Numerical simulations of aerosol delivery to the human lung with an idealized laryngeal model, image-based airway model, and automatic meshing algorithm. (22nd April 2017)

• Record Type: Journal Article
• Authors: Miyawaki, Shinjiro; Hoffman, Eric A.; Lin, Ching-Long
• Is Part Of: Computers & fluids. Volume 149(2017)
• Page Start: 1
• Page End: 9
• Publication Date: 2017-04-22
• Subjects: Centerline-based model -- Laryngeal model -- Non-uniform mesh -- Branch-by-branch analysis -- Multiple subjects; Fluid dynamics -- Data processing -- Periodicals; 532.050285
• DOI: 10.1016/j.compfluid.2017.02.008
• Languages: English
• ISSNs: 0045-7930

Abstract

Highlights: We developed an idealized laryngeal model to replace missing extra-thoracic images. The new method automatically meshed human airway geometries. The new method accurately predicted the aerosol depositions in the central airways. The laryngeal model generated a realistic level of turbulence in the trachea. The new method is suited for analysis of aerosol distribution in multiple subjects. Abstract: The authors proposed a new method to automatically mesh computed tomography (CT)-based three-dimensional human airway geometry for computational fluid dynamics (CFD)-based simulations of pulmonary gas-flow and aerosol delivery. Traditional methods to construct and mesh realistic geometry were time-consuming, because they were done manually using image-processing and mesh-generating programs. Furthermore, most of CT thoracic image data sets do not include the upper airway structures. To overcome these issues, the proposed method consists of CFD grid-size distribution, an automatic meshing algorithm, and the addition of a laryngeal model along with turbulent velocity inflow boundary condition attached to the proximal end of the trachea. The method is based on our previously developed geometric model with irregular centerlines and cross-sections fitted to CT segmented airway surfaces, dubbed the "fitted-surface model." The new method utilizes anatomical information obtained from the one-dimensional tree, e.g., skeleton connectivity and branch diameters, to efficiently generate optimal CFD mesh, automatically impose boundary conditions, and systematically reduce simulation results. The aerosol deposition predicted by the proposed method agreed well with the prediction by a traditional CT-based model, and the laryngeal model generated a realistic level of turbulence in the trachea. Furthermore, the computational time was reduced by factor of two without losing accuracy by using the proposed grid-size distribution. The new method is well suited for branch-by-branch analyses of gas-flow and aerosol distribution in multiple subjects due to embedded anatomical information.