On the extrapolation of CFD results for smoke and heat control in reduced-scale set-ups to full scale: Atrium configuration

Abstract

It is a common practice to use reduced-scale experiments to develop formulae for the design of smoke and heat exhaust ventilation systems. Implicitly, up-scaling of results is assumed justified. A similar approach can be adopted with numerical simulations, i.e. a reduced-scale setup can be up-scaled to compare results to full-scale observations. However, both in numerical simulations and in experiments, scaling must be done in a proper way. The classical method for up-scaling results obtained in fire related experiments is based on preservation of the Froude number only. The present paper, focusing on the up-scaling of results by means of a series of CFD simulations of fire-induced flows in an atrium configuration, confirms this to be justified as long as the flows in both the reduced-scale and full-scale configurations are sufficiently turbulent. If this is not the case, it is illustrated that other dimensionless numbers must also be preserved in scaling, in addition to the Froude number.