Applying computational fluid dynamics methods on nasal flow investigations based on a real domain generated from CT data

Abstract

Modern nasal surgery raises questions on the improvement of nasal airways, providing a patient with healthy, comfortable breathing. In fact, presently available and realizable measurements are not sufficient to define standard shapes for optimized flow behavior. Airway shapes determine local flow velocities and the degree of turbulence. Transition areas between laminar and turbulent flow regions affect particle transport and consequently the sensation of smell. Numerical simulation approaches (Computational Fluid Dynamics/CFD), particularly Large Eddy Simulation (LES), might help to understand the flow behavior of such complex geometries. Present clinical scanning techniques, e.g. CT, allow digital reconstructions of a patient's nasal geometry. These can be used to construct a simulation grid spanning the air flow domain required for breath flow calculations. From this point, investigations of varying geometries seem easy to realize. Even virtual surgery rooms with the aid of CFD seem possible. Within this interdisciplinary contribution, the principal steps for setting up such simulations will be outlined and compared to the steps taken within this project. Exemplary flow calculations based on a real nasal geometry will be shown, using a parallelized, compressible Navier-Stokes code. Applicability and limitations of this approach will be outlined. Focus is laid on questions related to grid generation/dependence and data validity.