FORWARD SMOLDER OF POLYURETHANE FOAM IN A FORCED-AIR FLOW

Citation
Jl. Torero et Ac. Fernandezpello, FORWARD SMOLDER OF POLYURETHANE FOAM IN A FORCED-AIR FLOW, Combustion and flame, 106(1-2), 1996, pp. 89-109
Citations number
27
Categorie Soggetti
Engineering,"Energy & Fuels",Thermodynamics
Journal title
ISSN journal
00102180
Volume
106
Issue
1-2
Year of publication
1996
Pages
89 - 109
Database
ISI
SICI code
0010-2180(1996)106:1-2<89:FSOPFI>2.0.ZU;2-F
Abstract
An experimental study is conducted of forward smolder of polyurethane foam. Air is used as oxidizer, and is forced in the direction of smold er propagation under conditions that produce approximately one-dimensi onal forward smolder propagation. The objective of the study is to pro vide further understanding of the mechanisms controlling forward smold er and verification of theoretical models of the problem. Upward and d ownward forward smolder are compared to also observe the effect of buo yancy on the process. Measurements of the temperature histories at sev eral locations throughout the foam sample are used to infer the charac teristics of the smolder process, and to calculate the smolder propaga tion velocity along the sample length as a function of the air flow ve locity and gravitational orientation. It is found that as the flow vel ocity is increased, there is a transition in the smolder characteristi cs from a smolder process that is characterized by the propagation of a single exothermic oxidation (smolder) reaction to one characterized by the propagation of two reactions, an oxidative smolder reaction pre ceded by an endothermic pyrolysis reaction. Buoyancy is observed to af fect this mode of smolder at very low air velocities, or when the smol der front approaches the sample end. The smolder velocity data are cor related well in terms of a nondimensional smolder velocity derived fro m previously developed theoretical models of forward smolder. The good agreement between theory and experiments verifies that the smolder co ntrolling mechanisms and simplifying assumptions implicit in the model s are appropriate at least for the present experimental conditions.