| Summary: | Previous studies have demonstrated upward flame spread on vertical surfaces to be one of the most serious fire scenarios due to the concurrent direction of flame propagation and air flow. The unburned zone of the wall ahead of the pyrolysis zone consequently receives more heat from the flame, causing the temperature of the unburned wall to rise quickly and the flame to spread more rapidly. In order to prevent the occurrence of hazardous wall fires, there is a need to select materials which satisfy performance-based regulations. This approach needs information from reliable fire models which simulate accurately the fire behaviour of materials in their end-use configuration. However, none of the existing models are used in this way. This is because of the limitations of the models themselves, in particular the assumptions involved and uncertainties in the empirical correlations used in the models. This study focuses on the early stages of the upward flame spread. An existing model which uses data directly from the Cone Calorimeter test was examined. A non-standard test procedure was developed which gives results capable of giving better predictions from the model. The flame height and heat feedback to the unburned wall were also examined and more information obtained. The flame height was measured from steady burning vertical fuels and from a gas-fuelled panel to examine the commonly accepted relation that the flame height is proportional to the 2/3 rds power of the heat release rate per unit width of the wall fire <i>Q’. </i>In the former experiment, the flame height and <i>Q’ </i>were measured directly in the same experiments, providing data for the early stages of fire spread. In addition, the effect of different geometric configurations was examined. The latter experiment revealed a width effect, questioning the validity of <i>Q’ </i>being the only parameter determining the flame height. Experiments were also carried out to measure the heat transfer to the unburnt material above the pyrolysing zone. A lower value was found to be more representative than the ones used previously. With these modifications, the model was found to give good agreement with experimental measurements of vertical spread on sheets of PMMA. Two additional cases were studied: wall fires influenced by an inert parallel wall and by the proximity of a corner. The measurement of flame height and heat feedback provides data for further investigation and modelling work.
|
|---|
