Soot yield values for modelling purposes - Residential occupancies
Robbins A. P. and Wade C. A. 2008.
BRANZ Study Report 185
BRANZ Ltd, Judgeford, New Zealand
Abstract
The Department of Building and Housing (DBH) has convened a working group to develop a fire engineering framework for performance-based design specifying design fire scenarios, design fire characteristics and acceptance criteria. They have been reviewing a number of case study examples on sample buildings, generally compliant with C/AS1 to test proposed fire design scenarios and input parameters. Handbook values for the smoke yield (the mass of smoke/soot produced per mass of fuel burned) based on flexible polyurethane foam have been assumed for design fires involving upholstered furniture. It is suspected that the predictions for the concentration of smoke are significantly higher than would be experienced in practice, and as a result these are contributing to the buildings failing the proposed acceptance criteria relating to visibility. The objectives of the research were to:
1. Conduct a sensitivity analysis on the smoke yield parameter when used with two commonly used fire models (BRANZFIRE and FDS) by simulating a well documented experiment and varying the input smoke yield between 0.05 and 0.2 kg/kg.
2. Compare model predictions for the optical density (which is related to smoke concentration and visibility) with the selected experimental data described in a report by the National Institute of Standards and Technology (NIST) (Bukowski et al 2007; Peacock et al 2002).
3. Present the results and make a recommendation on an appropriate smoke yield (kg/kg) to be included in the design fire specifications to be published by the DBH.
For the range of smoke/soot yields considered in this study (0.05–0.20 g/g) both FDS and BRANZFIRE provided generally conservative predictions of smoke optical density for the flaming upholstered armchair located within a manufactured home, described in detail in this report. In modelling the same experiment, a soot/smoke yield of 0.05 g/g generally provided the closest but conservative agreement with the experimental measurements of smoke density. With a soot/smoke yield of 0.20 kg/kg the models over-predicted the smoke optical density by up to a factor of 10. For the upholstered armchair used in this experiment, assuming a handbook value for soot yield based on flexible polyurethane foam did not provide a good prediction of smoke optical density. Recommendations based on the results of this research included selection of a design fire value for soot/smoke yields derived from optical measurements taken during flaming combustion of full-sized items of upholstered furniture. In these measurements, the upper 95th percentile value of the smoke yield based on 24 items of upholstered furniture (with outlying data points removed) was 0.07 kg/kg.