Effectiveness of RANS in predicting indoor airborne viral transmission: A critical evaluation against LES

We investigate the dispersal of droplet nuclei inside a canonical room of size 10×10×3.2m 3 with a four-way cassette air-conditioning unit placed at the center of the ceiling. We use Reynolds averaged Navier–Stokes (RANS) simulations with three flow rates corresponding to air changes per hour (ACH) values of 2.5, 5, and 10. The room setup as well as the operating conditions are chosen to match those of a recent high-fidelity large eddy simulation (LES) study. We use statistical overloading with a total of one million droplet nuclei being initially distributed randomly with uniform probability within the room. Six nuclei sizes are considered ranging in radius from 0.1 to 10μm (166,667 nuclei per size). The simulations are one-way coupled and employ the Langevin equations to model sub-grid motion. The flow and particle statistics are compared against the reference LES simulations, and we find that the RANS k−ɛ realizable model may be used as a computationally cheaper alternative to LES for predicting pathogen concentration in confined spaces albeit, with potentially increased statistical discrepancy.