No abstract

A CFD study is reported which attempts to shed light on the problem of estimating air entrainment into buoyant balcony "spill plumes". This topic has been of some controversy for several years but is of substantial practical importance to the design of smoke control systems in atria and other multi-level complexes. The study described here concerns numerical simulation of the reduced scale physical model from which much of the semi-empirical formulae have been derived. Excellent agreement between measured and predicted mass fluxes are demonstrated in the regions of lower heights of plume rise, but the two apparently diverge at greater heights of rise. Two possible explanations involve the apparently unrealistic assumption of a horizontal smoke layer surface and the validity of the turbulence modelling of this flow. The study does support recent dimensional analyses which suggest only a small degree of air entrainment into the turning region of the smoke flow from the horizontal to the vertical .

This paper presents the first truly 'blind' test of a CFD model used for the prediction of fire conditions in an enclosure. This test formed part of an auditable 'round robin' test of models conducted by CIB W14, sub group 2. The paper demonstrates that CFD models containing the same sub-models as those used in JASMINE are fit for the purpose of predicting gas phase conditions to better than 15% in a flashed over enclosure of the dimensions used here. For accurate prediction of surface heat fluxes in these conditions, however. the quasi-steady one dimensional conduction approximation, satisfactory for smoke movement problems. is not adequate.

The paper describes progress in the development of a numerical modelling methodology for predicting the environmental consequences of fires in warehouses. Two separate computational fluid dynamics models have been used to predict the emission of combustion products through warehouse roof openings and then to predict the atmospheric dispersion of the effluent. At this stage of development the dispersion model employs as boundary conditions the outflows, through the roof vents, predicted by the enclosure fire model.Results are presented for a representative warehouse containing a burning high-rack of storage with the hot gases vented to the external environment through nine roof vents. The key parameters for dispersion modelling are identified, and an example is provided using the emission from the nine roof vents. Further work is identified for the development of a practical hazard analysis tool. KEYWORDS

No abstract