About the choice of a turbulence model in building physics simulations

Environmental concerns and growing computer resources have promoted the emergence of new simulation tools and especially Computational Fluid Dynamics (CFD). A lot of academic studies have proved the interest of this approach. Unfortunately, for non-expert users, it is sometimes difficult to deal with turbulence modelling. Indeed, in this area, conclusions of previous studies are very fickle and sometimes contradictory. This paper aims to evaluate the impact of the turbulence modelling with different Reynolds-Average Navier-Stokes (RANS) techniques applied to airflows inside buildings. It appears that differences are often very small and completely insignificant at the design stage. However, when numerical resources requirements and computing time are considered, important differences appear in favour of the simplest model.