Vertical flame spread with horizontal projection

This report consists of two main parts. The first part focuses on the results from several small
scale experiments. A vertical and combustible wall was exposed to external venting flames 
from a lower compartment. A non-combustible horizontal projection was used to reduce the 
direct impingement of the venting plume on the wall. The wall consisted of particle board and 
the horizontal projection was made of calcium silicate. The effectiveness of the horizontal 
projection in preventing vertical fire spread along the wall was studied for six different heat
release rates (33, 37, 42, 47, 51 and 56 kW).  
The study investigated how different projection sizes impacted flame height, ignition, vertical 
temperature development, and char formation. In this study, 0.075 m and 0.1 m projections 
were used principally, while other projections (0.05 m and 0.2 m) were used for specific cases. 
The 0.1 m projection was more effective in reducing vertical flame heights and temperature 
development compared with the 0.075 m projection. The 0.1 m projection successfully 
prevented ignition on wall at 33 kW and 37 kW fires during the 10 min experimental duration, 
while the 0.075 m projection prevented ignition at 33 kW fire only. However, at higher heat
release rates, the wall ignited with both projections. Once ignition occurred, flame spread 
across the wall surface due to the combined effect of the burner and combustible gases from 
the pyrolysis process. The flame spread decreased after char formation on the wall. Char area 
and depth increased with higher heat-release rates for both projections, although there were no 
significant changes in vertical char height. A longer projection (0.2 m) was more effective by 
preventing ignition against 47 kW fire. The longer deflection distance restricted continuous 
attachment of flames and heat transfer to the wall.  
The second part of this study involves a computational simulation study. A three-storey 
building model was constructed using the FDS software ‘PyroSim’, where the flame started at 
the bottom compartment. A 3.1 MW fire was simulated and spread to the upper floors through 
the window. Two types of projections of variable lengths (0, 0.4, 0.6, and 0.8 m) were placed 
over each floor. One projection type was an open-ended ceiling projection, while the other one 
was solid wall balustrade (such as balcony with railing). The exterior wall or façade of the 
building, along with the projection didn’t include any combustible material. The simulation 
study allowed visualization and comparison of the flame trajectory and thermal behaviour of 
the ascending plume as it interacted with each projection type. Both projection types of several 
lengths (0.4 m, 0.6 m and 0.8 m) effectively reduced thermal impact, by 39% to 91%, on the 
upper floors compared with no projection, while the reduction increased with projection 
lengths. However, the temperature distribution at upper floor walls was affected by the 
projection types and lengths. Additionally, the thermal impact on a nearby building was also 
studied in these fire scenarios.