When fires occur in domestic or commercial buildings it is the smoke from the fire that leads to far more injury and death than the heat produced from the flames. Understanding the movement of smoke within the fire compartment and through openings in the enclosure is critical for designing buildings to prevent fire fatalities. Prediction of the movement of smoke is a complex phenomenon and is a continued focus of research throughout the world.

Work has been conducted in the past on the exchange flow rates through vertical openings, but very little has been done on horizontal ceiling openings. Current smoke transport calculations are most often carried out using standard vent flow models that do not accurately take in to account the buoyancy component of the flow. The fire zone model BRANZFire was developed with a ceiling vent flow algorithm based on the work of Cooper who found there was very little data on which to base his predictions. This report aims to provide additional experimental data on exchange flow rates through horizontal ceiling openings through the use of saltwater modelling and compare this to the work previously undertaken by Cooper.

Taking measurements of fire phenomena in hot and smoky environments can be difficult and expensive because the sooty environment and high temperatures involved can damage equipment and make taking accurate readings a challenge. Herein this problem is overcome through the use of a saltwater analogue system to model the conditions in a real fire scenario. The density difference created by a fire between the hot fire gases and the ambient air is replicated by using fresh and saltwater. The orientation of the experiment is inverted compared to the real life scenario as the saltwater which has the higher density is added to the fresh water. The saltwater is injected from a source on the ‘floor’ of the compartment into a tank of fresh water which generates a buoyant plume that ‘rises’ to the ceiling forming a distinct upper layer. Fluid in this layer exchanges with the ambient fluid through the ceiling opening.

The saltwater is dyed and Light Attenuation (LA) is used to discern the density of the fluid and hence the amount of mixing that has occurred. This can then be used to determine the amount of exchange flow through the ceiling vent.

An integral model for the descent of the interface between the hot smoky zone and the cool ambient zone has been developed and was found to perform well when compared with the saltwater experiments and another predictive model developed by Turner and Baines. The model was then developed further using mass conservation conventions to calculate the exchange flow through the ceiling opening.

The exchange rate through the ceiling opening was calculated and was found to compare well with Cooper’s algorithm when an equivalent fire size of 323 kW was used but differed significantly when a fire twice this size was considered. It was found that Cooper’s method did not adequately take into account the difference in fire sizes as the exchange flow predicted was almost identical between fire sizes for a particular ceiling vent. The implications of this are that the exchange, and hence the mixing and the amount of smoke, may be under predicted using larger fires in BRANZFire and this could lead to non-conservative design.