate the flow issued from the PV diffuser, we used data from a previous study [52]. The experimental approach used in order to obtain the distributions of the velocity magnitude fields is detailed in [52].

2.2. Numerical Simulation

The present study makes extensive use of the CFD method in the study of the flows generated by the general ventilation system, the PV system, and the thermal plumes of the heated manikins. Numerical study was carried out in Ansys Fluent. Calculations were performed using a double-precision pressure-based coupled solver. To solve the turbulent flow, a RANS approach was chosen. To simulate natural convection, a relatively large density of cells for the fluid domain mesh had to be used near solid surfaces and in the field. Additionally, the low velocities in the climatic chamber implied low values for y+ when using a robust mesh in term of spent computational resources.

These considerations imposed a low-Re turbulence model. Thus, the k-ω SST was adopted for all performed numerical simulations [53,54,55,56]. The gradients were computed using a least-square cell-based method. Second-order discretization schemes were used for pressure, momentum, and energy equations, while for turbulence quantities, first-order upwind schemes were chosen.

To initialize the flow field, a standard method was set. To ensure better convergence in the first steps, the whole fluid domain was patched using a low-velocity value along the Y axis (in the opposite direction of the gravity force) and a temperature value close to the equilibrium value measured during the experimental tests.

The mesh consists of 8 million mixed 3D poly-hexcore cells with polygonal 2D cells on the solid surfaces. Near the solid surfaces, to correctly capture the velocity and temperature variation in the boundary layer, 10 layers of prismatic cells were grown using a constant aspect ratio equal to 10 for the first cells near the walls. To ensure an optimal distribution of cells in the fluid domain, different characteristic lengths were used for different zones. The characteristic dimensions of the cells are detailed in

Table 1. Characteristic dimensions of the cells in the computational domain.

Climatic Chamber Walls [cm]		Manikin and Chair Surfaces [cm]		Climatic Chamber Velocity Inlet, Pressure Outlet [cm]	PV System Velocity Inlet [cm]	System Walls [cm]		PV Jet Zone [cm]		Field [cm]
Min	Max	Min	Max			Min	Max	Min	Max	Min	Max
1	2	0.1	0.5	1	0.1	0.1	0.5	0.1	0.5	0.1	5

. Values were chosen considering a previous study [57] so that the grid would correctly capture the experimentally determined temperature and velocity fields and satisfy the requirements imposed by the numerical model with respect to y+ values near the solid surfaces.

To avoid large leaps between cells, a growth rate of 1.1 was used, except in the boundary layer zone, where a growth rate of 1.2 was applied. Additionally, in the hexcore zone, to avoid 1:8 cell jumps, the transition regions from hex cells with different characteristic lengths were filled with polyhedral cells. For simulations in which a PV system was considered, a rectangular zone with dimensions of 5D_e × 2D_e × 2D_e was refined so that the influence of the PV jet could be better captured. The maximum characteristic length in this region was equal to 0.5 cm. Within this zone, another rectangular region was defined, wherein the mesh was refined so that the cell length was equal to 0.25 cm.

Figure 6 presents details of the mesh, and Figure 7 illustrates the distributions of y+ values on the manikins and seats surfaces. The maximum wall y+ value was 2, ranging from 0.29 to 2 on the surface of the manikins and seats [58].

Temperature-boundary conditions on solid surfaces and inlets are detailed in

Table 2. Boundary conditions and temperature conditions of the walls and the human body.

Parameters	Boundaries
Surface Temperature [°C]	ta1-Air	ta2-Air Diffuser	t1-Wall 1	t2-Wall 2	t3-Wall 3	t4-Wall 4	t5-Ceiling	t6-Floor
	17.3	22.2	22.7	22.7	25.0	23.5	22.7	21.1
Manikin temperature [°C]	torso		head and neck		hands		legs
	32		34		30		27
Air-diffuser flow rate [m3/h]	286

. A schematic related to the surface-temperature values of the virtual manikins is presented in Figure 8.

3. Results and Discussion

Numerical validation based on the comparison of experimental measurements and numerical results for the velocity and temperature fields were first carried out in the central, sagittal planes of the thermal plumes. As previously mentioned, the thermal plumes were characterized with and without the functioning of the general ventilation system using infrared (IR) thermography, arrays of thermocouples, and particle image velocimetry (PIV) measurements. Some studies have shown that the higher the temperature differences between the surface of the manikin and the air in the environment, the higher the velocity of the thermal plume [59]. For an increase in ambient air temperature of 6 °C (20 to 26 °C), a decrease in peak velocity of the thermal plume of a standing subject of 30% was observed, with no influence on the shape of the convective boundary layer [60]. Based on PIV measurements, Yuki et. al. [61] studied the thermal plume generated by a seated manikin in a room with variable air temperature. The authors varied the velocities in the enclosure, obtaining airflows transversal to the thermal plume characterized by velocity magnitude values of up to 1 m/s. In this case, the airflow generated by the thermal plume was no longer observable. Consequently, in the present study, the validation of the numerical thermal plume results by their experimental counterparts will be done using a test case the general ventilation system was turned off for both the experimental and numerical of approaches.

In a second step, we compared velocity distributions obtained at different locations of the transverse planes of the flow issued from the PV nozzle with the experimental results from [52].

3.1. Validation of Thermal Plumes

3.1.1. Particle Image Velocimetry

In this paragraph, we propose a comparison between the experimental velocity magnitude distributions obtained by PIV in the sagittal and coronal planes of the thermal manikin and the same distributions obtained from the CFD model. Although we performed the same type of comparison for each thermal dummy, all data are not presented for the sake of brevity. The comparison between the experimental and numerical results represents a situation where the thermal plume develops naturally, without interaction with the ventilation system.

Figure 9 presents the velocity magnitude distributions from PIV and CFD in the coronal plane, above the head of the thermal manikin. Figure 10 presents the velocity distributions in the sagittal plane of the same manikin. Numerical results and experimental measurements for the velocity field in the sagittal plane have the same ranges at the same height. A higher value of the velocity magnitude in the back part of the manikin can be seen in the numerical model, and this is due to a better capture of the natural convective boundary layer that forms from the lower part of the manikin, which tends to grow in thickness and velocity.

3.1.2. Thermal Measurements

In a first step, using a Flir E50 Infrared camera and a black cardboard sheet, we obtained a qualitative distribution of the temperature measurements in the thermal plume. Figure 11a and Figure 12a are present thermographic images captured in this way, allowing for qualitative inspection of the air-temperature distributions within the thermal plume. Figure 11b and Figure 12b present their numerical counterparts.

Their comparison results in similar distributions, showing that the temperature gradient and values have almost the same allure and the same height for both sets of results. The comparison between the experimental and numerical results represents, as previously, a situation with a natural convective thermal plume.

As we wanted to provide a quantitative comparison as well, we used a network of 24 thermocouples installed in sagittal and coronal plane above the manikin head, as explained previously, within a grid with six rows and four columns, keeping a distance of 40 mm between them (see Figure 13). Each thermocouple had an ID composed of its row number and column number, starting from the bottom-left thermocouple, as displayed in Figure 13.

Figure 14a and Figure 15a present a comparison of the temperature values obtained from the thermocouples and those obtained from CFD modeling at the same locations above the heads of the manikins. Blue bars represent experimental values, and orange bars represent numerical values. For the most part, the experimental results show slightly higher values than the numerical results (Figure 14a and Figure 15a). Figure 14b and Figure 15b show relative error distribution between numerical and experimental results. As these differences stay within the accuracy range of the measuring technique, we therefore we find that results are satisfactory.

3.2. Validation of PV Fields

We used for the validation experimental results from one of our previous studies [52]. During these experimental measurements, in isothermal conditions, velocity magnitude fields of the PV jet were measured with a hot-sphere anemometer at three distances in front of the PV diffuser: 0.5D_e, 1D_e, and 1.5D_e (Figure 16(a1–c1), respectively). In this figure, the velocity magnitude distributions are centered on the PV jet and were measured in a 6 × 6 cm grid with a measurement resolution of 2 mm (31 × 31 measurement points). Figure 16 presents a comparison from the present numerical model, in isothermal conditions, at the same distances from the PV nozzle (Figure 16(a2–c2), respectively) and from another numerical model of the same diffuser (Figure 16(a3–c3), respectively), also from the same previous study [52]. The main difference between the two numerical models is the density of the mesh dedicated to the PV nozzle itself and to the corresponding flow-domain region. Additionally, in the previous study, the flow was entirely resolved in the upstream ducting of the air distribution system. Another difference is the position of the virtual human body used in [52]. We placed the PV nozzle at the same position in relation to the face in order to cover the breathing zone, but the manikin was not seated like in the current study.

In all considered cases, results show that the lobed form of the PV jet is clearly seen at 0.5D_e. At 1D_e, relics of the lobed jet are still present, while at 1.5D_e, those same remnants have almost disappeared as the lobed jet transitions into a round jet. Peak velocities at all three distances (Figure 16a–c) are around 0.5 m/s. The jet expands from its width and height (W × H) of approximately 4 × 4 cm at 0.5D_e to a W × H of 5 × 5 cm at 1.5D_e (an increase of around 20%).

The numerical model from the current study presents a less symmetrical flow distribution, which is explained by the proximity of the PV jet to various surfaces of the virtual human body in the seated position, such as the left arm of the manikin and other surfaces, like the right upper cushion of the seat (see Figure 6). Figure 16d, compares the velocity magnitude values along the axis of the three PV jets (numerical and experimental from study [52] and the numerical PV jet from the present study) between the PV nozzle and the occupant’s face.

3.3. Numerical Results

The numerical model reproduced the entire experimental chamber with the general ventilation system. Figure 17 presents the distributions of velocity and temperature inside the modeled domain in the case without a PV diffuser in different vertical planes passing through the passengers in both longitudinal (A, B, C) and transverse direction (1, 2) and one horizontal plane passing at the level of the breathing zone (D).

Figure 17 displays the flow fields in the cross-section across the two rows and in the longitudinal section along the seats. A large circulation patten is formed on the sides of the cabin, with the vortex center above the heads of the passengers. The velocity in the jet regions is higher, while the velocity in the center of the cabin is lower. In the region of the passengers, the maximum velocity magnitude is 0.2 m/s, which is consistent with the requirements for an aircraft cabin. The passengers on the right side are found in a downward flow of the circulation pattern, while the passengers on the left side are found to be in the upward flow. With the downward air supply from the ceiling above the seats, the airflow was not vertical but shifted towards the front seats. In the longitudinal section, most of the airflow is downward because of the circulation pattern. The velocity magnitude near the passengers’ heads is low because the upward thermal plume counteracts the downward general flow. The velocity distribution is not exactly the same in front of all the virtual passengers.

Figure 17 also displays the temperature distributions in the same sections as were presented for velocity. The temperature distribution is relatively uniform, except in the ventilation jet and above the heads of the manikins. The air temperature is slightly higher above the manikins because of the thermal plume created by the released heat.

The thermal plumes above the virtual passengers in different seats varies slightly in the longitudinal section, as shown in Figure 17, which indicates the unstable nature of the flows in these regions.

Figure 18 presents the velocity and temperature distributions for the numerical cases with the PV diffuser installed near the breathing zone of one of the passengers. Contours are presented in vertical planes passing through the passengers in transverse direction (1, 2) and longitudinal direction on the second seats row (A). Another vertical plane is slicing through the PV diffuser (B).The general pattern of the flow within a recirculation region is maintained in this case, and the temperature and velocity distributions are similar to those of the previous case, except for the region around the virtual passenger with a personalized ventilation system. In this case, the influence of the PV flow in front of the face of the targeted occupant can be clearly distinguished.

The jet flow delivered from the PV diffuser has low momentum and is entrained by the convective layer around the body of the virtual manikin in such way that it delivers air to the face of the virtual passenger. Given the fact that the purpose of the PV diffuser is to supply fresh air to the occupant, it is of interest to place the flow in the context of the definition of the physical boundaries of the breathing zone.

Previous research of our team concerning CO₂ accumulation in the crew quarters of the International Space Station [62] allowed us to define a breathing zone (BZ) for the occupant by applying Fast Fourier Transforms to the simulated periodic human breath flow. Later, in [52], using the same simulated breathing function, a comparison was proposed of breathing zone CO₂ concentration levels with and without a PV system to determine whether the impact of a lateral PV nozzle is noticeable in the breathing zone during a full breathing cycle. As mentioned previously, the positioning of the PV nozzle from the head/face of the virtual manikin in [52] is similar to the one in this study, however the body position is slightly different in this previous work (i.e., zero g neutral position). In this case the proposed definition of the breathing zone concerns zero gravity conditions–no buoyancy, no gravity influence on the CO₂ presence. In these particular conditions, the PV solution proved to be efficient, showing reduced CO₂ content in the breathing zone of the occupant across the whole breathing cycle. Although both the lateral PV solution and the general ventilation system alone were capable of removing the exhaled CO₂ from the BZ across a breathing cycle, closer inspection revealed a supplementary 8% reduction in inhaled CO₂ over each breath in the case of the PV system.

In the current study, we implemented the same type of PV strategy (i.e., a lobed nozzle on a lateral diffuser). However, the present analysis represents the preliminary step within a more complex analysis. We did not take into consideration CO₂ emissions from breathing, for instance, nor did we simulate aerosols.

Figure 19 shows close-up images of the distributions of velocity and temperature around the passengers with and without PV. The presence of the PV diffuser itself modifies the distribution of the thermal boundary layer of the passenger (Figure 18) at the head and torso levels. Because the thermal boundary layer in the frontal part of the torso and of the head partially intersect with the breathing zone, if the PV jet modifies the boundary layer (as seen in Figure 19), it follows that it will influence air distribution in the breathing zone as well. The degree to which this influence is manifest depends on the PV jet’s initial velocity, as well as its orientation. This is illustrated in Figure 20, where we tried to put in evidence the interaction between the convective boundary layer of the human body and the PV flow with and without a PV diffuser. We superposed the frontiers of the breathing zone in front of the passenger (as defined in [62]) on the velocity and temperature fields.

As visible in Figure 17, Figure 18, Figure 19 and Figure 20, the boundary layer of each passenger develops around their bodies and is visible at the upper part of the shoulders and heads. In the transverse planes, the convective flows are also attached to the vertical surfaces of the seats extending towards the neighboring seat.

In Figure 20b, in the region of the PV diffuser, higher values of velocity magnitude are observable, starting with a small region corresponding to the initial region of the low-momentum PV jet and developing further in what could be the interaction between the convective boundary layer and the farther region of the PV flow. In Figure 20c, the air-temperature distributions show that the small region corresponding to the potential core located in front of the face of the passenger is colder than its surroundings. This could be explained by the fact that the clean air issuing from the PV nozzle has not yet mixed with the ambient air, thus fulfilling its objective. Since the PV jet brings fresh air to the passenger, the PV solution has the potential to improve the air quality in the passenger’s breathing zone.

4. Conclusions

In the last decade, flow and thermal analyses of airplane cabins have been performed with experimental and numerical data. Three methods are available for aircraft ventilation: mixing, displacement, and personalized ventilation. Mixing ventilation achieves good air-flow distribution but does little to combat airborne disease transport. In the present study, the authors numerically investigated the possibility of introducing a personalized ventilation diffuser, combined with mixing ventilation, to improve air quality and thermal comfort and reduce the risk of airborne diseases.

Humidity is an important part of airborne pathogen transmission. Dry air allows an intensive dehydration of mucosalivary droplets and their transformation into droplet nuclei that can float for a very long time, possibly entering the breathing zone of other passengers. That is why we believe that personalized-ventilation-based solutions that provide fresh and clean air to this breathing zone represent a strategy that is worthy of consideration. A personalized ventilation circuit could also be equipped with a local humidification solution adapted to the needs of each occupant of every seat in the cabin.

The aim of the present paper is to study the possibility of using a low-momentum PV system that can penetrate a passenger’s convective boundary layer, improving air quality in their breathing zone. A successful implementation of such a PV solution is the first step towards increasing passenger protection against airborne pathogens. This study is part of a larger project including experimental and numerical investigations.

Evaluation of the proposed PV solution was analyzed using numerical simulations supported by experimental measurements for validation purposes, allowing us to make preliminary conclusions about the efficiency of such systems. The numerical approach proposes the comparison of two models reproducing a small region inside an aircraft cabin containing four seats with virtual passengers.

The jet flow delivered from the PV diffuser has low momentum and is entrained by the convective layer around the body of the virtual manikin in such way that it delivers air to the face of the virtual passenger. The presence of the PV diffuser modifies the distribution of the thermal boundary layer of the passenger at the level of the torso and head. The presence of the convective layer formed around the human body, in combination with the effect of the interaction between the PV jet flow, modifies the air flow patterns in the breathing zone in terms of velocity and temperature. The air-temperature distributions show that the small region corresponding to the potential core, located in front of the face of the passenger, is colder than its surroundings. This could be explained by the fact that the clean air issuing from the PV nozzle has not yet mixed with the ambient air, thus fulfilling its objective. This allows us to make a hypothesis related to the distribution of clean-air concentration.

This study represents the preliminary step within a more complex analysis. In our study, we did not take into consideration CO₂ emissions from breathing, for instance, nor did we simulate aerosols. These two factors are to be considered in the next phase of our study.