Comparing LES and URANS results with a reference DNS of the transitional airflow in a patient-specific larynx geometry during exhalation

Abstract

Listen

Translate article

Air exhalation in the laryngeal area produces a complex transitional flow. A detailed investigation of the resulting flow features is very interesting from a fundamental point of view and useful to support medical researchers regarding treatment options. In the current study, the highly complex, unsteady, transitional flow in a patient-specific geometry has been described by three different numerical approaches, (1) using the unsteady Reynolds-averaged Navier–Stokes (URANS) equations with the k−ω shear stress transport (SST) turbulence model, (2) by large-eddy simulation (LES) with the WALE subgrid-scale model, and finally (3) using direct numerical simulation (DNS) – without any model needed. These three approaches deliver different pictures of the resulting flow, obviously with noticeable differences in terms of accuracy, but also regarding the associated computational efforts. The present article is devoted to the detailed comparison of these three approaches, using the DNS results as a reference. It has been found that even though URANS is able to predict with a fair accuracy the large-scale features of velocity and pressure in an average sense, it fails to predict correctly most relevant turbulent features; large differences are found concerning fluctuations, power spectral density, or spectral entropy. On the other hand, LES is able to reproduce both small-scale and large-scale flow features in a good agreement compared to DNS; additionally, LES delivers an appropriate description of the evolution of spectral entropy and can thus be used to predict flow transition. Obviously, DNS delivers an even more detailed and accurate view of the flow properties, but this comes with a tremendous increase in computing effort. For this configuration, a high-quality LES appears as best compromise between accuracy and computational requirements and, therefore, best suitable for medical studies of transitional airflows. All data-sets are publicly available under Abdelsamie et al. [1].