Citation

Van Maele K, Merci B. Fire Safety J. 2006; 41(2): 122-138.

Copyright

(Copyright © 2006, Elsevier Publishing)

DOI

unavailable

PMID

unavailable

Abstract

The accuracy of results of computational fluid dynamics simulations of fires strongly depends on the turbulence model applied when the Reynolds-averaged Navier-Stokes approach is used. In particular, the effect of buoyancy on turbulence is important for fire-driven flows. In this work, the standard and a realizable k-[epsilon] model are addressed. Both the simple and the generalized gradient diffusion hypothesis are applied for the calculation of the buoyancy production of turbulent kinetic energy. Simulation results are presented for the axisymmetric free buoyant plume and the plane buoyant wall plume. The buoyancy modification based on the simple gradient diffusion hypothesis has a negligible influence on the results for both test cases. The realizable k-[epsilon] model with modifications based on the generalized gradient diffusion hypothesis performs well for the test cases considered.