Sample width and thickness effects on horizontal flame spread over a thin PMMA surface

Author:  Jiang, L., Miller, C. H., Gollner, M. J., Sun, J. H.

Journal: Proceedings of the Combustion Institute

DOI:  10.1016/j.proci.2016.06.157

Keywords: Horizontal flame spread, Convection, Radiation, Width, Thickness, cross air-flow, pool fires, fuel surface, feedback, Thermodynamics

Abstract: 

In previous studies, it was found that there exists a minimum flame spread rate under a certain range of sample widths for steady burning horizontal flame spread. While this was hypothesized to occur due to a transition between convectively-dominated to radiation-dominated flame spread, no measurements were performed to quantify this process. This paper presents a detailed experimental study investigating sample width and thickness effects on steady horizontal flame spread, including detailed measurements of the components of radiation, convection, and conduction. Water-cooled heat flux gauges, R-type micro-thermocouples traversed through the gas phase, and K-type thermocouples embedded in the solid phase were all used to deduce these heat transfer components. Results show that convective heat transfer decreases with increasing sample width as the shape of the flame front is on average farther from the fuel surface, while radiation increases as the view factor from the fire to unignited fuel increases with larger sample size. Conduction measured within the fuel sample is, as expected, confirmed to be negligible. Comparing a combination of these components, the total heat flux first decreases as the competition between radiation and convection changes, followed by steadily increasing heat fluxes as the width of the sample increases. Heat feedback also influences the sample pyrolysis rate, so there was a coupled response following this trend. The apparent dip followed by an increase in total heat flux can now explain why a period of minimum flame spread rate exists. Modification of an existing theory also matches experimental results very closely. Finally, a dimensionless heat-release rate for different sample configurations is used to scale the dimensionless flame heights with a power-law correlation having exponents 0.39 for Q* > 1 and 0.6 for Q* < 1, closely resembling the 2/5 and 2/3 predicted by Zukoski's model. (C) 2016 by The Combustion Institute. Published by Elsevier Inc.