Validation of different Eulerian and Lagrangian solvers of aerosol dispersion in indoor spaces

Abstract

The corona pandemic accelerated a lot of studies about aerosol dispersion and different aerosol-generating tasks ranging in intensity from sneezing to breathing. Both measurements and numerical simulations were used to understand the behaviour of aerosols. For numerical simulations, Computational Fluid Dynamic (CFD) simulations were used; however, the set-up of cases varied between studies. Different solvers, methods, turbulence models and steadiness are used depending on the scope and aim of each study. The aim of this study is to compare different set-ups and solvers and validate them against measurements conducted in the Senselab at the Delft University of Technology. The purpose is to find the best approach that balances between accuracy and computational cost to use afterwards in ventilation design decision-making. Consequently, we set up several numerical cases with different levels of complexities (e.g.: eulerian-eulerian to eulerian-lagrangian, including/excluding temperature and relative humidity, steady/unsteady). We then compare those cases to the experiments of a breathing manikin in the Senselab. The performance of each case is determined depending on how well it predicts aerosol dispersion and the run time cost.