Investigating the impact of a window air conditioner with H‑14 HEPA filter on lessening SARS‑COV‑2 aerosols

,

the World Health Organization (WHO) including quarantine, social distancing, breakdown in the health management system, and scarcity of personal protective equipment (PPE) such as oxygen cylinders; persons with symptoms that match the case description for suspected should be provided a medical mask and sent to a single room as soon as possible.If a single room is not practicable, segregate patients with comparable clinical diagnoses and epidemiological risk factors in separate rooms.Suspected and confirmed cases should not be grouped together.Clinicians should be aware of differences in clinical presentation based on the patient's age, since physiologic, pathophysiologic, and cognitive-psychological responses and capacities differ dramatically across individuals.People living in deprived areas of the world, where poverty and inadequate healthcare services exist, are more prone to infection and disease transmission [2].WHO also approved the emergency use of eight COVID-19 vaccines, offering a glimmer of hope for the global population amidst the pandemic.These vaccines included AstraZeneca, Janssen, Serum, Sino Pharm, Sinovac, Bharat Biotech, Moderna, and Pfizer [3].
Canadian Committee on Indoor Air Quality (CCIAQ) recommends maintaining a relative humidity level between 40 and 60% in indoor environments during the COVID-19 outbreak.This recommendation is based on evidence that a relative humidity below 40% may increase the susceptibility to infection, while a relative humidity above 60% may lead to water condensation, chilly walls, and the growth of mildew.The Federation of European Heating Ventilation and Air Conditioning Associations (REHVA) stated that adjusting temperature and relative humidity set-points is ineffective because the virus can readily survive within the occupants' thermal comfort zone, which pointed out that the virus has been shown to remain viable for 14 days at 4 °C, 1 day at 37 °C, and 30 min at 56 °C, while temperatures above 37 °C are impractical in indoor settings, and emphasized the importance of operating toilet ventilation systems 24 h a day.Additionally, residents were advised to flush toilets with the lids closed and to check water seals every 3 weeks to prevent virus spread [4].
Engineering procedures revealed by American Society of Heating, Refrigerating and Air Conditioning Engineers (ASHRAE) including peer-reviewed studies with recommendations for HVAC systems that "open minimum exterior air dampers, as high as 100%, thus avoiding recirculation, and when the outside air quality and weather conditions are suitable, opening windows may be a viable alternative, but this leads to faster air flow and higher energy consumption because it raises the rate of air exchange during the hour.As well as the relative humidity and temperature of the outside air cannot be regulated, such natural ventilation cannot assure the thermal comfort of interior inhabitants [5].Nina Szczepanik [6] revealed a study, which examined how a faulty gas furnace leaking carbon monoxide (CO) could affect people in a house.The furnace was in an attached garage.Computer simulations to see if a special air terminal device (ATD) with an adjustable design could improve air quality are used.This ATD aimed to keep airflow strong in the garage even when the overall ventilation system changed.A strong airflow would remove CO from the furnace and stop it from entering the house.The study found that the new ATD design maintained good airflow and reduced CO infiltration.
Senha Edupuganti et al. [7] addressed a study related to an Internet-of-Things (IoT)based air pollution monitoring system as a response to the critical concern of air pollution in many urban areas, which caused serious health problems and environmental damage.This mechanism was designed to measure various air quality parameters in real time, including temperature, humidity, various gases, microbes, and light intensity.The system consists of sensor nodes, a gateway, a WIFI module, an LCD display, and a cloud server.Sensor nodes were placed at different locations to take measurements of air quality parameters, with the data then being transmitted to the gateway via wireless communication.The system could even send alerts to users when air quality is poor, allowing them to take precautions.By providing real-time information about air quality, this system could aid decision-makers in taking appropriate measures to reduce air pollution and safeguard the health of the community.
Salama et al. [8] demonstrated that HEPA filtration, in line with Centers for Disease Control (CDC) recommendations, can help curb the spread of COVID-19 infection.By enhancing the physicochemical properties of HEPA filters, researchers have explored the potential of incorporating various nanomaterials as disinfectants.The effectiveness of silver and titanium dioxide nanoparticles in viral inactivation is attributed to their unique size, shape, and phases.These nanoparticles exhibit significantly higher antimicrobial activity compared to other materials due to their exceptional chemical stability, catalytic properties, and enhanced thermal conductivity.Adding these nanoparticles to a HEPA filter will improve air quality without reducing air filtering effectiveness when there were infectious aerosols present.
Ans Al Rashid et al. [9] investigated the use of numerical modeling to simulate an automobile fuel cell system that employed a two-stage turbo-compressor for supplying air.Their numerical model incorporated key input parameters related to both air and hydrogen flow.Recognizing the limitations of single-stage compressors in terms of efficiency and pressure ratio, the compressor design process, aided by the Cf-turbo tool, began with a numerical analysis of a preliminary design for a highly efficient two-stage turbo-compressor with an expander.The final compressor design achieved a total pressure and temperature of 4.2 bar and 149.3 °C, respectively.It required 20.08 kW of power, with 3.18 kW of power losses, resulting in a combined efficiency of 70.8%.
Narjisse Amahjour et al. [10] proposed a framework to assess the influence of HVAC and sanitizer system design on COVID-19 transmission within an isolation unit at Saniat Rmel Hospital in Tetouan, Morocco.This framework utilized CFD and the realizable turbulence model.The study further investigated the impact of varying inlet air velocities on the dispersion of airborne viral particles inside the isolation unit.The findings indicated that the air change rate played a critical role in managing the spread and concentration of pathogens within the unit.A high inlet air velocity could expedite the movement of airborne particles, including viruses, towards the exhaust port.
Shamim et al. [11] demonstrated that using ultraviolet germicidal irradiation (UVGI) in building HVAC systems is an effective method for inactivating biological pollutants in indoor air.The concept of using UVGI to kill airborne viruses has been around for a while.UVGI is categorized into three subcategories based on irradiation wavelength: UV-A (320-400 nm), UV-B (280-320 nm), and UV-C (200-280 nm).Due to their potent germicidal properties, UV-C rays can effectively inactivate COVID-19 particles.Consequently, incorporating UV-C technology into HVAC systems holds immense potential for providing occupants with clean air free from viruses and bacteria.Saw et al. [12] revealed the effectiveness of the indoor air purifier in preventing the dispersion of SARS-COV-2 aerosols to characterize the dispersion of these particles virus which was difficult to observe in the experimental visualization.Assumptions are used to model the dispersion of the aerosol particles from the patients in the common ward of a hospital in Malaysia with multiple air purifiers at floor level.Starting with unsteady Lagrangian particle tracking, neglect evaporation and mass transfer of the particles expelled from the patients since the particle sizes were very small with range 0.07, 0.1, 0.3, 0.5, 0.8, 1.0, 2.0, 5.0, 8.0, and 10 µm in diameter, and the particles with ideal sphere shape.
Nazarious et al. [13] revealed an experiment at Imperial College, England, on a fan coil unit contains centrifugal blower, hood meter, breathing tube, power supply, and power bank used in cleanroom constructions.The basic working principle of a fan coil unit (FCU) with a high-grade H-14 HEPA filter is to get the purified air eliminated from SARS-COV-2 aerosols.
Subramanya [14] established a parametric model of an impeller of a vacuum cleaner with a HEPA filter to capture the SARS-COV-2 aerosols over surfaces and grounds to keep the houses secure.The design was constrained by the dimensions needed to fit the fan into the machine with CAD program CATIA V. 5 to model the entire simulation geometry, which was then loaded into the commercial CFD program STARCCM + V. 2020.2.The full simulation model was applied with a prism layer region to capture the flow field in the boundary layer region.The flow near the boundary region could be captured more correctly by the MENTER SST k-Omega turbulence model than by the k-epsilon model.Because of the presence of unfavorable pressure gradients, which was predicted in this application, it could capture flow separation that happens in the boundary region of the blade channels.Therefore, the SST k-Omega turbulence model was employed.
The objective of this study is to integrate the H14-HEPA filter with window AC, model GJC07AF-K3RNB9D through incorporating special features for the indoor blower.A new design of the blower will be achieved through CF-TURBO tool, and the CFD tool available, STARCCM + , approved by the German SIEMENS firm to simulate this blower with H-14 HEPA filter to capture SARS-COV-2 aerosols.

Case study
Setting up a modified version of the window AC, model GJC07AF-K3RNB9D, to be compatible with a HEPA filter aims to provide continuous protection for the mentioned office against COVID-19 particles.Figure 1 illustrates a familiar sight in homes and businesses around the world.The front panel houses the air intake louvers, which allow room air to be drawn in the unit.The side panels enclose the internal components of the window AC, protect them from dust and debris, and have a ventilation opening to facilitate airflow and prevent overheating.
Table 1 illustrates additional components integrated into the original window AC to transform it into an air purifier, capable of eliminating COVID-19 particles within the mentioned office.The inclusion of a diffuser before the HEPA filter is designed to reduce the velocity of the discharged air, thereby mitigating the risk of a significant pressure drop through the filter.Given that the maximum possible Mach number through the modified window AC is less than 0.3, it is reasonable to consider the air as incompressible [16].When installing HEPA filter, it is preferable to do that at the discharge rather than after the pre-filter.If done in the opposite order, the effectiveness of the HEPA filter may be compromised.There is a possibility that if a technician with an infection performs maintenance on the air conditioner, virus particles could be transferred to the internal parts of the unit.Conversely, the air may carry another infection from inside the air conditioner, leading to the recirculation of particles back into the office.

Indoor modified impeller design procedures
Table 2 illustrates the procedures employed to determine the parametric model of the modified blower by applying appropriate dimensions to optimize its performance.
The CFTURBO tool treats the flow as a one-dimensional streamline, despite the actual flow being unsteady and three-dimensional.Input operation conditions related to the selected type of turbomachine are required and can be manually adjusted.The main dimensions menu serves as the foundational basis for all impeller design steps, showcasing various blade shapes, including free-form three dimensions, ruled surface three dimensions, and circular form two dimensions [19].

Blower design governing equations
Numerous empirical models have been developed to calculate the slip coefficient, indicating that a smaller slip factor results in a higher deviation of the flow compared to the ideal direction given by the blades [19], starting with inlet and outlet meridional flow coefficient, φ as a ratio of radial component of absolute velocity C, and peripheral velocity U as expressed in Eq. ( 1).
Peripheral velocity is tangential with rotation direction of impeller as expressed in Eq. ( 2).
The slip coefficient is a function of outlet blade angle, β 2 number of blades, Z and/or the meridional geometry.Wiesner developed an empirical formula for the estimation of the slip coefficient, σ WI as expressed in Eq. (3).Gulich modified Eq. ( 3) by considering the size of the impeller hub and shroud in radial flow direction for the estimation of the corrected slip factor, σ GU as expressed in Eq. ( 4).
(1) Aungier adjusted the empirical formula of Wiesner for the estimation of the corrected slip factor, σ AU as expressed in Eq. ( 5) without considering the size of the impeller hub and shroud.
Von Backstroem developed an empirical formula σ VB as expressed in Eq. ( 6) for the estimation of the slip coefficient, assuming one single relative eddy in the rotor.

CFD validation of the modified blower
Exporting a process to STARCCM + involves several preparatory steps that ensure valid meshing, such as the following: (1) geometry repair in which undesired sections on the design should be eliminated to ensure a smooth and consistent geometry, (2) surface repair to remove any duplicate faces, and (3) assign parts to appropriate regions to initialize the meshing process with the correct settings as shown in Table 3. Meshing is a crucial foundation of CFD analysis.It involves discretizing the computational domain into a finite number of control volume cells to accurately predict the physics of fluid flow.Applying surface remeshing before mesh cell generation helps prepare a high-quality surface for the required blower and its volute casing [14].
As shown in Fig. 2, unstructured cells are applied to the entire design.This type of mesh is efficient to generate and uses less memory than structured mesh, which is much more time-consuming to generate.Polyhedral cells are used because they have very good mesh properties and provide better approximations of gradients.Prism (4) Table 3 Meshing properties over the whole modified blower

Meshing type Automated unstructured
Meshing pre-processing Surface remesher

Cell shape Polyhedral
Base size Stop at convergent results

Prism layers
To capture velocity gradients layers are applied to the entire model to capture the flow field in the boundary layer region, making it easier for the user to obtain an adequate solution [14].
As shown in Table 4, the segregated flow solver is utilized because it is suitable for incompressible fluids and solves the governing equations individually, leading to easier convergence compared to the coupled solver.The selection of the turbulence model depends on factors such as the physics involved in the flow, established practices for specific problem classes, the required accuracy level, available computational resources, and the time constraints for the simulation.
The Reynolds-Averaged Navier-Stokes (RANS) simulation method is employed due to its computational efficiency, reliability, and widespread adoption in industrial applications [20,21].The Menter SST k-ω turbulence model is particularly wellsuited for capturing near-wall flow behavior and flow separation in boundary regions, which is anticipated to occur in the current work due to the presence of adverse pressure gradients.This makes it a more precise choice compared to the k-epsilon model.
Boundary conditions inform the CFD solver regarding the interaction between the computational domain and its surroundings.The inlet condition specifies a velocity inlet, allowing fluid mass and momentum to enter the computational domain.The outlet condition is set as a pressure outlet to monitor for any reversed flow at the discharge section caused by the blower's high pressure.
Component motion is simulated using two reference frames: a fixed frame for stationary components like the volute and a moving frame for rotating components like the  blower impeller.This approach avoids the need to move mesh vertices, simplifying the solution process [14].Ffowcs Williams-Hawkings (FW-H) is an exact rearrangement of the continuity and momentum equations into the form of an inhomogeneous wave equation.It provides accurate results even when the integration surface lies within the nonlinear flow region [22].

CFD governing equations
The fundamental governing equations of fluid dynamics, including continuity and momentum equations, are employed in CFD to analyze complex fluid-solid interaction problems as presented here.Turbulence models, such as the Menter SST k-ω model, are utilized to capture the intricacies of turbulent flow.Approaches such as FW-H and discrete phase modeling techniques like Lagrangian particle tracking are also employed to address specific flow phenomena [23].

Continuity equation
The continuity equation, a statement of mass conservation, describes the relationship between density and velocity, as expressed in Eq. (7) where ρ Ai [kg m −3 ] is the air density, t [s] is the time, V i [m s −1 ] is the velocity vector component, and x i [m] is the position vector component.

Momentum equation
The momentum equation, which is a manifestation of Newton's second law of motion, describes the relationship between pressure and velocity as expressed in Eq. (8) where p Ai [pa] is the air pressure, υ [m 2 s −1 ] is the kinematic viscosity, and τ ij [pa] is the stress components.

Menter SST k-ω equation
The Menter SST k-ω model is a hybrid model that combines the strengths of the k-ω model in the near-wall region and the k-epsilon model in the bulk flow region.This combination provides satisfactory performance for boundary layers with adverse pressure gradients.(7) where k [J] is the kinetic energy, µ [pa s] is the dynamic viscosity, ρ [-] is the stress vor- tex stretching modification factor, γ is the production of turbulence eddy [J kg −1 s −1 ], ω [J kg −1 s −1 ] is the dissipation rate, is the free shear modification factor [-], F [-] is the blending factor, and ϕ [-] is the model coefficient.

Energy equation
The energy equation, as expressed in Eq. ( 13), is used in various fields as thermodynamics and fluid dynamics principles.Although there is no heat transfer in the present work, there is a work done by the rotated impeller on the airflow through the entire blower, which makes the total pressure (static, dynamic, and elevation) not constant.
where E [J] is the total energy, ∇[-] is the Nabla operator,̺ [pa] is the viscous stress ten- sor, and R [J] is the work done.

FW-H approach for acoustic wave propagation
It is important to set limitations on rotational speed of the required blower to get reasonable sound levels, suitable for the comfort of persons inside the room.Inserting a considered value of speed in the CFD tool specifies the sources of noise, which are the blower region and point receiver, which are different locations inside the mentioned office.STARCCM + tool provides an Excel sheet, which contains discretized points related to the sound levels, then inserting this sheet again in CFD tool, and specify the data related to the sound pressures to represent the results as graphical visual information.These procedures are indicated as: where P s [pa] is the total surface term of the receiver, based on the free space to compute the sound pressure at the observer locations as P T [pa] is the summation of the mono- pole term,P L [pa] is the dipole term, and P Q [pa] is the quadrupole term, as expressed in Eq. ( 14) [23].
For general flow and for flows with rigid body motion or moving reference frames, the displacement of fluid as the body passes generates the noise in monopole term, as expressed in Eq. ( 15) [23].
where Q o [m 3 s −1 ] is the flow noise due to the operation of the blower, θ [m] is the dis- tance from source to receiver, and M [-] is the Mach number.
As sound is a propagation pressure wave, it is important to obtain the hydrodynamic pressure oscillations in the air in dipole term, as expressed in Eq. ( 16) to calculate the propeller noise.(12) where L i [pa m −1 ] is the loading blade component.Non-linearities in the flow generate quadrupole noise, as expressed in Eq. ( 17).
where T ij [pa] is the Lighthill stress tensor components.
The flow noise due to the operation of the blower is expressed in Eq. ( 18).
where ρ o [kg m −3 ] is the far field density, V Ai is the air velocity component in axial direc- tion, V Sur [m s −1 ] is the surface velocity component in axial direction, and A n [m 2 ] is the surface normal area vector.
As expressed in Eqs. ( 19) and ( 20), the loading blade component, which depends on compressive stress tensor, is as follows where L i [pa] is the loading blade component, V n [m s −1 ] is the surface normal velocity vector,P ij [pa] is the compressive stress tensor, p o [pa] is the far field pressure, ξ ij [-] is the Dirac wave function, and ̺ ij [pa] is the viscous stress tensor component.

Lagrangian particle tracking model
The continuous phase of aerosol transport and deposition is air, while the discrete phase, which is tracked using a Lagrangian technique with input features as shown in Table 5, is the droplets with different ranges of 0.07, 0.1, 0.3, 0.5, 0.8, 1.0, 2.0, 5.0, 8.0, and 10 µm.The interaction between fluid phase and the discrete phase is assumed to be one-way coupling, and the impact from the discrete phase to the fluid phase was negligible, since the COVID-19 particles have very small sizes [12].
Lagrangian particles [16] are introduced into the flow field using injectors.The droplets are injected randomly from the inlet of H-14 HEPA filter with part injection mode (16 and Nukiyama-Tanazawa method with 0.01 s time step.The droplets in the Lagrangian move as a result of forces, specifically the forces of buoyancy, drag, and Brownian.The difference in densities between droplets and air determines the buoyancy force as expressed in Eq. ( 21).
where F b [kg m s −2 ] is the buoyancy force, d par [m] is the injected particle diameter, ρ par [kg m −3 ] is the particle density, ρ Ai [kg m −3 ] is the air density, and g [m s −2 ] is the gravi- tational acceleration.
Drag force mainly depends on the relative velocity of the droplets, their size, and drag coefficient as expressed in Eqs. ( 22) and (23) where [-] is the drag coefficient, and m par [kg] is the injected particle mass.
The random movement of the smaller droplets due to Brownian motion is described as expressed in Eq. (24).
where F k [kg m s −2 ] is the Brownian force, [-] is the vector whose components are ran- dom numbers which follow a Gaussian distribution with mean zero and a variance of 1, and S [m 2 s −3 ] is the spectral intensity of the noise process.

Results and discussion
Upon analyzing the materials data using the two previously defined programs, the findings are presented in this chapter.To evaluate its performance prior to manufacturing, the designed blower will be simulated, and all extracted data from the design and CFD tools will be presented in the form of graphs or tables.These engineering visualization techniques provide insights into the performance of the existing system.

Design performance of modified blowers by CF-TURBO
As shown in Table 6, the AUNGIER model is a more effective choice for fan design due to its suitability for this application [19].Employing an odd number of blades helps to disrupt air waves, which exist naturally in a sinusoidal pattern, leading to significantly reduced sound levels caused by air turbulence.This, in turn, contributes to balanced blower operation [24].
Table 7 illustrates the final modified operating point of the desired blower.The CFturbo tool enables the input of the desired operating point of the designed blower, thereby facilitating the determination of its complete geometry.
Table 8 presents the final modified operating point of the desired blower after designing a new one with the same blade shape as described in Table 7.Additional modifications include increasing the number of blades, enlarging the suction diameter by 20 mm, and reducing the width by 24 mm.Slip factor is a crucial parameter in centrifugal blower design, and an optimal value of 0.9 [25] is employed to achieve an accurate solution for the overall blower performance, ensuring efficient energy transfer to the air.Modifying the original blower's dimensions will necessitate changes to the air conditioner's dimensions, requiring the adaptation of internal components to accommodate the modified blower.
Figure 3 shows the geometry of both modified blowers with the squirrel cage impeller, free-form three-dimensional backward curved blades, hub, and shroud sections.These blowers have a good quality surface to remove undesired duplicate faces to be prepared for meshing.
In turbomachinery, velocity diagrams are graphical representations of the various components of the working fluid's velocities at both the inlet and outlet sections of the indoor blower.These components include the peripheral velocity U, relative velocity W, and absolute velocity C, as illustrated in Fig. 4. Additionally, several crucial parameters are associated with these velocities, including the absolute angle α, blade angle β , and meridional flow coefficient φ [19].Table 9 illustrates that after extracting the velocity triangles from CF-TURBO, the airflow enters the blower radially as evidenced by the inlet absolute angle of 90°.The outlet blade angle being less than 90° indicates the presence of a backward-curved blade, which generates higher pressure compared to a forward-curved blade.Furthermore, the low meridional flow coefficient is a consequence of modifying a blower for a window AC unit, which typically has a low airflow rate requirement [19].

CFD STARCCM +
As shown in Fig. 5, part A reveals an airflow of blower 1; part B reveals a reversed flow at discharge of blower 1, especially at the wall, which will cause high-pressure waves, and hence make the blower louder; part C shows a nozzle 1 with length 15 mm, end width 155 mm, and 65 mm end height is attached to volute of the blower 1 in as an attempt to estimate the pressure drop through it to add this value to the new design of blower 2 to satisfy the total pressure drop of all components; and part D shows a six sections with 3 mm step, which describes that decreasing the static pressure with a pressure drop equals around 100 Pa.
To prevent the occurrence of the reversed airflow, a nozzle should be installed at the discharge of the volute.However, the nozzle will introduce a pressure drop that was not considered during the calculation of total pressure across the indoor components of the window AC, so the blower 1 does not perform its function.

Parametric study of the blower 2 outlets with different nozzles
Figure 8 shows that a nozzle 1 of blower 1 and different four nozzles of blower 2 were simulated to predict the levels to estimate the best one [26] through using the point receiver analysis inside the office at 2 m from the blower while the air conditioner was operating, and also the total pressure at exit of each nozzle to ensure the ability to maintain the required total pressure of the new version of the selected window AC.

Volute geometry of blower 1 and blower 2
Figure 9 illustrates a single-spiral volute of the blower 1, and blower 2, which also have a smooth surface to facilitate the removal of any remaining duplicate faces in preparation for meshing.To minimize air leakage, maintaining a recommended clearance between the blower and the volute is crucial [27].Table 11 illustrates the effective design of the volute casing of blower 1 and blower 2. Selecting the nozzle 5 with blower 2 as the best case, as it maintains suitable sound pressure levels in the office unlike the rest of the nozzles; in addition, it could prevent the reversed flow at the discharge of its volute, and satisfies approximately reasonable pressure drop across it as mentioned previously.

Description of blower 2 with nozzle 5 and new version of window AC
Random particles were injected at the inlet of the H-14 HEPA filter with a solution time of 150 s, and the results became approximately constant.COVID-19 particles with a diameter of 0.1 µm, as shown in Fig. 10, are able to pass through this filter due to its very high porosity.The total pressure over the entire blower 2 with nozzle 5, as shown in Fig. 11, is high at the side with a small volute thickness due to airflow friction with its blades.The  pressure then starts to decrease as the volute thickness increases.The negative pressure value at the suction side indicates that the blower draws air at this section.
As shown in Figs. 12 and 13, a performance analysis from CFTURBO tool reveals that both modified blowers can achieve the required pressure drop.However, blower 1 exhibits problems with reversed flow at the discharge and excessive sound pressure levels.Consequently, blower 2 is the optimal solution for the modified window AC.The findings indicate that blower 2 exhibits an 18.3% increase in average total pressure and a 0.2% increase in average output power compared to blower 1.However, its average overall efficiency shows a slight decrease of 0.06%.This decrease in efficiency is attributed to the higher rotational speed of blower 2.
Modeling a design for modified version of window AC with AUTOCAD 2016 tool as shown in Fig. 14, in addition to export the whole design of blower 2 with nozzle 5 from Cf-turbo tool to AutoCAD, and combine the whole system together.

Conclusion
In summary, this study conclusively demonstrated, through meticulously designed computational fluid dynamics (CFD) simulations, that a high-pressure backward-curved blades centrifugal blower offers a clear solution to the pressure drop impediment hindering the integration of HEPA filters into window air conditioners.This research presents a paradigm shift for improving indoor air quality (IAQ) and safeguarding public health, which translates to a significant reduction in the spread of SARS-COV-2 aerosols, fostering healthier and safer indoor environments for homes, offices, and public spaces.
In this paper, performance enhancement of a centrifugal blower by designing discharge nozzles is carried out by numerical simulation.It is found that the best one for blower 2 is nozzle 5, as it has better flow profile at discharge, and would make it operate with a reasonable sound pressure level of 47 dB.The findings also indicated that the blower 2 would exhibit an 18.3% increase in average total pressure and a 27% increase in input power compared to blower 1.

Appendix
As shown in Fig. 15, a diagram analysis explains the steps followed to reach the required blower of the selected window AC to make the reader feel with the achievement and be a roadmap to the procedures of using CF-TURBO in design other different turbo machines.

Fig. 5 1 Table 10 LengthFig. 6 5 Fig. 7
Fig. 5 Schematic of blower 1 of 350 Pa.A Velocity vector analysis at whole blower, (B) velocity vector analysis at discharge, static pressure analysis over the nozzle 1, and (D) average static pressure analysis over the nozzle 1

Fig. 8 Fig. 9
Fig. 8 Parametric performance of the different nozzles modified blowers

Fig.Fig. 13
Fig. Performance curves of both modified

Fig. 14
Fig. 14 Analysis of the new version of the window AC with H-14 HEPA filter

Fig. 15
Fig. 15 Diagram of the modified blowers from design to simulation

Table 1
Extra components for the window AC, model GJC07AF-K3RNB9D

Table 2
Initial description of the modified blower

Table 4
Physics and boundary conditions

Table 5
Analysis of Lagrangian particle tracking applied for the H-14 HEPA filter

Table 6
Quantitative description of modified blowers

Table 7
Final description of 350 Pa, blower 1

Table 8
Final description of 450 Pa, blower 2 Table 10 reveals four different nozzles, applied at discharge of volute of blower 2 as shown in Figs. 6 and 7 to select the best one.

Table 9
Quantitative description of the velocity triangles of both blowers

Table 11
Effective design of the volute casing of both blowers