Citation

Himoto K, Iwami T. Fire Safety J. 2021; 125: e103418.

Copyright

(Copyright © 2021, Elsevier Publishing)

DOI

10.1016/j.firesaf.2021.103418

PMID

unavailable

Abstract

Spot ignition by wind-dispersed firebrands is an important factor in the spread of fire in large outdoor fires. However, the large probabilistic variability in its generation, transport, and ignition processes causes difficulty in quantifying its risk. In this study, a series of wind tunnel experiments was conducted to understand the probabilistically varying characteristics of two among three processes involved in spot ignition, i.e., the generation and transport of firebrands. Analysis samples were generated by combusting a wood crib in an apparatus that simulates the behavior of a fully developed structural fire. The experimental parameters were the ejection height of firebrands, H, crosswind velocity, U∞, and thickness of wood stick used for assembling the wood crib, d. Utilizing the experimental results, the projected area of a firebrand and its transport distance downwind of a fire source were formulated based on simple physical considerations. However, because the obtained correlations represent deterministic relationships of the parameters, they were further incorporated into probability distributions (a truncated exponential distribution for the projected area and a truncated normal distribution for the transport distance) to enable the consideration of intrinsic variations. The obtained models reasonably reproduced the experimental results for both the projected area and transport distance. Although structural fire is the targeted type of fire source, the proposed models exhibit forms extensible to other types of fire sources.

Language: en

Keywords

Firebrand generation; Firebrand transport; Large outdoor fires; Statistical analysis; Wind tunnel experiment