Experimental Study Of Heat Transfer To Ceiling From An Impinging Diffusion Flame

Abstract

A one-zone model (NRC C) and a two-zone model (CFAST) have been tested against experiments and used in the fire research community. However, the selection of either of them as a sub-model for a fire safety system model needs further verification against the experimental results under a variety of fire scenarios. In the present study, the performance of the above two models is evaluated with respect to accuracy, efficiency and simplicity. This study indicates that significant discrepancies exist between the experimental results and the results from both the CFAST model and the NRCC model. The CFAST model generally over-predicts the upper layer temperatures in the burn room but provides reasonable predictions in the adjacent enclosures. The CFAST model over-predicts CO concentrations when the air-handling system is turned on; the model under-predicts CO when the system is off. The NRCC model correlates the burning characteristics of the fuel with the burn room conditions. The results give a good agreement with the experimental results in the burn room for some fire scenarios but discrepancies exist for others. However, the NRCC fire model is applicable to the bum room only. It was found that the room-averaged temperatures for the bum room obtained from the CFAST model were in a good agreement with the experimental results and the NRCC model results. For the purpose of the fire safety system model, the major weakness of the CFAST model is the requirement of prescription of the mass or heat release rate, while the system model simulates hundreds or thousands of fire scenarios. Considering the efficiency and simplicity of the fire growth model required for the system model and the accuracy of the prediction, the one-zone NRCC model is recommended for the system model to predict the bum room conditions and a simple two-zone model for the adjacent enclosures.