Numerical Analysis of the Impact of the Use of Personal Protective Equipment on the Face in the Process of Pollutants Spreading Emitted During Breathing

Personal protective equipment (PPE) plays an important role in reducing the airborne transmission risk of contaminants exhaled by infected people, but also protects occupants from the risk of infection. During the pandemic, people started wearing face shields of different shapes and construction and face masks of different protection quality, from high protection given by FFP2 type to simple homemade material masks. A wide variety of parameters like material type, shape, fit to the user’s face and user habits play an important role of PPE efficiency in the reduction of virus transmission in the air [1,2]. A deeper understanding of the efficiency provided by PPE is still needed.

Airborne transmission occurs mainly indoors due to the spreading of exhaled jet that contains infectious droplet nuclei. Exhalation can occur as breathing, sneezing, and coughing but also when talking. Each of that expiration process characterizes different velocities from 20 m/s for sneezing and from 1 to 5 m/s for breathing and talking. Droplets’ sizes range from 0.1 to 1000 µm [3, 4]. Droplets exhaled by people evaporate quickly to form droplet nuclei. Droplets nuclei of size 10 µm and less are airborne and may remain suspended in the air for hours. They can contain viruses that can be transported for a very long distance [5].

Airborne transmission is investigated using experimental and CFD studies [5]. Both experimental and CFD studies use thermal breathing manikins [6–11] or Computer Simulated Person (CSP) that rely on the human breathing process [9, 12–16]. Basic knowledge of the human breathing process, physical parameters and developed mathematical formulas are based on experimental measurements with people [17–19].

A cross-infection risk is simulated in the presence of a Computer Simulated Person (CSP) with different breathing mode [12, 14–16]. CFD simulations require a well-developed numerical model with properly designed boundary conditions. An extensive review of the BC for CFD models including characteristics of the thermal body, the geometry of mouth, nose, characteristics of breathing mode and breathing flow can be found in [5]. A number of CFD study focuses on the importance of ventilation strategy like: Mixing Ventilation, Displacement Ventilation, Personal Ventilation on contaminated flow distribution and cross-infection risk [9, 15, 20]. Villafruela et al. [16] show that exhaling through the nose or mouth disperses exhaled contaminants in a completely different way. The influence of the thermal plume on exhaled by people aerosols was analysed by Feng et al. [14]. In a number of numerical simulations, a cross-infection risk analysis is based on the pathways of airborne droplets nuclei. A simplified model includes only a carrier phase where exhaled contaminants are simulated using tracer gas: CO₂, N₂O or SF₆. A spatial distribution of ejected tracer gas concentration indoors was studied by [9, 15, 16, 21]. A discrete phase is simulated by particle generation in the exhaled air using DPM (Discrete Phase Model). These models are much more realistic, however, they generate problems with a proper prediction of particle size distribution and its concentration in exhaled air. The transport of particles was included in the study of [12, 14, 22]. Simplified numerical models usually calculate steady-state conditions with constant exhalation [23]. In transient calculation, a full breathing model with inhalation and exhalation must be assumed [24].

A current topic of scientific research in the field of ventilation, medicine and occupational health is the efficiency of PPE in reducing cross-infection risk. Wearing personal protective equipment has become our daily routine during the pandemic. There is available a wide variety of face masks and shields, characterized by different effectiveness in removing viruses and contaminants. Manufacturers declare only the effectiveness of removing impurities for specialized masks and face shields that fulfil the requirements of the dedicated standards. Sai Saran et. al. [25] revised existing standards and recommendations on the technical aspects of PPE. Many researchers focused their numerical studies on the issue of PPE application [2, 26–28]. Dbouk and Dirkakis [29,30] showed in their study the transition of respiratory droplets through a face mask. They proved that using the face mask will reduce airborne transmission, however, the mask does not provide complete protection due to leaks that appear during coughing. Lia et al. [28] calculated that most leaks appear in the regions of the nose (40%). Deng and Chen [2] focused their work on the analysis of suitable social distancing for people wearing face masks during the pandemic. They have found that social distancing could be reduced to 0.5 m when people wear the masks. Numerical modelling of face protection against a sneeze was analysed by Ugarte-Anero et al. [27]. They proved that face protection does not provide accurate protection. The only protected person is that one is sneezing. Tretiakow et al. [26] analysed the time of contact with contaminated air. They have found that when the air starts to move the time of contact with contaminated air significantly decreases compared to still air. In case of wearing PPE also a risk of re-inhalation of contaminated air is important. Jian and Ai [28] found that oxygen fraction (O₂) in the inhaled air once wearing a mask is 2% less than without wearing a face mask. This discovery shows how important is an accurate design of PPE in order to protect the user’s wellbeing when wearing masks or shields for many hours.

Our study presents a comparison of a few selected face protective equipment in terms of the spatial distribution of contaminants in the close vicinity of the face under the different phases of the breathing cycle. A few face shields of a different shapes and distances from the face as well as a simple face mask were analyzed. We have chosen CO₂ as a tracer of contaminants distribution. Application of the CO₂ as an indicator of airborne cross-infection risk was an issue of a few publications [31–33]. They have found that CO₂ can be successfully used as an indicator of airborne cross-infection risk. A risk of re-inhalation of exhaled contaminated air was also analyzed in our study. Many people after wearing face protection for many hours, complain that they fill a deterioration in their well-being. It is suspected that this is a result of inhalation of increased CO₂ concentration and acidification of the body which causes symptoms of headache, irritability and general malaise.

COMPUTATIONAL MODEL

2.1.

Geometric model

A three-dimensional geometry of the human head with torso was prepared. Human stature was located in the environment simulated in the shape of a cube with a volume equal to 1 m³ (edge length is 1 m), (Fig. 1). Human is breathing through the nose. The surface of one nasal opening was 0.65 cm² [33]. Five types of geometrical models were analysed. In each model human head was equipped with different personal protective equipment (PPE). Types of PPE considered in the study were: two types of face shields worn at a different distance from the face and fabric face masks with shape characteristics for chosen FFP masks. The fifth reference geometry did not include any personal protective equipment (Fig. 2).

Example of geometry showing the torso and head of a human with the fabric mask

Examples of Personal Protective Equipment used in research, A – full face shield, B – small face shield, C – material mask, D- no PPE

Table 1.

List of parameters describing the quality of meshes

Model	Number of elements	Average skewness	Average orthogonal quality
Reference model without PPE	2 485 096	0.181	0.817
Model with shield covering nose and mouth	2 637 907	0.183	0.814
Model with face shield, 15 mm far from face	2 779 952	0.185	0.812
Model with face shield, 40 mm far from face	2 811 042	0.186	0.811
Model with a fabric mask	907 944	0 (skewness parameter is not applicable for polyhedral mashes)	0.96526

2.2.1.

Numerical discretization

The computational model was discretized with Ansys Meshing software. In the central part of the domain, near head, neck and torso model was discretised with tetrahedral elements. The numerical grid in the rest of the domain was discretised with hexahedral elements. Between tetrahedral and hexahedral elements, a layer of pyramidal elements was created (Fig. 3).

Refined computational mesh on the cross-section of the numerical model. Geometry with a small face shield

The size of particular elements of the grid was globally set as 11.5 mm. The Sizing functions were used on the surface of the head, neck and torso, to increase the number of the grid elements in the vicinity of the human body. The elements size was set locally as 2.95 mm. The growth rate in the whole model was defined as 1.2. Particular parameters describing individual meshes are shown in Tab. 1. The values of average skewness and orthogonal quality of mesh elements in each model indicate the good quality of numerical meshes used in this study [34].

The computational cubic domain for the computational case with the fabric mask was discretized with the Fluent Meshing software due to the higher complexity of geometry. This numerical mesh contains only polyhedral elements in its structure (Fig. 4). The minimal size of surface mesh elements was set as 0.05 mm where the maximum element size was equal to 10 mm. The growth rate in the whole model was defined equal to 1.2. Boundary layers were added on the surfaces of the body and mask. Their transition ratio was set as 0.272. The maximum size of volume mesh elements was set to 13 mm. The number of elements and their average orthogonal quality are summarized in Tab. 1.

Refined computational mesh on the cross-section of the numerical model. Geometry with a fabric mask

2.2.2.

Grid sensitivity study

The grid independence study was carried out on the model with a face shield covering the entire face. For this purpose, 4 numerical meshes were created using Ansys Meshing software. The number of elements of each of the four meshes is summarized in Tab. 2.

Fig. 6 shows the course of the mean values of carbon dioxide concentration as a function of time on the upper surface of the computational domain (Fig. 5), obtained for two consecutive breathing cycles for which a pseudo-steady state was reached.

Computing domain with marked location of the top surface

Graph of CO₂ concentration changes in time on the upper surface of the computational domain for 4 different numerical mesh for two breathing periods at pseudo-steady state

The lowest concentration value was obtained for the first mesh with the smallest number of elements: mesh (1.1) in Tab. 2. The results obtained for the successive mesh show a similar course of the CO₂ concentration function. The fundamental method of verifying simulation results is by showing asymptotic convergence of the numerical solution. For this purpose, the average CO₂ concentration at the upper domain boundary for a pseudo-steady state was computed as follows (1):

Table 2.

Grid parameters used for grid independence study

Mesh:	1.1	1.2	1.3	1.4
Element Size [mm]:	20	15.4	11.5	8.8
Sizing [mm]:	5	3.8	2.9	2.2
Number of elements:	887 932	1 643 675	2 886 498	5 251 038

{\bar{Y}}_{{CO}_{2}} = \frac{1}{T A_{u p p e r}} \int_{τ}^{τ + T} \iint_{A_{u p p e r}} Y_{{CO}_{2}} d S

${\bar{Y}}_{{CO}_{2}}$ – averaged mass concentration of carbon dioxide,

${\bar{Y}}_{{CO}_{2}}$ – time and position-dependent mass concentration of carbon dioxide, -

A_upper – an area of upper domain surface, m²

T – breathing period duration, s

Fig. 7 presents the value of the average concentration of carbon dioxide at the upper domain surface as a function of a number of mesh cells, it can be clearly noticed that the solution is within the asymptotic region of the mesh size convergence.

Graph of average CO₂ concentration at the top surface as a function of a number of mesh cells

Results presented in Figure 7 were used to carry out verification of the numerical solution. First of all for three consecutively refined meshes the observed discretisation order was computed (2) [35, 36]:

p = \frac{\ln (\frac{f Δ x, 3 - f Δ x, 1}{f Δ x, 2 - f Δ x, 1} - 1)}{\ln (r)}

f_Δx,i – numerical solution at i-th refinement level,

$r = \frac{Δ x_{2}}{Δ x_{1}} = \frac{Δ x_{3}}{Δ x_{2}} = \frac{Δ x_{4}}{Δ x_{3}} ≅ 0.82$ – average mesh refinement ratio,

Having a value of the observed discretisation order, the Grid Convergence Index (GCI) can be computed, which is interpreted as an estimation of the relative discretisation error (3):

E = | \frac{f_{e} - f_{Δ x, i}}{f_{Δ x, i}} | ≅ GCI= | \frac{f_{Δ x, i + 1} - f_{Δ x, i}}{f_{Δ x, i} (r^{p} - 1)} |

f_e – exact solution to the problem,

E – relative discretisation error,

The results of the solution verification procedure are presented in Tab. 3. It can be noticed that the observed discretisation order is equal to around 1.6 which is slightly lower than the theoretical discretisation order of the applied discretisation scheme. Moreover, it can be seen that the GCI decreases for subsequently refined meshes which proves that the solution is within the asymptotic region of mesh convergence.

Table 3.

Results of the mesh independence study

Mesh	Number of cells, Millions	Mesh refinement ratio	Solution, ppm	Observed discretisation order	GCI
1.1	0.888	-	725.5	-	-
1.2	1.644	0.81	750.0	-	0.123
1.3	2.886	0.83	767.7	1.63	0.086
1.4	5.251	0.82	784.0	-	0.077

Based on the carried out mesh dependence study, numerical mesh number 1.3 was chosen for further simulations, because it offers good accuracy in reasonable computing time.

2.3.

Equations and simulation parameters

The mathematical model consists of the following governing equations: continuity (4), momentum (5) and energy conservations (8), turbulence model equations (6, 7) and species of transport (9). Four species transport was included in the model: O₂,CO₂, H₂O and N₂. The time-dependent calculations were solved due to a full breathing model including breath in, breath out and a pause between breaths. The airflow turbulence was simulated using SST k-ω model. The enhanced wall treatment was used to resolve the boundary layer in the near walls. Heat transfer by means of radiation was neglected in the mathematical model. These assumptions result in the following set of differential equations:

\frac{\partial ρ}{\partial t} + \nabla \cdot (ρ u)

\frac{\partial ρ}{\partial t} + \nabla \cdot (ρ u u) = - \nabla p + \nabla \cdot T_{e f f} + ρ g

ρ – fluid density, kg/m³

t – time, s

u – fluid velocity vector, m/s

p – pressure, Pa

T_eff – tensor of molecular and turbulent stresses, N/m²

g – gravitational acceleration, m/s²

Table 4.

List of boundary conditions with specifications

Element:	Type of boundary conditions	Conditions related to the conservation of momentum	Thermal conditions
Nostrils	Mass-flow inlet	Mass flux defined by the user	Temperature defined by the user, T = 307 K
Face shield surface	Wall	Standard fixed walls	Zero heat flux
Fabric mask	Porous Jump	Porous materialThickness: 0.5 mmFace permeability: 9.8⁻¹⁰ m²	Zero heat flux
Body surface	Wall	Standard fixed walls	Heat flux 40 W/m²
The outer surface of the numerical domain	Pressure-outlet	Standard conditions	T = 293 K

\frac{\partial}{\partial t} (ρ k) + \nabla \cdot (ρ k u) = \nabla \cdot (Γ_{k} \nabla k) + G_{k} - Y_{k}

\frac{\partial}{\partial t} (ρω) + \nabla \cdot (ρω u) = \nabla \cdot (Γ_{ω} \nabla ω) + G_{ω} - Y_{ω}

G_k – turbulence kinetic energy production term due to average velocity gradient

G_ω – specific dissipation rate production term

Γ_k – effective diffusion coefficient of turbulence kinetic energy

Γ_ω – effective diffusion coefficient of specific dissipation rate

Y_k – dissipation term of turbulence kinetic energy

Y_ω – dissipation term of specific dissipation rate

\frac{\partial (ρ E)}{\partial t} + \nabla \cdot (u (ρ E + p)) = \nabla \cdot (k_{e ff} \nabla T - \sum_{j} h_{i} J_{i} + T_{e f f} u

E – specific internal energy, J/kg

k_eff – effective thermal conduction coefficient, W/mK

J_i – diffusion mass flux of i-th component, kg/m²s

h_i – specific enthalpy of i-th component, J/kg

\frac{\partial (ρ Y i)}{\partial t} + \nabla \cdot (ρ u Y_{i}) = - \nabla \cdot J_{i}

Y_i – a mass fraction of i-th component.

The above set of equations was discretized in space using second order upwind and second order interpolation scheme was applied to integrate equations in time. The pressure-velocity coupling model was solved by the SIMPLE scheme. The transient calculations were performed for 126 seconds of real-time which included twenty breathing cycles that allow one to reach a pseudo-steady state. A time step in the model was assumed as 0.1 s, with 30 iterations per one-time step. The simulations were run on a Windows Server 2019 Standard platform (64-bit operating system), on a workstation with 160 GB RAM memory. The host processor was Intel(R) Xeon(R) E5-2670V2 2.5GHz. Three other Intel(R) Xeon(R) E5-2670V2 2.5GHz processors were used for compute-node processes.

2.4.

Boundary conditions and initial parameters

In the numerical model the human silhouette is surrounded by air at a temperature of 20°C.

Human breathing was assumed only through the nose and a full breathing model was applied at nostrils surfaces (see section 2.5 for details). Constant heat flux was assumed at the human body surface.

2.5.

Breathing model – UDF

A realistic breathing model was assumed in the simulation. A single breath cycle lasts 6 seconds and consists of inhalation (2.5 s), exhalation (2.5 s), and pause (1 s) according to the literature [33]. Molar fractions of exhaled mixture made up of O₂, N₂, H₂O and CO₂, were defined based on the procedure described in [24]. The model of human breathing with a complete respiratory cycle allows the mapping of the actual human respiration process and enables monitoring of carbon dioxide concentration changes during inhalations. A boundary condition on the surface of the nostrils was described using the User Defined Function (UDF) described in [24]. It replaces the constant volume flow that can be defined in the Ansys Fluent software with a realistic breathing cycle (Fig. 8). The UDF function was used to define the temperature variability during individual respiratory phases, the qualitative and quantitative composition of the exhaled air, as well as the values of the inhaled and exhaled air flow rate.

Example of CO₂ concentration changes in time at the surface of the nostrils, during breathing, as it is described by developed UDF

RESULTS

3.1.

Qualitative analysis

This section shows the results of the comparison of different PPE applications by means of the contours and air flow paths coloured by CO₂ concentration, velocity and temperature during different stages of human breath. For the case with material mask, calculation took 11 hours, for the rest cases calculation took approx. 20 hours and 20 minutes. The graphics present results obtained for the 21^st breathing cycle, i.e. after two minutes of real-time simulation. The calculation results, shown in Figures 11–15, were analysed for the time equal to 0.1 seconds before the end of each, of the three characteristic phases in the cycle: inhalation, exhalation, and pause as shown in Fig. 9. The sagittal plane was used as shown in Fig. 10 to present results of CO₂ concentration, velocity and temperature profiles.

Characteristic time points in the breathing cycle used in results analysis

Location of section plane for results visualization

Contours of CO₂ concentration in the numerical domain during inhalation for the 21^st breathing cycle

Contours of CO₂ concentration in the numerical domain during exhalation

Pathlines coloured with CO₂ concentration in the numerical domain during exhalation

Contours of CO₂ concentration in the numerical domain during the pause between exhalation and inhalation

Contours of temperature (K) and velocity (m/s) on the example of models without any PPE and with a face shield located 40 mm from the face, during the pause between exhalation and inhalation

Figure 11 presents contours of carbon dioxide during inhalation for the 21^st berating cycle, (122.5 s of simulation) for five analysed cases with different personal protective equipment. Only in the base case with no personal protective equipment (A), CO₂ in exhaled air dilute in the surrounding and descends towards the bottom surface of the domain. In this case high concentration of CO₂ in the close vicinity of human face is not observed. This is caused by a lack of re-inhalation of used-up air. In other cases, inhaled air contains a significant part of previously exhaled air. That phenomenon is clearly visible for models supplied with face shields (C, D). Below the human nostrils area of CO₂, the concentration at a level of 3 500 ppm is visible, while in the area between the shield and face, CO₂ concentration stays in the range between 4 800–5 800 ppm. Depending on the distance of the shield from the face, the airflow sticks to the surface of the face shield (model C) or flows downwards along the human body (model B and D) with a small face shield and face shield at distance of 40 mm from the face respectively. The lowest carbon dioxide concentration is visible around the model equipped with a fabric mask (E).

During exhalation (Fig. 12) an instantaneous value of CO₂ concentration exceeds 53 000 ppm. After the exhaled air contact the surface of the face shield (B, C, D) carbon dioxide concentration decreases rapidly and reaches a value of 13 000–16 000 ppm. The contours of the carbon dioxide concentration take similar shapes in the cases of all three face shields (B, C, D).

In the case of the fabric mask, only a small fraction of CO₂ passes through the porous fabric of the mask. Majority of exhaled air leaks between the edge of the mask and the human face, losing speed at the same time (Fig.12 E). Exhaled air streams with low concentrations of carbon dioxide leave the mask mainly through leaks around the cheeks and nose. In the case of the small face shield (B) and the face shield moved away from the face by 40 mm (D), the stream of exhaled air behaves similarly, i.e. after contact with the face shield, it flows down to the level of the torso. In the case of the face shield located 15 mm from the face, the exhaled air hits the face shield and some of the carbon dioxide particles also scatter horizontally to the sides, which is visible in the frontal view of the CO₂ concentration pathlines (Fig. 12 C).

The flow pathlines show the directions of the spread of CO₂ and their disappearing concentration as they move away from the nostrils (Fig. 13), it can be stated that in the case of a small shield covering the nose and mouth and a full shield located 40 mm away from the face, CO₂ particles reach the lowest parts of the numerical domain, without being lifted up by the convective flux. The situation is different in the case with the shield located 15 mm away from the face, where carbon dioxide particles are begin raised upwards before they reach the bottom of the computing domain.

Maps of carbon dioxide concentrations during the pause period are presented in Fig. 14. Attention should be drawn to the high value of CO₂ concentration in the close vicinity of the human face, which takes place in the case of face shield moved 15 mm from the face (C) and for the medical mask (E). The mean concentration of CO₂ in the facial area is 13 500 ppm, with very low carbon dioxide concentration in the rest of the computing domain. In the case of a small face shield (B) covering only the nose and mouth and a face shield (D) placed 40 mm away from the face, the carbon dioxide flows down along the face and its concentration at the chest is in the range of 6 000–9 000 ppm.

The observed distribution of pollutants around the human body is also influenced by the generated heat flux and convection flux formed as a result of emitted heat by the human body. Contours of air temperature and velocity are shown in Fig. 15. After two minutes of simulation the velocity over the human head reaches values up to 0.3 m/s for model (C) and 0.1 m/s for model (D).

3.2.

Quantitative analysis

Figure 16 presents the results of varying carbon dioxide concentrations on the nostrils surface during breathing cycles. Results include the thirteen breathing cycles (from 40 to 120 seconds of simulation). During the inhalation process, significant differences in CO₂ concentrations are observed. Carbon dioxide concentration values quoted at the last moment of inspiration increase with subsequent respiratory cycles. The rate of changes and level of CO₂ concentration is different depending on the considered case. Carbon dioxide concentrations during inhalation remain constant in the range of 800–1 100 ppm for the model with no PPE (Fig 16 A). Fluctuations beyond that range occur only around 80’s – 90’s seconds of calculation. In the case of the small face shield model (Fig. 16 B), carbon dioxide concentration changes decrease from the level of 6 800 ppm to 2 600 ppm during a single inhalation phase. This can be observed in each subsequent breathing cycle.

Time-varying CO₂ concentration monitored on the surface of the nostrils

The greatest changes in CO₂ concentration during inspiration occur in the model with a fabric mask (Fig. 16 E) and in the model with a face shield located 15 mm from the face (Fig. 16 C). The observed decrease in CO₂ concentration during the single breath-in was from approx. from 12 400 to 1 300 ppm for a model with fabric mask and approx. from 9 900 to 2 200 ppm for a model with a face shield located 15 mm from the face. In those models, fabric masks and face shields are placed close to the face. This causes the lack of space between the human face and the face shield to dilute exhaled air sufficiently during the pause in the breath. This cause the inhaled air contains a large amount of exhaled CO₂. This may affect the user’s comfort.

In the model with a face shield located 40 mm away from the face (Fig. 16 D) CO₂ concertation slightly increases at the beginning of each inhalation phase and then decreases in concentration in surrounding air, which does not take place in the remaining cases. This is caused by ambient air, which flows in at a large amount during the pause phase and replaces the exhaled air from the space adjacent to the inner wall of the face shield. CO₂ concentration in the inhaled air does not exceeds 7000 ppm. The lowest inhaled CO₂ concentration significantly exceeds 2000 ppm compared to the remaining cases.

Figure 17 presents histograms of CO₂ concentration distributions in inhaled air recorded for each inhalation phase throughout the whole simulation (lasting 122.5 s). The sampling frequency is equal to 0.1 seconds (1-time step). For the model without any PPE (Fig. 17 A) 91% of recorded concentrations in inhaled air do not exceed 1500 ppm. Only 9% of records indicate concentrations above 1500 ppm, moreover, the highest recorded concentration is 1843 ppm.

Distributions of CO₂ concentration ranges during the inhalation phase

For other cases, concentration below 4500 ppm makes up about 30% of inhaled air in cases: (Fig. 17D) – model with face shield located 40 mm from the face (29.2%) and (Fig. 17 C) - model with face shield located 15 mm from the face (30.6%) or about 50% in cases: (Fig. 17 B) – model with a small face shield (47.2%) and (Fig. 17 E) – model with fabric mask (51.8%). However, in models B and D, the range of occurrence of carbon dioxide concentrations is narrower and the highest concentration values do not exceed 7500 ppm. For case B, 52.8% of registered values of concentration is within the range of 4501–7500 ppm, for case D it is 70.8%. For models C and E, concentrations of CO₂ in the range 4501–7500 ppm occur at 30.6% and 24% respectively. Concentration in the range above 7501 ppm appears in the model (Fig. 17 C) with a face shield located 15 mm from the face and in the model with a fabric mask (Fig. 17 E) and constitutes 32.8% and 24.2% of inhaled air in remaining cases respectively. There is a difference between the cases mentioned last (Fig. 17 C and E) regarding concentrations above the 9500 ppm value. While in the case of model E, concentrations above the 9500 ppm value represent only 12% of the recorded values, in the case of model C it is 21.8%.

Medians for recorded CO₂ concentrations in the inhaled air equal to:

• 1098 ppm for case (A) - model without any PPE,

• 4563 ppm for case (B) – model with a small shield,

• 6210 ppm for case (C) - face shield located 15 mm from the human face,

• 5522 ppm for case (D) with the shield moved 40 mm from the human face,

• 4343 ppm for case (E) with a fabric mask.

CONCLUSIONS AND SUMMARY

Application of different PPE worn by a human can affect the airflow pattern and contaminants distribution around the human face. Performed analysis shows that exhaled CO₂ is a good indicator of infection risk as well as re-inhalation risk. The numerical simulation is an efficient tool for analysing different scenarios of PPE application as well as for optimizing the design of PPE.

Performed simulations show that the use of personal protective equipment changes the trajectory of exhaled air and CO₂. Application of PPE reduces the concentration of contaminants marked with CO₂ in front of the face, outside the PPE contour, as they spread up and down the face. This significantly extends the time needed for contamination, originating from an infected person wearing PPE, to reach another person standing in front of an infected human.

Different PPE generates different airflow patterns in the vicinity of the human being. The most efficient in reducing infection risk is wearing a mask or face shield located close to the face, as it most efficiently reduces the spreading of CO₂. However, in that cases, the CO₂ re-inhalation risk with very high CO2 concentration was the highest. Wearing a face mask or face shield very close to the face reduces the comfort of the human being. It should be noted that the calculations covered a period of only two minutes. To draw far-reaching conclusions, it is necessary to extend the simulation time.

Optimisation of the PPE as a compromise between reducing infection risk and user comfort is a further challenge.

eISSN:: 2720-6947
Language:: English

Publication timeframe:: 4 times per year
Journal Subjects:: Architecture and Design, Architecture, Architects, Buildings

Journal RSS Feed

Numerical Analysis of the Impact of the Use of Personal Protective Equipment on the Face in the Process of Pollutants Spreading Emitted During Breathing

Published Online: Apr 24, 2023

Page range: 113 - 130

Received: Nov 10, 2022

Accepted: Feb 27, 2023

DOI: https://doi.org/10.2478/acee-2023-0009

Keywords
CFD, human breathing, SARS-Cov-2 pandemic, infection risk, personal protective equipment

© 2023 Anna Bulińska et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Figure 11.

Figure 12.

Figure 13.

Figure 14.

Figure 15.

Figure 16.

Figure 17.

Numerical Analysis of the Impact of the Use of Personal Protective Equipment on the Face in the Process of Pollutants Spreading Emitted During Breathing

Published Online: Apr 24, 2023

Page range: 113 - 130

Received: Nov 10, 2022

Accepted: Feb 27, 2023

DOI: https://doi.org/10.2478/acee-2023-0009

KeywordsCFD, human breathing, SARS-Cov-2 pandemic, infection risk, personal protective equipment

© 2023 Anna Bulińska et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Figure 11.

Figure 12.

Figure 13.

Figure 14.

Figure 15.

Figure 16.

Figure 17.

Keywords
CFD, human breathing, SARS-Cov-2 pandemic, infection risk, personal protective equipment