Assessment of turbulent particle dispersion in the respiratory airway using a hybrid RANS-LES model

Turbulent particle dispersion in wall-bounded flows plays a critical role in a range of industrial and biomedical applications, including respiratory drug delivery and inhalation toxicology. This study investigates the performance of a hybrid Stress-Blended Eddy Simulation (SBES) turbulence model coupled with the Discrete Random Walk (DRW) particle dispersion model to predict particle transport and deposition in simple cylindrical and complex anatomical geometries. Simulations were conducted using CFD for three configurations: a vertical tube, a 90° pipe bend, and a realistic human nasal airway, under fully turbulent conditions (Re = 10,000) and two inhalation flow rates (30 and 60 L/min). The SBES model provided improved resolution of turbulent structures and consistent agreement with experimental correlations for particle deposition in all geometries. In the vertical tube where particle deposition is driven by turbulence dispersion, the k - ω SST model coupled with the Low-Reynolds Number (LRN) correction provided the most accurate results. However, in the 90 ° pipe bend and nasal cavity, where deposition was dominated by inertial impaction, sensitivity to the LRN correction was reduced. For nasal airflow simulations, the SBES model with or without the LRN correction produced accurate deposition trends across all particle sizes and flow rates. Results underscore that model selection should be tailored to the dominant deposition mechanism, turbulence-driven or inertia-driven. Whilst the SBES model with the LRN correction offered the most consistent performance in resolving turbulent dispersion for all three cases, the standard k - ω SST model with LRN correction was more accurate for turbulence driven deposition. Future work should evaluate additional dispersion models such as the Continuous Random Walk (CRW) to further improve physical realism in particle transport simulations. This study reinforces the value of the SBES model in simulating complex, aerosol behaviour in respiratory applications. • A hybrid SBES (Stress-Blended Eddy Simulation) turbulence model was assessed for turbulent particle dispersion in respiratory airways. • Simulations were conducted in a vertical tube, a 90° pipe bend, and a realistic nasal airway at turbulent conditions. • The SBES model accurately resolved turbulence structures, providing good agreement with experimental deposition data. • Deposition driven by turbulence dispersion was best captured by the k- ω SST model with Low-Reynolds Number (LRN) correction. • Inertial impaction-dominated cases showed reduced sensitivity to the LRN correction, especially in nasal airway geometries. • Optimal mesh configurations and particle tracking parameters were identified for respiratory particle deposition simulations. • The study highlights model selection importance based on the dominant particle deposition mechanism.