Computational modelling of flow in idealised and realistic airway bifurcations

Understanding of the flow of air and particles in the lung is essential to the success of pulmonary drug delivery. The main objective of this work is to characterize the flow in a single geometrically realistic tracheobronchial bifurcation by computational methods. Much research to date (e.g. Comer et al [1]) is based on idealised geometry put forward by Weibel [2]. Another aim of this project, therefore, is to compare observed flows in realistic and idealised geometries in order to evaluate the validity of the simplified models. A computational model of a realistic geometry was generated using images obtained from the Visible Human data set (Banvard [3]), which comprises a three-dimensional anatomical picture of a human cadaver. Extracts from this data set were used with image processing and surface modelling software to generate a geometric description of a 4^th generation bifurcation. The geometry of the idealised model is based on the Weibel model A. The unsteady flow of air through the bifurcation was then modelled computationally. A general purpose computational fluid dynamics package, CFX5®, was used to predict the entire three-dimensional unsteady flow field inside the bifurcation. It was assumed that the airflow was unsteady, incompressible and laminar at the location of the 4^th generation airway in the lung. Computations were carried out for a Reynolds number of 510 (based on the average velocity in the parent branch of the bifurcation). The walls of the airway were assumed to be rigid. Computational results for air flow highlight the complex flow structure in the realistic geometry, and indicate major discrepancies between realistic and idealised geometry models.