Arun Mahalingam
2007
Combustion products generated in enclosure fires can be transported throughout
the enclosure causing death and injury to occupants and a great deal of damage
to property and the environment. The ability to estimate the generation and
transport of toxic combustion products in real fire scenarios involving common
building materials is of great importance to fire protection engineers in
producing detailed quantified risk assessment and in the design of fire-safe
buildings. Most common building materials are polymer based. Thus toxic products
evolving from burning polymers is the single most important factor in fire
fatalities. Fire hazard calculations require modelling of heat generation, toxic
combustion products generation and its transport in realistic building scenarios
involving common building materials. However, the thermal decomposition,
combustion behaviour and chemical kinetics for common polymers like wood,
plastics, rubber and textiles are extremely complex. In the present study, a
methodology (STEM-LER: the Scalar Transport Equation
based Model using the Local Equivalence Ratio
concept) based on solving separate transport equations for the species and using
the yield correlations obtained from bench-scale experiments to model the source
terms is proposed to predict the products generation and its transport during
enclosure fires. Modelling of complex solid phase degradation and chemical
kinetics of polymers is bypassed by measuring the product yields as a function
of equivalence ratio by burning the samples in a bench-scale combustion
apparatus called Purser furnace. Since the accuracy of prediction depends upon
the quality of the yield data obtained from the Purser furnace, attempts were
also made to numerically investigate this bench-scale toxicity test method in
order to understand its modus operandi.
Also, large-scale fire tests were carried out involving combustible cable
materials to generate the validation data for the combustion and toxicity models
developed in the current work. Simulations were carried out using the proposed
methodology to simulate the large-scale cable fires and validated with the
experiments. Irrespective of the errors involved in approximating the complex
physical and chemical processes, the STEM-LER methodology is able to
model the combustion products generation near the fire and its transport to
distant locations with reasonable accuracy.
Finally, a preliminary assessment on the effect of cable fires on building
evacuation for the simulated fire scenarios were carried out using a
sophisticated evacuation model.