Comparisons of 'Blind Predictions' of a CFD Model with Experimental Data

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.