Prediction of pollutant dispersion in buildings: analysis of the gradient-diffusion hypothesis

T. van Hooff, B.J.E. Blocken, P. Gousseau, G.J.F. van Heijst

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review


Numerical simulations of room airflow using Computational Fluid Dynamics (CFD) are often based on the Reynolds-averaged Navier-Stokes (RANS) approach. Using this approach, only the averaged quantities are computed, whereas the effect of turbulence on the mean flow is modeled. Since the RANS approach does not provide information on velocity and concentration fluctuations, also the turbulent mass fluxes should be modeled. In the majority of the cases this is done by employing the standard gradient-diffusion hypothesis, which relates the turbulent mass flux to the mean concentration derivatives. In this paper a CFD analysis of pollutant dispersion in an enclosure ventilated by a transitional wall jet (Re˜ 2,500) is presented, using validated high-resolution RANS and Large Eddy Simulations (LES). The LES simulation shows that a counter-gradient turbulent mass flux is present, indicating that the standard gradient-diffusion hypothesis used in RANS is not valid in the entire flow domain. However, it is shown that for this particular case, the convective mass fluxes dominate over the turbulent mass fluxes, and that the predicted pollutant concentrations by RANS will therefore not differ significantly from those by LES. Paper_ID613
Original languageEnglish
Title of host publicationProceedings of Healthy Buildings 2015 Europe, 18-20 May 2015, Eindhoven, The Netherlands
Place of PublicationEindhoven
Publication statusPublished - 2015
EventHealthy Buildings Europe 2015, HB 2015 - Eindhoven University of Technology, Eindhoven, Netherlands
Duration: 18 May 201520 May 2015


ConferenceHealthy Buildings Europe 2015, HB 2015
Abbreviated titleHB 2015 Europe
Internet address


Dive into the research topics of 'Prediction of pollutant dispersion in buildings: analysis of the gradient-diffusion hypothesis'. Together they form a unique fingerprint.

Cite this