Towards an accurate CFD prediction of airflow and dispersion through face mask

Given the difficulty of experimental measurement of respiratory airflow and dispersion through a face mask, accurate numerical simulation is an important method to increase the understanding of the health effect of face masks and to develop high-performance ones. The objective of this study is to develop such an accurate modeling framework based on computational fluid dynamics (CFD) theory and method. For model validation, the flow characteristics through the face mask were tested experimentally, and the air speed and exhaled pollutant concentration in the breathing zone were measured with human subjects. The influence of gird division, time step size, and turbulence model on simulation accuracy were investigated. The result shows that the viscous resistance coefficient and inertial resistance coefficient of face masks (surgical masks) were 3.65 × 10⁹ and 1.69 × 10⁶, respectively. The cell size on the surface of face masks should not be larger than 1.0 mm; the height of the first layer cells near the face masks should not be larger than 0.1 mm; and the time step sizes discretizing the breathing and coughing periods should not be more than 0.01 s and 0.001 s, respectively. The results given by LES model show closer agreement with the experimental data than RANS models, with approximately 10% relative deviation for the air speed near the face mask. Overall, the SST k-ω model performs the best among the RANS models, especially for the air speed. The findings obtained form a CFD modeling framework for an accurate prediction of airflow and dispersion problems involving face masks.