Limitations on Use
Recommendations regarding room sizes
Wade, C.A. and Robbins, A.P. 2008. Smoke filling in large spaces using BRANZFIRE. BRANZ Study Report No. 195. BRANZ, Judgeford, New Zealand. [Download]
This study provides guidance to fire engineers, design reviewers and regulators regarding the limitations and appropriateness of using a two-zone fire model to simulate smoke development in large spaces. Smoke filling of seven enclosures ranging in area from 625–5,000 square metres and height 6–12 m was simulated using the zone model BRANZFIRE and the computational fluids dynamics model FDS and the results were compared. For the cases examined, it was found that the BRANZFIRE zone model provides good agreement with FDS for room areas up to about 1200 sqm.
Multicell (virtual room) simulations were also investigated using the zone model and were found to provide more realistic representation of the smoke layer position compared to a zone model single room simulation; however, average gas temperatures may be overestimated close to the fire plume and underestimated far from the plume. It was recommended that, for larger enclosures up to 5000 sqm, additional simulations and sensitivity analysis should be conducted subdividing the enclosure into virtual rooms as well as a single zone simulation.
This study did not consider compartments larger than 5000 m2 or higher than 12 m. For these cases, the model user should consider the use of a CFD model, or conduct further analysis of their own to justify the use of a zone model. The virtual room approach may still be useful for these cases.
Where the building geometry and arrangement of ventilation openings is complex and could significantly impact on the characteristics of the smoke flows, then users should consider whether simulation using only a zone model is sufficient. In these cases, CFD simulation may be more appropriate.
Refer to the above report for further information and explanation.
Plume Entrainment in Tall Spaces
The McCaffrey plume correlation (used in branzfire) has known limitations in tall spaces, and will overestimate the plume entrainment. The Morton-Turner-Taylor theory suggests that entrainment in the buoyant plume region should depend on H^(5/3), whereas the McCaffrey correlations uses the 1.895 power for the buoyant plume. According to Rockett [1], the McCaffrey correlations result in a severe over-prediction of plume flows far above the fire (i.e in tall spaces) but provides good predictions in immediate over-fire region more relevant for small rooms. Over estimating the entrainment will result in upper layers that are deeper and cooler than otherwise expected. This could have a large impact with added cost implications if used to size a smoke extract system based on raising the layer height.
1. Rockett, John A. 1995. Zone Model Plume Algorithm Performance. Fire Science & Technology Vol 15 No 1 & No 2. p1-15.