CFD-DEM Simulation of the Deposition of Pharmaceutical Aerosols in Human Airways

Recent advances in the characterization of the human respiratory system and in multi-phase flow dynamics in complex geometries have led numerical simulations to play an expanding role for exploring aerosol deposition mechanisms in the lungs. However, the development of an efficient numerical and mathematical description is far from unique, and determining which aspects of the modeling are critical and which details are essentially irrelevant is indeed a difficult task. With the aim of addressing this lack of a rationalized framework, we propose: 1. A systematic analysis of pharmaceutical aerosols deposition in the extra-thoracic airways, focusing on several important modeling aspects whose related assumptions and approximations have not always been sufficiently discussed and clarified. We consider the importance of intrinsic time dependent fluctuations of the air flow, highlighting how their contribution in aerosol deposition is as important as the particle-turbulence interaction one. We show how sensitive the turbulence intensity can be to the meshing strategy and how aerosol deposition can be influenced by the latter choice. We demonstrate how a swirling air jet can enhance extra-thoracic deposition compared to a straight one, and how different the deposition patterns can be in case a realistic inhalation profile and aerosol plume are employed. 2. A critical analysis of some crucial computational aspects of aerosols deposition, we address the issues of velocity fluctuations propagation in the upper intra-thoracic airways and of the persistence of secondary flows using the SimInhale reference benchmark. We complement the investigation by describing how methodologies used to drive the flow through a truncated lung model may affect numerical results and how small discrepancies are observed in velocity profiles when comparing simulations based on different meshing strategies. 3. Comparison to experimental and numerical data available in the literature to study and quantify the impact of modeling parameters and numerical assumassumptions. Even if total deposition compares very well with reference data, it is clear from the present work how local deposition results can depend significantly upon spatial discretization and boundary conditions adopted to represent the respiratory act. The modeling of turbulent fluctuations in the airflow is also found to impact local deposition and, to a minor extent, the flow characteristics at the inlet of the computational domain. Using the CFD-DEM model, it was also possible to calculate the airflow and particles splitting at bifurcations, which were found to depart from the assumption of being equally distributed among branches adopted by some of the simplified deposition models. The results thus suggest the need for further studies towards improving the quantitative prediction of aerosol transport and deposition in the human airways.