Thomas J. Ohlemiller

An experimental comparison of forward and reverse smolder propagation in permeable fuel beds

Abstract An experimental comparison of forward and reverse smolder propagation is presented; these two cases correspond, respectively, to reaction wave movement in the same and in the opposite direction of the air flow. Two types of permeable fuels were examined, a cellulosic loose fill insulation (wood fibers) and a particulate polymer material (polyisocyanurate). Air flow velocity was varied from 0.04 and 0.77 cm/s in the reverse smolder studies, yielding propagation velocities from 4·10−3 to 4·10−2 cm/s; air flow velocity was varied 0.15 to 0.49 cm/s in the forward smolder studies, yielding propagation velocities from 1·10−3 to 2.5·10−3 cm/s. Thermocouples and gas sampling (O2, CO, CO2) were used to probe the structure of the two types of smolder wave. There are marked qualitative and quantitative differences in the two smolder propagation modes. Reverse smolder quickly reaches a steady propagation rate determined largely by heat transfer processes; forward smolder propagation is unsteady and moves at a lower rate that appears limited by the stoichiometry of char oxidation. Both modes of propagation are ultimately limited by the rate of oxygen supply.

Smoldering combustion propagation through a permeable horizontal fuel layer

Abstract Although the propagation rate of smoldering through porous horizontal fuel layers has been measured for a variety of materials, there has been little work on the structure of the smolder reaction zone and the factors controlling it. These latter aspects are the focus here for the case of thick (18 cm) layers of wood-based fibers in the form of cellulosic insulation smoldering under natural convection air supply conditions. Two-dimensional profiles of temperature, oxygen mole fraction, and residual organic material have been measured both for unretarded insulation and for insulation having 25 wt% of the smolder retardant, boric acid, added on. It is inferred that the overall wave structure is dominated by oxygen diffusion from above. The heat release chemistry appears to involve both oxidative pyrolysis and char oxidation in a shifting balance depending on depth in the layer. Boric acid is unable to halt the smolder process in these thick fuel layers but it slows its spread by about a factor of 2 by a combination of endothermic and kinetic effects.

Wood gasification at fire level heat fluxes

Abstract This study was motivated by a need to understand the source of pollutant species emitted by wood burning stoves; the results are relevant also to fire research. The study examines the products generated when wood is heated in controlled conditions, without flaming, in atmospheres of varying oxygen concentration (0–21 % O 2 in N 2 ). Small wood samples (typically 4 × 4 cm exposed face, 2–4 cm thick; mainly white pine and red oak, but also two tests with yellow pine) were subjected to uniform radiative heat fluxes (2–7.8 W/cm 2 ) on one face. Sample weight was followed in some tests and sample temperature (5 thermocouples in depth) in others since the two measurements could not be made together. In all tests, all evolved products were either monitored (H 2 O, CO, CO 2 , total hydrocarbons not condensible at −40°C) or trapped and analyzed (condensible organic species) by gas chromatography and mass spectroscopy. Many of the trends of the major products (CO, CO 2 , H 2 O, THC, total organic condensible or tar) are qualitatively intelligible in terms of the expected impact of varying temperature or oxygen level, for example. The extent of change in these major products is rather limited (factor of two to four) over the range of variables explored here. The organic condensate was difficult to analyze; it is estimated that only 20% of it would pass through a gas chromatograph. More than 40 species in this analyzable portion were positively or tentatively identified and quantified. Chromatographic fingerprints of the organic condensate indicated that its composition does not vary a great deal for the conditions examined here. The fingerprints from the radiative heating tests bear a strong resemblance to those of the smoke condensate from a wood stove.

Effects of external radiant flux and ambient oxygen concentration on nonflaming gasification rates and evolved products of white pine

Abstract Time-dependent gasification mass flux, sample temperatures, and evolved products (CO, CO2, H2O, and total hydrocarbons) of thermally thick white pine (approximately 3.8 cm cubes) were measured under the nonflaming condition at thermal radiant fluxes of 2.5–6.9 W/cm2 in three different atmospheres of N2, 10.5% O 2 89.5% N 2 , and air. Results indicate that ambient oxygen significantly increases the gasification mass flux (for example, at 2.5 W/cm2, the mass flux in air increases about three times over the mass flux in nitrogen), sample temperatures (surface temperature in air increases as much as 200°C over that in nitrogen), and char depth, and changes the evolved products distribution (CO2 yields in air are about 6–8 times larger than that in nitrogen and CO yields in air are about twice as large as that in nitrogen). Therefore, extreme care is needed to apply results obtained in an inert atmosphere to cases, such as ignition, flame spread, and smoldering, in which a wood sample is exposed to oxygen containing atmospheres.

Forced smolder propagation and the transition to flaming in cellulosic insulation

Abstract It is well known that a smoldering fuel responds to an increased oxygen supply by becoming faster and hotter until, eventually, flames erupt. This sequence was examined quantitatively for thick horizontal layers of a permeable fuel, i.e., cellulosic insulation. Two configurations are possible, forward and reverse smolder; both were investigated experimentally. Reverse smolder was shown to respond only weakly to an increased air flow and it exhibited no transition to flaming at flow velocities up to 5 m/s. Forward smolder responded strongly to increased air flow and yielded transition to flaming at about 2 m/s for unretarded material. The influence of combustion retardants was also examined; these included boric acid, a smolder retardant, and borax, a flaming retardant. Both prevented the transition to flaming in the absence of adjacent flammable material but were less effective in its presence. The overall response of these various fuel mixtures and configurations suggests that both kinetics (via leading edge heat transfer effects) and oxygen supply rate (not the latter alone) play substantial roles in dictating smolder response to an air flow.

Upward flame spread on composite materials

Three existing models of upward flame spread were tested against intermediate-scale experiments on a vinyl ester/glass composite. Characterization of rate of heat release per unit area, needed as input to the models, was obtained at external radiant fluxes below the minimum for ignition by adaptation of a method due to Kulkarni. There are several limitations on the accuracy of the material characterization when applied to composites. Each of the flame spread models has definite limitations as well. Nevertheless, all three models produced predictions of spread behavior in sufficiently quantitative agreement with the experiments that they should prove useful for engineering analyses of flame spread potential.

The effect of surface coatings on fire growth over composite materials in a corner configuration

Structural composites are vulnerable to fire in two respects: (1) their resin content may ignite and enable the spread of flames over the surface of the structure; (2) the resin may degrade from the heat of a localized fire exposure thus weakening the composite structure. The present study focuses mainly on the first issue, in particular, on the ability of various commercial coatings to prevent flame spread. The second issue is examined briefly by applying thermocouples to the back surface of test specimens. Four commercial coatings have been tested over an unretarded vinyl ester/glass composite. In addition an uncoated phenolic/glass composite and a polyester/glass composite coated with a fire retarded resin were tested. In all cases the configuration was a 3.3 m high corner with a 53 cm square propane gas burner at its base, operated at 250 kW as the fire exposure. The results show that, with the proper choice of coating and coating thickness, fire growth can be suppressed quite effectively. Two of the coatings, applied at a substantial thickness, were reasonably effective at slowing the penetration of heat to the back of the composite panels. The other coatings, much thinner in application, were notably less effective at slowing heat penetration.

Effect of ignition conditions on upward flame spread on a composite material in a corner configuration

This paper focuses on the issue of fire growth on composite materials beyond the region immediately subjected to an ignition source. Suppression of this growth is one of the key issues in realizing the safe usage of composite structural materials. A vinyl ester/glass composite was tested in the form of a 90° corner configuration with an inert ceiling segment 2.44 m above the top of the fire source. The igniter was a square propane burner at the base of the corner, either 23 or 38 cm in width, with power output varied from 30 to 150 kW. Upward flame spread rate and heat release rate were measured mainly for a brominated vinyl ester resin but limited results were also obtained for a non-flame retarded vinyl ester and a similar composite coated with an intumescent paint. Rapid fire growth to the top of the sample was seen in replicate tests for the largest igniter power case; the intumescent coating successfully prevented fire growth for this case.

Low heat-flux measurements: Some precautions

Simple experiments are described for the purpose of illustrating measurement errors and their avoidance during use of Gardon or Schmidt-Boelter total heat flux sensors. These errors can assume serious proportions of the observed signal at flux levels below about 15 kW/m2. They result from two sources, both a consequence of the flux gages temperature relative to its surroundings: firstly, convective heating of the gage by the boundary layer from hot surfaces surrounding it and second, heat exchange with the ambient environment by radiation and convection. Some proposals are made for standardizing measurement methods, but it seems unlikely that errors in measurement can be completely eliminated. It thus becomes important for users of experimentally derived data to understand the limitations which may exist in the data reported.

Characterization of the California technical bulletin 133 ignition source and a comparable gas burner

The California Bulletin (CB) 133 upholstery ignition source is based on the use of crumpled newsprint. The present work examined the reproducibility of several aspects of this source when placed on an inert chair mock-up. The tendency of this source to heat the side arms of a chair, the area of the seat back subjected to high heat fluxes, the peak flux there and the flux duration all showed substantial variability. For inherently lesser variability a gas burner is preferred. A gas burner, derived from that developed at the British Fire Research Station, was shaped so as to deposit a similar pattern of heat to that of the CB 133 source. The two sources were tested for comparability both on chair mock-ups and on full-scale chairs made from a wide variety of materials. The results indicate that the gas burner, as used here, is a somewhat less severe ignition source than is the CB 133 igniter.