Application of High Performance Computational Fluid Dynamics to Nose Flow

Modern nasal surgery aims at improving airways for healthy, comfortable breathing. Presently available measurements are not sufficient to describe optimized shapes, particle transport and flow. Computational Fluid Dynamics (CFD), particularly Large Eddy simulation (LES), might help to understand flow details and define standards. However, the human nose flow is challenging state-of-the-art CFD methods. It is generally not as clear and as well investigated as technical flows where CFD has originally been developed for. The challenging aspects are: First, the geometrical complexity of the nasal airways is much higher. Thin, long channels exist with multiple junctions and separations, 3-dimensionally contorted, requiring high resolution. Second, flow conditions are an interaction of several physical phenomena and tend to challenge numerical schemes in terms of viscosity, Reynolds number, Mach number, wall roughness, turbulence, heat transfer, humidity, fluid-tissue interaction etc. Third, only few validable experimental data of sufficient accuracy are available. Dealing with humans, either no standardized measurement conditions exist due to the bodies’ uniqueness or standard measurement procedures of engineering type cannot be applied. This causes a lack of comparability, limiting conclusions for surgery. Within this contribution, exemplary flow simulations through a real nose geometry under average conditions will be shown using a MPICH-parallelized, compressible Navier–Stokes scheme. The emphasis is on investigating fast, small-scale flow fluctuations near the regio olfactoria. The intention is to present a first step and to show in which direction developments must turn in order to perform reliable simulations. A 3D compressible CFD research code will be used, which is developed at the Institute of Fluid Machinery, University of Karlsruhe, Germany.