Computational analysis of human upper airway aerodynamics

Abstract

There is a considerable interest in understanding transient human upper airway aerodynamics, especially in view of assessing the effects of various ventilation therapies. Experimental analyses in a patient-specific manner pose challenges as the upper airway consists of a narrow confined region with complex anatomy. Pressure measurements are feasible, but, for example, PIV experiments require special measures to accommodate for the light refraction by the model. Computational fluid dynamics can bridge the gap between limited experimental data and detailed flow features. This work aims to validate the use of combined lattice Boltzmann method and a large eddy scale model for simulating respiration, and to identify clinical features of the flow and show the clinical potential of the method. Airflow was computationally analyzed during a realistic, transient, breathing profile in an upper airway geometry ranging from nose to trachea, and the resulting pressure calculations were compared against in vitro experiments. Simulations were conducted on meshes containing about 1 billion cells to ensure accuracy and to capture intrinsic flow features. Airway pressures obtained from simulations and in vitro experiments are in good agreement both during inhalation and exhalation. High velocity pharyngeal and laryngeal jets and recirculation in the region of the olfactory cleft are observed.

Computational Rhinology: Unraveling Discrepancies between In Silico and In Vivo Nasal Airflow Assessments for Enhanced Clinical Decision Support

“…Few publications have investigated the effects of imposing unsteady flow boundary conditions, such as respiratory tidal flow, or applied these turbulence models to simulate nasal airflow. In addition to the present study, recent investigations utilizing finely resolved LES models with time-varying flow have been conducted by Lu et al [ 248 ], Calmet et al [ 249 , 250 ], Bradshaw et al [ 159 ], and Hebbink et al [ 127 ]. Other publications have also utilized high-fidelity CFD models under steady flow conditions; refer to Table 1 for an overview.…”

“…Although a few studies report a good match between in vivo and in silico RMM, the general impression is that CFD-based models struggle to reproduce in vivo RMM pressure–flow curves [ 115 ]. It appears that in silico studies agree better with in vitro studies in rigid nasal cavity replicas, however [ 117 , 127 ].…”

“…The alignment reported between in silico RMM and in vitro RMM in physical replicas of nasal cavities [ 127 , 129 , 130 ] starkly contrasts the observed disagreement with in vivo RMM [ 117 , 125 ]. This suggests that CFD models are correctly configured, but somehow fail to adequately represent the nasal cavity’s actual state and function during in vivo RMM examination.…”

Computational Rhinology: Unraveling Discrepancies between In Silico and In Vivo Nasal Airflow Assessments for Enhanced Clinical Decision Support

“…Although a few studies report a good match between in vivo and in silico RMM, the general impression is that CFD-based models struggle to reproduce in vivo RMM pressureflow curves [115]. It appears that in silico studies agree better with in vitro studies in rigid nasal cavity replicas, however [117,127].…”

Computational Fluid Dynamics and Its Potential Applications for the ENT Clinician

“…These studies have rigorously assessed and verified the accuracy and reliability of CFD in simulating nasal sinus airflow and irrigations, providing valuable insights into its applicability in this domain. 8,[20][21][22][23][24][25][26][27][28][29][30][31][32][33] Li et al conducted a comprehensive study comparing various numerical models used in CFD analysis with experimental measurements obtained from a physical model. 9 The evaluated models included unsteady large eddy simulations (LES) and steady Reynoldsaveraged Navier-Stokes (RANS) simulations employing different turbulence models, such as the κ-ε model, standard κ-ω model, and κ-ω shear stress transport (SST) model.…”