In many industries, cyclone separators are frequently employed to remove solid particles from the fluid flow. Cut-off diameter is recognized as a significant parameter to evaluate the performance of cyclone separators in addition to pressure drop. Computational Fluid Dynamics (CFD), a powerful computer-based method, can precisely estimate the cut-off diameter of cyclone separators. There is no arguing, however, that the CFD technique is computationally expensive and practically difficult. This research has suggested a more precise, computationally proficient hybrid CFD–DL method to improve the accuracy of cut-off diameter prediction in coarse-grid simulations for cyclone separators. It has been demonstrated that the proposed method not only requires less computational cost than typical CFD, but also delivers more accuracy results (with mean error less than 5.1% compared to experimental data). In other words, it takes advantage of the promise of a novel approach to decrease computational time while enhancing accuracy for CFD simulations.

A liquid jet injected from a nozzle stretches in the direction of the applied electric field and subsequently emits multiple droplets on pinch-off. It is hypothesized that the size and frequency of the drops can be manipulated by varying the electric field strength and flow rate, which is essential in electrohydrodynamic jet printing. Numerical simulations of the electrified jet breakup are performed using an in-house code based on the Dual Grid Level Set Method. To comprehend the impact of the electric and hydrodynamic forces, three dimensionless parameters are introduced: injection velocity (0.01≤Vinj≤0.15), Reynolds number (10≤Re≤100) and electric Bond number (0≤Boe≤30).Our results indicate the transition from dripping to micro-dripping and micro-jetting (at high Vinj) with increasing Boe. The critical Boe, at which this transition occurs, is found to escalate with an increase in Re. The stability of the micro-dripping mode is characterized by the balancing of the electrical and viscous shear stresses. Although the influence of the electric Bond number on the jet breakup length is marginal, it has a significant effect on the droplet diameter and formation frequency. A re-formulated scaling law suggests that the jet breakup length is mainly affected by injection velocity and Reynolds number.

Liquid jet impinging onto a surface occurs in many industrial process such as nuclear facilities where a part of radioactive material is handled in liquid form. In the case of accidental leak of this liquid, the airborne particle release, in droplets form, is important to quantify since it is the vector of radioactive air contamination. In the literature, while droplets splashing by drop impact have been extensively studied, only few data are available concerning the airborne particle release fraction and the case of liquid jet impact is even less studied. The purpose of this work is to measure aerosol airborne release when a circular liquid jet impacts a solid surface. We found, when the liquid jet is in the Rayleigh regime, so that the jet is broken into multiple drops before impact, the inertia of the impacting drops influences the amplitude of the aerosols mass size distribution but does not change its shape and consequently the aerodynamic mass median diameter. We also show that particle airborne release depends on the impacting Weber and Ohnesorge numbers through the so-called splashing number K which characterizes the splashing transition. We finally propose a quantitative prediction of the aerosol airborne release fraction, valid for Re ∼ O(103−104) and We ∼ O(102−103), opening the way to a more general model.

Despite the significant increase in research on mask filtration testing since the COVID-19 pandemic, there remains considerable ambiguity regarding which parameters affect particle filtration efficiency (PFE) and how differences in standard testing protocols can lead to divergent PFE values. We evaluated the PFE (and differential pressure) of several common face masks and community face mask materials including woven cotton, spunbond polypropylene, and meltblown polypropylene, testing in accordance with ASTM F2100/2299 standards for medical masks, using polystyrene latex (PSL) aerosol, as well as NIOSH standards for respirators, using NaCl aerosol. In both cases, PFE was measured with and without aerosol charge neutralization, which is used to bring the particle population to a known, equilibrium bipolar charge distribution. Aerosols of either composition that were untreated (not neutralized) led to significant increases in measured PFE, especially in the case of PSL. In contrast, effective neutralization led to lower PFE measurements that also showed little to no dependence on aerosol composition across most materials. To investigate further, the bipolar charge distributions of PSL and NaCl aerosols, both neutralized and untreated, were characterized using an aerodynamic aerosol classifier operated in tandem with a scanning mobility particle sizer (AAC-SMPS). This technique illustrated the differences in the distribution of particle charge states between PSL and NaCl aerosols of the same size, and between PSL particles of different sizes, revealing the presence of highly charged particles in many cases. Most importantly, the equilibrium charge distribution after neutralization is shown to be independent of particle composition or initial charge distribution, highlighting the crucial role of aerosol charge neutralizers in preventing overestimates of mask performance (due to electrostatic effects) and promoting consistency in standard testing procedures.

The main objective of this study is to compare the soot properties and pressure sensitivity of the high-carbon-number n-alkane and cycloalkane in laminar diffusion flames at elevated pressures. As typical high-carbon-number components of aviation kerosene, n-dodecane and decalin have been used in surrogate fuel kinetic models for aviation kerosene such as Jet-A and Chinese RP-3. In this study, quantitative experiments on soot properties are conducted in nitrogen-diluted n-dodecane and decalin laminar diffusion flames at up to 5.0 atm, which has scarcely been reported in the literature. Soot morphology and detailed soot volume fraction distribution are studied by transmission electron microscopy analysis and laser extinction method. Based on the soot volume fraction results, the soot yield is evaluated to compare the pressure sensitivity of the soot propensity between the two fuels. The results show that from 1.0 to 5.0 atm, the primary soot particle diameter, maximum soot volume fraction and soot yield in decalin flames are higher than those in n-dodecane flames. However, with increasing pressure, the growth rates of these soot properties are higher for n-dodecane flames than for decalin flames. The ratio of maximum soot yield of decalin to n-dodecane decreases from 2.2 at 1.0 atm to approximately 1.2 at 5.0 atm. The quantitative results indicate that the pressure sensitivity of soot properties of n-dodecane is stronger than that of decalin. The quantitative experimental results of this work are helpful in understanding the relevant fuel kinetic mechanisms suitable for simulating soot formation at elevated pressures.

To reduce the adverse impact of civil aviation on local air quality and human health, a new international standard for non-volatile Particulate Matter (nvPM) number and mass emissions was recently adopted. A system loss correction method, which accounts for the significant size-dependent particle loss, is also detailed to predict nvPM emissions representative of those at engine exit for emissions inventory purposes. As Particle-Size-Distribution (PSD) measurement is currently not prescribed, the existing loss correction method uses the nvPM number and mass measurements along with several assumptions to predict a PSD, resulting in significant uncertainty.Three new system loss correction methodologies using measured PSD were developed and compared with the existing regulatory method using certification-like nvPM data reported by the Swiss and European nvPM reference systems for thirty-two civil turbofan engines representative of the current fleet. Additionally, the PSD statistics of three sizing instruments typically used in these systems (SMPS, DMS500 and EEPS) were compared on a generic aero-engine combustor rig.General agreement between the three new PSD loss correction methods was observed, with both nvPM number- and mass-based system loss correction factors (kSL_num and kSL_mass) within ±10% reported across the engines tested. By comparison, the existing regulatory method was seen to underpredict kSL_num by up to 67% and overpredict kSL_mass by up to 49% when compared with the measured-PSD-based methods, typically driven by low nvPM mass concentrations and small particle size. In terms of the particle sizing instrument inter-comparison, an agreement of ±2 nm for the GMD and ±0.08 for the GSD was observed across a range of particle sizes on the combustor rig. However, it was seen that these differences can result in a 19% bias for kSL_num and 8% for kSL_mass for the measured-PSD-based methods, highlighting the need for further work towards the standardisation of PSD measurement for regulatory purposes.

This paper demonstrates the effects of C3H3/C4H5/i-C4H5/CN radicals on the formation and growth of aromatic hydrocarbons, particular types of polycyclic aromatic hydrocarbons (PAHs) such as heavier-PAHs, cyclopentafused PAHs (CP-PAHs), through reactive force field molecular dynamics simulations under 2250K and 2500K. Adding non-acetylene radicals enhanced the early-stage mass growth of aromatics to varying degrees. However, the later-stage mass growth and aromaticity degree were weakened by adding radicals except for the C3H3 radical. The reactivity of acetylene and additional radicals showed that the additional C3H3 radicals accelerated the consumption of radical pools, and the interaction reaction between C3H3 was responsible for the consumption of radical pools for the C3H3 additive case. The consumption of i-C4H5 was significantly faster than that of C4H5, whereas the consumption of acetylene in the two simulation cases was essentially the same. Thus, the i-C4H5 simulation case formed more intermediate species C6H7 than C4H5 radical. As for the formation of less-studied PAHs, i-C4H5 radicals generated more heavier PAHs and CP-PAHs than straight butyl C4H5 isomers. Finally, the effect of nitrogen-containing radical CN, which normally resides in ammonia fuel, was analyzed. Several factors explained the stronger enhancement in aromatic mass growth, but the relatively lowest aromatization degree resulted from the additional CN radicals. The reactions between CN and acetylene formed carbon–nitrogen species that participated less in the formation and growth of aromatics. The potential reactions between CN radical and aromatics led to ring-opening reactions and the formation of nitrogen-membered rings, which accelerated the mass growth of carbon clusters but inhibited the development of aromaticity degree. The least amount of CP-PAHs was found in CN cases compared with hydrocarbon radical cases.

Small air ions have the ability to charge airborne particles, thereby increasing their accumulation on surfaces. Indoor air purification by applying ionization uses electrostatic particle deposition. Respiratory pathogens, including viruses and respiratory droplets carrying viruses or other pathogens, represent bioaerosols, whose particle size distributions contain increasingly larger proportion of fine and ultrafine particles, as the evaporation process proceeds. We have generated two model aerosols: the nebulized NaCl solution, resembling human saliva, and the cigarette smoke, having relatively low water content. We have conducted real life experiments of such surrogate aerosol particle deposition without ionization, using bipolar ionization, as well as using unipolar negative air ions. Particle number concentrations have been measured in the 10 nm–10 μm particle size range. The calculated deposition rates and aerosol particle half-life times were correlated with bioaerosol pathogens based on the core pathogen sizes. Bipolar ionizers emitting equal concentrations of positive and negative ions had low impact to the particle concentration decrease. Intense negative air ionization resulted in pronounced deposition rate increases, particularly in the particle size range of viruses including the SARS-CoV-2. The impact of negative air ionization was most pronounced in the same size range where the deposition rates without ionization were the lowest. Therefore, the results are very promising from the standpoint of air purification and bioaerosol pathogen removal, bearing in mind that the effect of ions will be most pronounced if the unipolar ion rich air stream is directed towards the breathing zone.

The heterogeneous nature of the size and shape of particulate matter (PM) deposited onto forest canopies is acknowledged. However, it is uncertain how PM interacts with the forest canopy and how it is transported and cycled via the hydrological processes of throughfall and stemflow. To improve our understanding of particulate cycling in forested watersheds, this study quantifies the geometric configuration of PM in bulk precipitation, throughfall, stemflow, and upper soil (Oa horizon) solution in both leafed and leafless periods in a European beech (Fagus sylvatica L.) forest in Germany. Circular equivalence diameter, circularity, elongation ratio, perimeter-to-area ratio, and fractal dimension were calculated for all 43,278 individual particulates in bulk precipitation, throughfall, stemflow, and Oa solutions. Loss on ignition measurements were also conducted to determine the organic matter content of the particulates. From a physical point of view, the opposite trends for circular equivalence diameter and perimeter-to-area ratio of PM between stemflow or throughfall and bulk precipitation during both leafless and leafed periods were the most striking. For bulk precipitation, the PM's mean circular equivalence diameter was significantly larger in the leafless period than the leafed period, with the reverse observed for throughfall and stemflow. Mean perimeter-to-area ratios (μm−1) of PM of both stemflow and throughfall were significantly larger in the leafless period than the leafed period. The opposite trend was observed for bulk precipitation and Oa solution. The percent organic matter of PM was not statistically significantly different across solutions or canopy state. Our results indicate that the differential routing of PM through the canopy indeed influences the geometry of PM among solution types as compared to the bulk precipitation. The effects of these changes on the chemistry of the PM and its impact of particulate cycling, and the impacts of shifting seasonality with climate change, warrants further investigation.

This paper presents an automatic particle detection algorithm to study the time-resolved resuspension of isolated microparticle mono-layers exposed to airflow acceleration followed by steady-state. The algorithm post-processes movies of the deposit behaviour and returns the particle number, the granulometry, and the homogeneity of each frame. It allows the detection and isolation of the particle clusters to process them separately. The algorithm is validated using both synthetic images, and experimental datasets corresponding to ventilated duct cases. The number of particles remaining on the surface over time is returned, and correlations can be made with instantaneous physical parameters of the flow (e.g., centre or friction velocity). The algorithms and data are available online: see Cazes et al. (2023); CAZES et al. (2023), respectively.

High levels of particulate matter (PM) are observed in indoor environments and are of great concern to human health. INCA-Indoor is an indoor air quality (IAQ) model that is able to simulate more than 1200 gaseous species in several rooms of a building. The present study details the aerosol module of this model, which is implemented to simulate aerosol formation by the nucleation of gaseous species, the growth of particles by coagulation and condensation, their deposition on surfaces, and their exchanges with the outdoors and between rooms. To assess the performance of the new modeling system, the simulations are compared with measurements from atmosphere simulation chambers obtained from the EUROCHAMP-2020 database. Two different types of processes are studied: the growth of diesel soot in the AIDA chamber and secondary organic aerosol (SOA) formation and evolution following the ozonolysis of α-pinene in the EUPHORE chamber. INCA-Indoor nicely reproduces the evolution of the number of diesel soot particles, their size distribution during growth by coagulation, and their deposition surfaces. The model also nicely reproduces the formation of aerosols from the ozonolysis of α-pinene, the competitive growth processes, condensation and coagulation, and aerosol deposition. The simulations confirm previous experimental findings and associated assumptions regarding aerosol formation by ozonolysis.

Nuclear safety is the lifeline for the development and application of nuclear energy. In severe nuclear power plant accidents, aerosols are the major carriers of fission products, which may leak into the environment and cause potential radioactive contamination. Containment spray is an important mitigation measure for severe accidents in nuclear power plants. It can effectively reduce containment atmospheric pressure and radioactive aerosol concentration. This paper improves the containment spray removal models so as to address the low accuracy of the original models of the ISAA code. Based on single droplet collection particle mechanism, new inertial impaction and interception models were used to accurately calculate the collection efficiency of large particles. Additionally, a new correlation was used to explain the Brownian diffusion collection mechanism of small particles. Moreover, thermophoresis and diffusiophoresis models were introduced to consider the contribution of steam condensation in the containment to the collection efficiency. A rear capture model was introduced to incorporate the influence of recirculation within the wake of large droplets on the collection efficiency. Furthermore, THAI, TOSQAN and CSE experiments were selected to validate and evaluate the improved ISAA code. According to the calculation results, the improved models can simulate the attenuation trend of suspended aerosols with higher accuracies and significantly improves the calculation accuracy of the spray removal constants. This work contributes to understand of the scavenging mechanism of aerosol particle by containment spray droplets, and provides an analysis method with higher simulation accuracy. At the same time, it points out shortcomings in the simulation of aerosol behavior in the current ISAA code and discusses future improvements to the code.

Propagation of respiratory particles, potentially containing viable viruses, plays a significant role in the transmission of respiratory diseases (e.g., COVID-19) from infected people. Particles are produced in the upper respiratory system and exit the mouth during expiratory events such as sneezing, coughing, talking, and singing. The importance of considering speaking and singing as vectors of particle transmission has been recognized by researchers. Recently, in a companion paper, dynamics of expiratory flow during fricative utterances were explored, and significant variations of airflow jet trajectories were reported. This study focuses on respiratory particle propagation during fricative productions and the effect of airflow variations on particle transport and dispersion as a function of particle size. The commercial ANSYS-Fluent computational fluid dynamics (CFD) software was employed to quantify the fluid flow and particle dispersion from a two-dimensional mouth model of sustained fricative [f] utterance as well as a horizontal jet flow model. The fluid velocity field and particle distributions estimated from the mouth model were compared with those of the horizontal jet flow model. The significant effects of the airflow jet trajectory variations on the pattern of particle transport and dispersion during fricative utterances were studied. Distinct differences between the estimations of the horizontal jet model for particle propagation with those of the mouth model were observed. The importance of considering the vocal tract geometry and the failure of a horizontal jet model to properly estimate the expiratory airflow and respiratory particle propagation during the production of fricative utterances were emphasized.

Flame synthesis of Carbon NanoParticle (CNP) films is gaining strong interest for novel industrial applications because of the easy tuning of the operating conditions that enables accurate control of the chemical and physical properties of the produced CNPs. This work proposes a novelty in the synthesis of CNP films, namely the possibility of applying electric fields in flame to trigger electrophoretic deposition phenomena. In this way, it is possible to add another degree of freedom to the harvesting process and potentially modify the properties of the CNP films, without changing the operating flame conditions. To investigate the physical mechanisms governing the thermo-electrophoretic deposition of CNPs, a numerical model to simulate the particle dynamics close to the collecting substrate has been developed, and experiments have been carried out to provide highly controlled test conditions that can be used to support model validation. The experimental results consist of Atomic Force Microscopy (AFM) measurements to determine the number of particles deposited after a controlled harvesting condition as a function of the applied electrophoretic force, imposed on the substrate by means of a DC voltage varied from 0 to -3kV. The AFM shows that the amount of deposited material increases up to six times when passing from an uncharged to the -3kV charged case. The model predictions are highly consistent with the AFM measurements and pointed out that the electric field in flame significantly alters the CNP deposition velocities and impact angles, which are likely to affect the properties of the film.

The prediction of nanoparticle agglomeration using population balance modelling is a difficult endeavour as continuously growing clusters undergo different collision regimes and exhibit complex, fractal-like structures. A key model parameter in this context is the collision kernel which governs the agglomeration dynamics within the particle population. In this work, we exploit a recently developed model for Brownian coagulation to solve the population balance equation for an initially monodisperse system of spherical nanoparticles in a quiescent gas. Since the model was originally developed for the limiting cases of ballistic and diffusive agglomeration, we incorporate a suitable interpolation scheme to extend the kernel expression to the more general transition regime.The results obtained from population balance calculations are validated by means of detailed Langevin Dynamics simulations where agglomerates and their individual trajectories are resolved. A comparison of agglomeration dynamics and evolving size distributions reveals strong agreement between the two approaches which corroborates the accuracy of the collision kernel model. In contrast, conventional models for the kernel are not able to reproduce the results from detailed Langevin Dynamics simulations as they noticeably underpredict the width of the agglomerate size distribution.Analysis of the agglomerates’ morphology distribution further reveals that fractal dimensions of clusters are normally distributed and can be characterized by a size-dependent mean value and standard deviation. We find, however, that the specifics of the morphology distribution do not need to be included in the kernel as omission causes only minor changes with respect to the collision dynamics. Using a size-dependent averaged fractal dimension is sufficient within PBE calculations.

Engineered, condensational growth of aerosol particles is a well-established technique incorporated into particle detection and sampling systems. It is typically accomplished through passage of particles through controlled temperature, supersaturated systems with moderate-to-high vapor pressure compounds, which are liquid at room temperature. However, there are instances where it is advantageous to intentionally grow particles with room temperature, low vapor pressure (500 nm was desired. Homogeneously nucleated particles were broadly distributed in size, and generally smaller than the test particles grown by condensation, but still with diameters in excess of 100 nm. While SEM images of grown particles revealed smooth surfaces not indicative of agglomeration with homogeneously nucleated particles, we cannot rule out the contribution to growth of homogeneously nucleated particle coagulation with sampled particles in the growth system. By comparing to numerical simulations, we use TDMA measurements to estimate effective vapor pressures for the tested sublimated solids, assuming growth occurred solely by vapor condensation. Simulations suggest that saturation ratios in excess of 102 are needed for nanoparticle growth.

Inhaled aerosols have wide-ranging applications in the treatment and diagnosis of respiratory diseases. Methacholine challenge testing (MCT) is a diagnostic test frequently used to evaluate airway hyper-reactivity. We hypothesize that a significant fraction of the inhaled dose of methacholine is exhaled during treatment, given the small droplet sizes produced by nebulizers traditionally used for MCT. Here, an in vitro – in silico approach was developed to predict respiratory tract deposition achieved with various nebulizers proposed for use with MCT.Emitted doses, particle sizes, and the temperature and humidity of the nebulized airstream were experimentally measured for three nebulizers (RX160, Roxon Meditech; Hudson RCI Micro Mist [HRCI], Teleflex; AirLife Misty Max 10 [MM10], CareFusion) at methacholine chloride concentrations of 0.0625, 1, and 16 mg/mL in 0.9% saline solutions. Emitted doses at a concentration of 1 mg/mL were measured to be 42.0 (SD 5.1) μg from the RX160 over 120 s, 96.3 (SD 33.7) μg from the HRCI over 60 s, and 162.3 (SD 38.4) μg from the MM10 over 60 s. For a typical adult tidal inhalation, the inhaled dose was found to be half of the emitted dose. Our hygroscopic lung deposition model predicted considerable condensational growth within the respiratory tract for aerosols used with MCT. For the 1 mg/mL methacholine chloride solution, the fraction of the inhaled dose predicted to deposit in the lungs was 0.40 (8.3 μg) for the RX160, 0.62 (29.6 μg) for the HRCI, and 0.60 (48.7 μg) for the MM10. Predicted exhaled dose fractions were greatest with the RX160 (0.60) and similar for the HCRI (0.34) and MM10 (0.36). Hygroscopic modeling thus suggests that the exhaled dose obtained during methacholine challenge testing is considerable, with the characteristics of the nebulizer influencing the relative proportion of the dose that is exhaled.

Traffic remains a major source of urban air pollution although emission regulations have led to significant reductions in exhaust emissions of new vehicles. In this work the photochemical transformation of exhaust emissions from gasoline and diesel passenger vehicles compliant with the current European ‘Euro6’ emission standard and operated with traditional and alternative fuels was investigated using an environmental chamber. By assessing four different engine operation conditions, we show that vehicle operation notably affects the exhaust composition and secondary aerosol formation potential. For the gasoline vehicle, secondary aerosols dominate the total particulate emissions. In contrast, we observe no substantial secondary aerosol formation for exhaust emissions of a Euro6-level diesel vehicle. High engine load operation and cold start of the gasoline vehicle led to 11–470-fold particulate mass enhancement, while for moderate driving conditions the enhancement ratio was below 2. High aerosol enhancements also led to strong increases in particle light absorption. The results underline the necessity for future directives to include the emission components leading to secondary pollution, in addition to the freshly emitted pollutants. The link observed between secondary organic aerosol (SOA) formation and gaseous aromatic hydrocarbon emissions suggests that monitoring and limiting these gaseous species can provide an indirect regulation for SOA. Additionally, ammonia released as a byproduct of the gasoline vehicle is confirmed as an important precursor for secondary aerosols.

To resolve the potential evaporation loss of about 14% NH4+ in the particle into liquid sampler (PILS), a novel two-stage semi-dry electrostatic precipitator (SDEP) was developed successfully for monitoring atmospheric PM2.5 water-soluble inorganic ions (WSIIs) with the high collection efficiency and good accuracy. The number-based collection efficiency was >95% for particle sizes ranging from 14 nm to 11 μm, and the mass-based collection efficiency for PM2.5 was also as high as 94.4–98.1%. With an efficient pulse water jet system, the SDEP had a high ion recovery rate of 93 ± 10% for Na+, 103 ± 15% for NH4+, 92 ± 5.6% for SO42−, and 96 ± 8% for NO3−, respectively. The SDEP was coupled with the parallel plate wet denuder (PPWD) and ion chromatography (IC) as the PPWD-SDEP-IC semi-automatic monitoring system for the hourly precursor gases and PM2.5 WSII concentrations. The field comparison test for three-month measurements showed that the daily average results of the PPWD-SDEP-IC had a good agreement with those of the porous-metal denuder sampler (PDS), showing the linear regression slopes of 0.99–1.05 and 0.98–1.05, respectively, for precursor gases (NH3, HONO, SO2 and HNO3) and PM2.5 WSIIs (Na+, NH4+, K+, F−, Cl−, NO3− and SO42−) with high R2 values of 0.92–0.99 and 0.96–0.99, respectively. Moreover, the results of the hourly concentration comparison between the PPWD-SDEP-IC and the PPWD-PILS-IC showed linear regression slopes of 0.90–0.99 and 0.84–0.97 for precursor gases and WSIIs with high R2 values of 0.93–0.96 and 0.90–0.96, respectively. The PILS showed an underestimation of NH4+ concentration by about 15.6 ± 6.3% when compared to NH4+ from the SDEP owing to the volatilization loss of NH4+ caused by the mixing of aerosols with high-temperature steam in the PILS. Thus, the current PPWD-SDEP-IC system can be a viable tool for accurate semi-continuous monitoring of precursor gases and PM2.5 WSII concentrations.

In the absence of experimental air filtration test data, it is difficult to predict the air filtration efficiency and airflow resistance of filter media. In this novel work, an effort is undertaken to predict the air filtration efficiency and airflow resistance of filter media based solely on scanning electron microscopy (SEM) images of the filter fabric. The analytical single fiber efficiency (SFE) model is incorporated into a machine-learning model designed to analyze digital imagery. A convolutional neural network (CNN) is developed and optimized using synthetic training data produced from digital replication of air filter media. Simulated air filter media were created to replicate the physical characteristics of actual nonwoven nanofibrous air filter media. Digital grayscale images of the simulated media provided the input data for the CNN. Analytical calculations of efficiency and resistance based on the SFE model provided target data for the CNN. The regression model included seven convolution layers and two hidden fully connected layers along with width reduction and depth expansion methods. Twelve model parameters were optimized using training and validation data of 2100 iterations of simulated media. The model was then employed to predict the air filtration behavior of actual air filter media based on an SEM image of the media. The resulting predictions of air filtration efficiency and airflow resistance based solely on an SEM image suggests the potential viability of using synthetic data derived from simulated media to train machine learning models for real-world applications. To the knowledge of the authors, this is the first such application of a CNN to predict air filtration efficiency and airflow resistance using an SEM image, synthetic data created from simulated media, and the SFE analytical model.