Session H01: Biological Fluid Dynamics: Physiological Respiratory Flows (5:45pm - 6:30pm CST)

Numerical study of oscillating non-Newtonian flows, application to a pulmonary airway clearance device.

Chronic respiratory diseases are responsible for an increasing number of deaths due to the deterioration of air quality in industrialized countries. To help patients suffering from these diseases, which in most cases produce an excess of mucus (severe asthma or cystic fibrosis), the use of external devices like Simeox can be recommended. This device produces pressure waves inside the bronchial tree to help to expectorate mucus more easily, and improving the actual comprehension of the mechanisms at play to enhance this expectoration is crucial. In this perspective, a numerical study of pulsated forcing of a 2D idealized bronchus filled with a non-Newtonian fluid has been performed. The Lattice Boltzmann method is used to solve the flow. The non-Newtonian behavior of the fluid is modeled by a Herschel-Bulkley law, taking into account two properties: shear-thinning/thickening and yield stress. A parametric study based on three types of signals, as well as on the variation of the 4 non-dimensional numbers governing the flow is carried out.   

Virus transmission: sensitivity study of the Nasal/Buccal Geometries and Saliva Properties.

The present work presents a sensitivity study of different geometries and fluid properties in the context of a sneeze to understand how these parameters drive virus transmission. This work aims to bring more comprehension of the mechanisms behind the transmission of viruses, such as the COVID-19, that was responsible for several deaths around the world. Numerical simulations, based on four geometrical conditions for nasal and buccal passages and three saliva properties, represent sneeze events using typical regimes and conditions. The numerical domain consists of a human body, with simplified geometry of throat, buccal and nasal passages. This geometry considers four different exit conditions: teeth and nasal passage open; without teeth and nasal passage open; without teeth and nasal passage closed and; teeth and nasal passage closed. Besides this, the physical properties of saliva were evaluated at three conditions: thinner saliva, standard saliva, and thicker saliva. The results showed that the spray formation characteristics are highly dependent on small geometrical changes and. In terms of fluid properties, the thicker saliva presented a reduction of at least twice in the number of droplets compared to the thinner saliva.   

Dynamics of inertial particles in the respiratory airways in the small-Stokes, small-P\'{e}clet regime

We investigate the Lagrangian transport of micron and sub-micron size particles in the human lung using a novel Maxey-Riley-Langevin model. The respiratory airways are divided into two primary zones: (i) conducting zone and (ii) acinar zone. The particle transport in the conducting zone is showed to be mainly governed by fluid convection (low Strouhal number). In the acinar zone, particle intrinsic motion (i.e.,Brownian transport) is shown to become increasingly important. The transition between a convection-dominant and diffusion-dominant motion in the Lagrangian reference frame is observed to occur at about the $17^{th}$ branching generation in the lung. The particle transport by hydrodynamics and by Brownian motion are shown to not obey superposition - an observation that bears importance in the context of disease propagation. Finally, Boussinesq-Basset history force is shown to not play a role in particle transport.   

Machine learning analysis of the exhaled human breath

Human exhaled breath is predominantly turbulent. During exhalation, air is forced to flow out of the lung through the trachea as a result of the contraction of diaphragm. The air passes through the throat and oral cavity. We attempt to validate the hypothesis that turbulent exhaled flow carries a signature of the source of generation i.e., the geometry of the upstream flow region. Based on this hypothesis, we explore the possibility whether humans can be classified based on the turbulent signatures in their exhaled breath. We employ machine learning practices to test the hypothesis. Features from the available time series data were extracted using MFDFA, a technique widely used for determining the fractal scaling properties and long-range correlations in the time series. The features are attributes of the multifractal spectrum of exhaled velocity data. Machine learning algorithms such as logistic regression, decision trees, boosting trees were used in an ensemble model for classification in our study. The accuracy in binary classification of a majority of subject combinations was above $75\%$ which signifies the existence of some uniqueness in an individual's exhaled breath. This could be used as a tool to characterize the person-to-person variation in extrathoracic morphology.   

Flow analysis of endotracheal tubes for the prevention of ventilator-associated pneumonia

Respiratory illnesses in the recent past, including COVID-19, have led to an increased use of mechanical ventilators in severely affected patients. Long-term use of these ventilators (over 14-21 days) is linked to mortality rates of up to 50\%, owing to the development of secondary infections. Ventilator-associated pneumonia (VAP) is a leading cause of these infections and is primarily caused by the pooling of mucus around the endotracheal tube (ETT) inserted in the upper trachea. Prediction and prevention of the pooling are challenging owing to the complex geometry and non-Newtonian flow interactions, thus requiring constant clinical monitoring of patients. This study aims at providing a prediction of the mucus behavior in intubated patients, via a flow analysis of an intubated upper trachea model. Large eddy simulations incorporating the respiratory cycle and mucus distribution are carried out and the effect of the ETT in mucus leakage is determined. A predictive model for the leakage is developed from this understanding of the primary flow physics in and around the ETT and its relation to patient-driven factors. The model will aid informed selection of the ETT, monitoring when needed, and thus reduced secondary infections in patients and reduced exposure for clinicians.   

Albuterol Delivery Through An Adult Ventilator Circuit To A Patient-Specific Tracheobronchial Airway Model

An integrated \textit{in vitro in silico} approach was employed to investigate albuterol administration to a patient-specific adult lung airway. The particle size distribution (PSD) of aerosolized albuterol generated by a jet nebulizer and a vibrating mesh nebulizer was evaluated by placing them at three different positions in an adult ventilator model -- dry side of the humidifier, before the Y-piece, and tested alone without the circuit. The circuit was connected to a Next Generation Impactor equipped with a standard USP induction port, through a 7-mm inner diameter ETT that is operated at a constant flow rate of 14L/min. The experimentally determined PSD served as the input to the \textit{in silico} model, which involved Reynolds-Averaged Navier Stokes simulations with Lagrangian particle tracking. Particle deposition was predicted in two patient-specific upper airway models - intubated model A (ETT-G0-G7), and a truncated model B (G0-G7). The deposition in the mouth-throat region and the effect of the laryngeal jet on the flow for model B was appropriately accounted for by comparing the deposition characteristics in the USP throat and an anatomical mouth-throat geometry. The vibrating mesh nebulizer delivered a higher dose of albuterol compared to the jet nebulizer, and the nebulizers were most efficient when placed on the dry side of the humidifier. The effect of type and location of nebulizer on deposition efficiency, regional deposition, and the fraction of drug available to the different lung lobes will be presented.   

Study of airflow and aerosol deposition characteristics within human airways using large-eddy simulations

Aerosolized drug delivery in human airways is typically used for the treatment of several pulmonary diseases. In this study, large-eddy simulation (LES) is used for the study of the airflow and the aerosol deposition characteristics within the upper human airways. The geometric model comprising of the extrathoracic and a part of the intrathoracic airways corresponds to a truncated portion of a realistic human airway model based on the SimInhale benchmark case. LES is performed using the well-established Eulerian-Lagrangian approach where the airflow is modeled as an incompressible flow using the Eulerian formulation, and the aerosol evolution is tracked in a Lagrangian manner under the dilute suspension conditions using a one-way coupled approach. The closure of the subgrid-scale stress in the Eulerian equations is performed using the locally dynamic one-equation based model, and for the subgrid dispersion in the Lagrangian equations, a stochastic model is used. The accuracy of the LES results is first assessed in terms of the prediction of the turbulence statistics and the regional/global aerosol deposition by comparing with the past results corresponding to the full geometry. Afterward, the effects of inlet Reynolds number on the flow and the aerosol deposition are analyzed.   

Fluid dynamic insights into the deposition of virus in deep lung

The impact of airborne viruses like SARS-CoV-2 in a human lung depends on the deposition and clearance of the inhaled viruses, among other factors. The viruses get deposited in the mucus layer of the lung. The mucus layer motion toward lower generations, caused by ciliary beating, clears the lung of the deposited viruses. Thus, the combined effect of the deposition and clearance mechanisms is critically important to determine infection by viruses. The deposition and clearance of airborne viruses has been individually investigated. In this work, a coupled analysis considering both mechanisms is done. We developed a coupled Weibel-like numerical model for the airways and the mucus which solves for virus-laden aerosol concentration in the airways and virus concentration in the mucus. The viruses are observed to be deposited in the deep lung for aerosol sizes less than 3 micron. Clearance rate of the deposited virus is observed to depend on the initial deposition location, mucus advection and breathing cycle. While depositions in the upper lung are cleared in a short duration, substantially longer times are required for clearance of the deep lung depositions. Weaker mucus advection in the deep lung and a longer breathing cycle is observed to substantially increase the clearance time.   

Analyzing the dominant SARS-CoV-2 transmission routes towards an ab-initio SEIR model

Different transmission routes of the SARS-CoV-2 virus, and their role in determining the evolution of the Covid-19 pandemic are analyzed, in this work. Probability of infection caused by inhaling virus-laden cough droplets (initial, ejection diameters between $0.5-750\mu m$) and the corresponding desiccated nuclei that mostly encapsulate the virions post droplet evaporation, are calculated. At typical air-conditioned yet quiescent indoor space, for average viral loading, droplets of initial diameter between $10-50 \mu m$ have the highest infection probability. However, by the time they are inhaled, the diameters reduce to about $1/6^{th}$ of their initial diameters. Combined with molecular collision theory adapted to calculate frequency of contact between the susceptible population and the droplet/nuclei cloud, infection rate constants are derived ab-initio, leading to a SEIR model applicable for any respiratory event – transmission vector combination. Viral load, minimum infectious dose, sensitivity of the virus half-life to the phase of its vector and dilution of the respiratory jet/puff by the entraining air are shown to mechanistically determine variation in the basic reproduction number $\mathcal{R}_0$, from first principle calculations. Ref: https://arxiv.org/abs/2007.13596   

Multi-plug instillation into a physiologically representative infant airway tree model - A computational study

We present 3D multiphase flow simulations of liquid surfactant plug transport through a physically representative model of the human infant lung airway tree. Liquid surfactant instillation into the lung airways is used to treat respiratory distress syndrome (RDS) in preterm infants. The procedure, commonly known as surfactant replacement therapy (SRT), is used in the targeted delivery of surfactant plugs with the goal of achieving a uniform film distribution. SRT's effectiveness is tied to the successful plug propagation through each branching airway network. We previously investigated such plug dynamics in individual airways and Y-Tube models with comparisons made between other experimental, numerical and mathematical works. By expanding the airway model to include further airway generations with rotations, our simulations can capture the dynamical effects of upstream plug propagation on downstream plug behavior across multiple length scales. The simulations investigate the effects of multi-plug instillation, plug blockages, and plug rupture on plug splitting and film distribution, with the goal of improving our understanding of SRT.   

Respiratory Droplets Transport via Vortex Dynamics during Expiration

Human-to-human transmission of upper respiratory diseases such as COVID-19 is primarily driven by the dispersion of virus-laden droplets that are expelled from the nose and mouth. Aerosolized droplets can accumulate in the air for hours and be present at sufficiently high concentrations to pose a significant health hazard especially in confined spaces. Large-scale coherent vortical structures, induced along the surfaces of the mouth and nose, play a particularly crucial role in determining the transport of aerosolized droplets. In this study, high-resolution particle imaging of a fully pulsed flow imitating a human cough reveals coherent motion and dispersion of droplets away from the source. The transient interactions of the droplets with vortical structures are detailed at different temporal phases including the initial starting pulse, the immediate trailing jet, and asymptotic behavior downstream. The coupling between the vortical structures and the droplets is investigated based on the distribution of droplet sizes and concentration. The penetration of the jet along with the lifetime or extinction of the droplets are visualized as a function of the distance from the source, droplet size, and droplet concentration.   

Simulations of the Interactions of Coherent Vortical Structures and Respiratory Droplets during Expiration

Two of the most critical questions pertaining to reducing transmission of respiratory infections and social distancing are what human-to-human separation distances are safe, and how can airborne transmission be mitigated. Coherent vortical structures induced by a cough on the surface of the mouth and nose affect the transport of respiratory droplets. Computational simulations are undertaken to capture the spatio-temporal interactions of the coherent vortical structures and the respiratory droplets that can aerosolized and become airborne. The simulations are validated with complementary experimental measurements to assess the ability of numerical models to capture the spatio-temporal evolution of the droplets and aerosolized particles convected in a fully pulsed jet, which is used to model a human cough. Comparisons of the high-resolution images and simulated droplets are discussed. Lagrangian droplets are tracked as they are expelled from a transient fully pulsed jet and interact with the incipient vortices as they disperse outward. The coupling between the droplets and the vortical structures is detailed as a function of the distribution of droplet sizes. The formation and evolution of the vortical structures are shown to transport and disperse the droplets.   

Flow Disturbances Reduce the Effectiveness of a Face-mounted Negative Pressure Antechamber for Endonasal Surgery

The COVID-19 outbreak has driven an increase in face-mask research. Despite many investigations concerning general-use face masks, few options are available to protect surgeons performing aerosol-generating procedures. A novel mask was designed based on the containment technique of biosafety cabinets using negative pressure. Cross-validation using a high-speed camera and an optical particle counter was done to find a threshold where no aerosol leakage occurred. Two masks of different opening areas were developed, and these resulted in different minimum pressure requirements. Flow disturbances impacted the ability to contain aerosols near the larger opening, leading to a higher pressure threshold. These experiments showed the mask was effective in containing aerosols, limiting the spread of diseases during surgical procedures.   

A Coupled Lagrangian Model for Flow-mediated Transmission of SARS-CoV-2 through Respiratory Ejecta in a Skilled Nursing Facility.

The SARS-CoV-2 pandemic has fundamentally altered societal structures and norms in ways that will continue for years to come. SARS-CoV-2 can be transmitted between individuals through respiratory ejecta from infected persons. Infection transmission risks are particularly high in occupied indoor spaces, especially nursing and care facilities with vulnerable older population. It is therefore important to understand the transport of respiratory particles in closed indoor spaces, to better characterize infection spread. These include aerosolized particles, surface-based particles, and the particles that potentially become re-suspended after being disturbed due to some human activity. In this work, we propose a Lagrangian model for virus-laden particle transport in indoor air flow. The model accounts for coupled transport and mass-transfer phenomena at the individual particle scale. This is combined with full-scale building air flow models to understand indoor viral particle transport. We used this model to conduct a study on designing and implementing a negative-pressure isolation space in a skilled nursing facility to control infection spread. Model results illustrate how particle transport was controlled within the isolation space.   

Airway flow generated in abnormal breathing patterns

Normal breathing is a combination of involuntary expansion and contraction of chest and diaphragm muscle moments. Normal respiratory rate (RR) is in the range of 10-20 breaths per minute. Physiological and pathological causes can lead to abnormal breathing patterns such as: tachypnea (approximately 1.5x increase in RR), bradypnea (1.5x decrease in RR), hyperpnea (deep breathing with abnormally large peak flow rate), and hypopnea (shallow breathing with abnormally low peak flow rate). To investigate the airway flow fields representative of these abnormal breathing patterns, 3D simulations were performed on an idealized airway model up to the 2$^{\mathrm{nd}}$ bifurcation, using k-$\omega $ model available in ANSYS Fluent v19. A parametric study was conducted across varying: (a) peak flow rates (during inhalation and exhalation) and (b) cycle duration. At the end of one-minute, cumulative mass flow rates at outlets followed the order of tachypnea \textgreater hyperpnea \textgreater normal \textgreater hypopnea \textgreater bradypnea. Airway flows during varying breath hold durations will also be discussed.   

Material filtration efficiency for effective respiratory protection in time of COVID-19

In the time of COVID-19, ad-hoc face-coverings and medical masks have become prevalent due to shortages or conservation of personal protective equipment (PPE) for healthcare workers. It is important to guide choice of materials and designs of face-coverings based on systematic and standardized detailed characterizations of particle penetration, flow, and breathability. In this talk, we present the results of systematic and multi-modal measurements of filtration efficiency,  breathability, in addition to flow visualization for a range of common materials and their combinations and discuss the implication for do-it-yourself face-covering or resource limited PPE designs in time of shortage.    

Fluid Dynamics of Speech and Cough Aerosols in the Context of COVID-19 Transmission

Production and dispersion of disease-carrying human aerosols from coughing and sneezing have been widely studied. However, an alternative origin of transmission (aerosol from speaking) have largely remained unexplored. The latter is especially important for COVID-19 where peak virus shedding is reached in the pre-symptomatic phase. Recent studies by Asadi et al. showed that speaking generates as much aerosol over time as coughing/sneezing, while vowels such as /i/ and consonants such as plosives and nasals have the highest production rates. Subsequent dispersion of these aerosols remained unstudied. Our investigation completes the knowledge gap by employing high-speed particle image velocimetry to quantify the dispersion of speech aerosols produced by a human subject at different intensities of vocalization. Our results indicate that phonetics with the most aerosol production did not always correlate with the furthest aerosol penetration. Certain vowel-consonant combinations such as ``ti'' (``tee'') can produce penetration comparable to coughing. Furthermore, the ejected aerosol clouds were observed to fluid dynamically resemble a puff jet, where vortex roll-up along the plume-front concentrated the ejected aerosols and prevented immediate dilution. We propose that assessment of risk burden from COVID-19 exposure during speech must take production, penetration, and vortex structure all into account.   

Pathogen clearance in the respiratory system and other self-cleaning surfaces

Our airways are continuously exposed to potentially harmful particles like dust and viruses. The first line of defence against these pathogens is a network of millions of cilia, whip-like organelles that pump flows by beating over a thousand times per minute. In this talk, I will discuss the connection between local cilia architecture and the topology of the flows they generate. We image the mouse airway from the sub-cellular (nm) to the organ scales (mm), characterising quantitatively its ciliary arrangement and the resulting flows. Interestingly, we find that disorder in the ciliary alignment can actually be beneficial for this pathogen clearance. More generally, I would also like to discuss how systems can be driven out of equilibrium by such active carpets. Combining techniques from statistical and fluid mechanics, I will demonstrate how we can derive the diffusivity of particles near an active carpet, and how we can generalise Fick's laws to describe their non-equilibrium transport. These results may be used for new self-cleaning materials, much like our airways.   

Hyper-realistic lung model for quantitative CT and CFD-based lung assessment of personalized exposure to air pollution

We propose a hyper-realistic lung model for personalized assessment of lung exposure to air pollution, based on inspiratory and expiratory computed tomography (CT) images, computational fluid dynamics (CFD) simulations, and individual exposure measurement. We collected inspiratory and expiratory chest CTs, pulmonary function test (PFT) data, and personal air pollution exposure measurements from 200 subjects, to date, with healthy, asthma, chronic obstructive pulmonary disease (COPD), and interstitial pulmonary fibrosis (IPF) lungs. We first apply exposure amount of particulate matters (PMs) to an individual, by five standard size ranges, PM1, PM2.5, PM4, PM10, and total suspended particles (TSP). Then, a PFT-based personalized tidal breathing pattern at the mouth and regional ventilation based on inspiration-expiration CT image matching with the compliance lung model provides realistic prediction of cumulative inhalation of PMs at different size ranges. Large eddy simulations (LES) and particle transport simulations resolve regional lung distribution of PMs throughout the entire conducting airways. The proposed model associates regional particle deposition hot spots with disease-specific structural and functional alteration associated with particulate air pollutions.   

Precipitation Dynamics in a Respiratory Droplet Evaporating on a Solid Surface

Expired respiratory droplets from an infected person dry and pose a threat to those in immediate vicinity. While smaller droplets desiccate in air, larger droplets will settle and evaporate on a surface creating fomites. The authors have used surrogate respiratory droplets consisting of salt in water (1% wt.), porcine mucin and DPPC surfactant, to understand the role of different surfaces such as glass, plastic, ceramic and metal on the final distribution of virus in fomites, which is reportedly different from airborne precipitates. Particles of same size as most well-known virions are loaded in an initial concentration of ˜109 particles/ml. Fluorescent labelling of particles visualizes their relative distribution within the surface precipitation. Crystals are precipitated at the end of the droplet evaporation; pre-dominantly large isolated crystals are observed on metals while dendrites, of varying sizes based on the substrate, are observed in other cases. The particles appear preferentially located in the dendritic formations which could be due to their faster evolution rates as compared to large crystals. This is linked to the flow and evaporation in sessile droplets.