Numerical modeling of cough airflow: Establishment of spatial–temporal experimental dataset and CFD simulation method

Abstract In this study, the boundary conditions to simulate the cough airflow of a human using computational fluid dynamics (CFD) simulation have been defined. Spatial–temporal datasets of cough airflow were obtained from particle image velocimetry (PIV) experiments conducted previously, and an ensemble-average analysis was performed for CFD calculations and to validate simulated results. The air velocity corresponding to the boundary conditions for the opening of the mouth for CFD simulation could not be obtained directly from PIV experiments. Hence, the experimental results of airflow near the mouth were used to explore the boundary conditions of the cough for CFD simulation, assuming that the velocities near the mouth and the boundary conditions are highly correlated. The time-series velocity boundary conditions for the simulation of cough were explored using two methods. The conditions for Method 1 were closer to that for the overall cough airflow distribution obtained from the experimental datasets, while the conditions for Method 2 most closely matched the maximum velocity at a specific location of the cough airflow. The cough airflow reproduced by CFD using Method 1 showed RMSE values of 0.24–0.32 m/s compared to the experimental results when the maximum velocities at boundary conditions were 15.5–15.8 m/s. Moreover, for Method 2, the maximum velocity was reproduced with an error of 0.02 m/s. This study obtains the detailed boundary conditions for cough simulation, which, coupled with droplets and droplet nuclei dispersion models, can be applied to investigate the risk of infection and transmission routes of human cough.