Evaluation of CFD Modeling Methods for Fire-induced Airflow in a Room

Methods for exploiting the technique of computational fluid dynamics (CFD) to model fires and the associated fundamental processes of combustion and radiation are appraised by applying the CFD model JASMINE to simulate the fire-induced airflow in a room-sized compartment. For a systematic evaluation of these methods describing the fundamental processes and the associated mathematical formulation, a two-stage-verification and validation procedure is adopted here. In the verification stage, a number of important model parameters, e.g., the choice of fire source model (e.g., combustion model vs. volumetric heat source (VHS) model), radiation transport model (e.g., six-flux vs. discrete transfer radiation model), gas property model (e.g., constant absorption coefficient vs. grey gas model), design and resolution of the numerical grid, etc. are assessed by comparing the predicted results against a specific fire scenario selected from the data of Steckler, K.D., Quintiere, J.G. and Rinkinen, W.J. (1982). Flow Induced by Fire in a Compartment, NBSIR B2-2520, National Bureau of Standards, Pittsburgh [2]. The validation stage has then focused on the detailed comparison of the CFD fire model against the comprehensive data set of Steckler et al. [2] for a wide range of fire scenarios covering a range of fire strengths (i.e., heat release rates), fire source locations, the sizes and shapes of the ventilation opening (i.e., door or window). Such a systematic study has demonstrated the predictive capability of the CFD methodology in reproducing detailed thermal and flow field behavior of the enclosure fires, and has also shown that the predictive accuracy of the CFD methodology can be significantly improved by careful selection of the fire models and model parameters. Some guidance on the use of CFD methodology has been provided for performing the fire modeling properly and effectively for performance-based fire safety design.