Airflow and Particle Transport Through Human Airways: A Systematic Review

This paper describes review of the relevant literature about two phase analysis of air and particle flow through human airways. An emphasis of the review is placed on elaborating the steps involved in two phase analysis, which are Geometric modelling methods and Mathematical models. The first two parts describes various approaches that are followed for constructing an Airway model upon which analysis are conducted. Broad two categories of geometric modelling viz. Simplified modelling and Accurate modelling using medical scans are discussed briefly. Ease and limitations of simplified models, then examples of CT based models are discussed. In later part of the review different mathematical models implemented by researchers for analysis are briefed. Mathematical models used for Air and Particle phases are elaborated separately.


1.INTRODUCTION
Respiration is one of the vital processes executed by almost all living beings. On an average, a healthy human being breaths about 20 times in a minute. The respiration process starts from Nasal passages from where air is taken in and passing through nasal cavity it enters trachea. From trachea air enters primary bronchus then secondary bronchus followed by segmented bronchus and at the end air reaches at alveolus. Alveolus consists alveoli sacs and alveoli ducts through which oxygen enters blood capillaries and carbon di oxide is taken out from blood capillaries into alveolus. While breathing along with air many small particles present in surroundings are also inhaled, which are deposited inside lungs as air is taken in through respiratory airways. Such external particles deposition may cause harmful effects on human health. Deposition of harmful particles leads to many respiratory health problems. But particle transport through airways also has beneficial application when inhaled in controlled conditions; This technique is called as Aerosol Drug Therapy [54]. For efficient Aerosol drug transmission, it is essential to understand Aerosol flow through the complex shapes of airways. Knowledge of governing parameters affecting Aerosol deposition is essential, such that varying these parameters optimum Aerosol deposition can be achieved. Also, it is essential to understand about harmful particle deposition like smoke and pollutants damaging human airways. It may also need to explore with the two-phase system to understand the behavior of deposition.

LITERATURE REVIEW
To analyze airflow and/or particle deposition inside human airway, researchers have implemented various methods. Classification of these methods can be done as numerical methods [1,2,3,4], experimental methods [5,10], and theoretical/analytical methods [11,12]. The available literature can be differentiate mainly based on types of Geometrical model constructed and mathematical method implemented. These models are discussed further according to literature.

Geometrical Modelling
According to explored literature, Geometric models of human airways can be categorized as Approximated models and Accurate models.

Approximate Modelling
In the early studies, most of the work utilizes simplified form of geometric model due to complexity of airway geometry. In simplified geometry (e.g. fig. 1.), dimensions and cross sections are approximated for reducing computational time and efforts [6,7]. This simplification eases the analysis but accuracy of results is compromised. Many of current research works [13,14] are based on such simplified models. These models are advantageous when global deposition values are to be determined but lacks in precision. Bora Sul et al. [45] compares airflow characteristics in normal and obstructive airways with the help of simplified models. These models are developed in AutoCad12. To investigate human tidal breathing through airways Azarnoosh et al. [49] constructed a CAD model using SolidWorks modelling package. For studying Deposition Fraction of Aerosol Particles in a Human Oral Airway approximated oral airway model is used [46]. Airflow structures and Nano-particle deposition [47] and Comparison of micro-and Nano-size particle depositions [44] are also studied based on such approximated models. For analyzing flow structure and particle deposition in asthmatic human airway [48], asthmatic airway is constructed by simply decreasing diameters of normal airways. For further simplification, some researchers assumed airways to be symmetric [22,26]. Whereas there some researchers considered asymmetric models [1,9,28,29].

Accurate Modelling
For better analysis, it is essential to use more accurate models. Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) scans [8,9,16] are helpful for accurate airway modelling. CT, HRCT and MRI are medical scan images which are imported into image processing software such as Mimics, 3d Slice, 3-D Doctor etc. These software processes images based on their grey values to construct a model for visualization and/or analysis. Human airways have complicated cross sections which are accurately modelled with such medical images [18,19]. For studying flow analyses in lower airways Backer et al. [50] implements patient specific model and boundary conditions. CT scan data of 73-year-old female patient suffering from chronic obstructive pulmonary disease (COPD) is processed using Mimics 10.0 to construct a 3D model. Thoracic CT scans of a 60-year-old male patient are processed using Mimics to model bifurcating flow in a human lung airway [51]. CFD simulation of airflow behavior and particle transport and deposition through the realistic model of human airways in different breathing conditions is conducted on a cad model developed from DICOM images using 3-D Doctor Cad package [52]. These medical scan images also can be used for Rapid prototyping. This can be done by generating. STL file from medical images using Imaging software mentioned above. Kim et al. [17] uses such method to study the airflow with PIV method. Cheng et al. [5] produced a human airway cast till three generations of bronchi to study the particle deposition in the oral airways. Jayaraju et al. [8] built a geometrical model of human upper airway with the help of DICOM images to study turbulent modeling methods [36] and the airflow fields in airways with the tracheal stenosis [20].For determining effect of geometrical modelling on the airflow field and particle transport, four different geometrical modes of upper airway are adopted by Xi et al. [16].The W. H. Finlay et al [25] carried out different geometric models for Child and adult patients with asthma or other pulmonary diseases may have different features. Rapid prototyped model developed from CT scan is used for experimental study to of turbulent particle transport in human airways under steady and cyclic flows [53]. Literature also revels analysis of bronchial airways at various generation levels. Flow field in the first generation largely influence the subsequent flow in the subsequent bifurcations [40]. In general, secondary flow still exists in the generations from G10 -G13 [41]. Nowak et al. [30] and Cebral et al. [37] uses CT scans to model bronchial airway from generation 0 to generation 4 (G0 -G4). Van Ertbruggen et al. [31] created a geometry model for seven generations based on the morphometrical data cited in Horsfield et al. [37]. Seventeen generations of human respiratory tract are abstracted topological graphical data from anatomical model defined in Schmidt et al. [38], was adopted by Gemci et al. [32] to study the airflow in the human respiratory system. It includes seventeen generations of the total twenty-six generations of the human tracheobronchial airways with 1453 bronchi [32].

Air Phase
The major difficulty in analysis of Gas phase is that due to geometrical complications and variations of Airways the flow changes from laminar to turbulent. Method should be efficient to formulate not only laminar flows, but also transitional and turbulent flow structures [19,23,36]. The complexity of the extra-thoracic airway includes the bends, sudden cross-sectional area change [18,19], this creates turbulence in the flow. Currently, there are many methods available for analyzing the laminar-transitional-turbulent flow in the human respiratory system for example direct numerical simulation (DNS), large eddy simulation (LES) [9,36], Reynolds-averaged equations [8,7,33], detached eddy simulation (DES) [36]. K-model is suitable for simulating the flow with small pressure gradient, while it cannot accurately predict the flow containing large adverse pressure gradient [37]. The K-model is another kind of the most commonly used two-equation turbulence models including the transport equations of kinetic turbulent energy and specific turbulent dissipation [38]. Some of the work [2,7,8,19,33] implements the RANS(Reynolds-averaged Navier-Stokes equation) method to simulate the flow field. Large eddy simulation (LES) predicts particle deposition. LES also predicts the large-scale flow structures by solving the filtered Navier-Stokes equations and modeling unresolved vortices with subgrid scale model. Jayaraju et al. [36] used RANS k -, detached eddy simulation (DES), and LES methods for comparing accuracy of these models corresponding to deposition inhaled aerosol medication. Until now, the governing equations have been solved with commercial software such as KIVA [39], FLUENT [52] and CFX [2]. Sometimes, it is accompanied with the user-defined program in C, Fortran or Openfoam.

Particle Phase
At present, there are basically two different approaches in the analysis of the phenomenon of particles dispersed in the airflow in the respiratory system such as Euler-Lagrange method [15] and Euler-Euler method [21]. In the Euler-Lagrange method, a particle trajectory is calculated by solving equations of the motion for each particle using Lagrangian approach [8,14,19,21,33]. On the other hand, in the Euler-Euler approach, a particle concentration distribution of the carrier fluid is calculated [34]. The parcel method assumes that there are more particles in one parcel with the same physical properties such as diameter and velocity [35]. If there are more particles, the volume fraction will become large, then it may be necessary to consider the influence of particle momentum on the airflow field, which means that two-way coupling should be adopted [35]. The inertial impaction of particle is decided by the air. Thus, the larger of IP means the larger of particle inertial impaction [15]. Gravitational settling is a function of particle size, particle density and time, with the rate of settling proportional to particle size and particle density micro-particle transport and deposition have been extensively studied by a lot of researchers in the oral airways [5, 4, 7, 12, 19,]. Other than total and regional particle deposition, ICMAEM-2017 IOP Publishing IOP Conf. Series: Materials Science and Engineering 225 (2017) 012132 doi:10.1088/1757-899X/225/1/012132 local particle deposition pattern is another important parameter for the assessment of particle deposition influence on the health [4]. The different particle trajectories have been obtained and discussed to understand the effects of airflow and Stokes numbers [19]. Particle deposition in the Tracheobronchial airways has been found to be contributed to the occurrence of asthma attacks [42]. Aerosol drug deposition in the Tracheobronchial region can reduce drug delivered into pulmonary region [25]. In contrast, some aerosol drug is targeted to Tracheobronchial region, such as bronchodilator and corticosteroids, to treat Tuberculosis airway asthma [43]. A series of research has been developed to study the particle transport and deposition in the lung with experimental and numerical methods from one to several bifurcations [1,2,24,30,31,34].

3.CONCLUSION
Literature revels development of accurate Airways model is feasible with CT data. For accurate analysis, deeper generations of bronchi are needed to be modelled. Two phase system intended to be analyzed for particle deposition inside airways. According to the flow conditions most applicable mathematical method is implemented in analysis.