Flammability Limits of Fire Resistant Fabrics

Transcription

1 Flammability Limits of Fire Resistant Fabrics Thomsen M. 1, *, Murphy D. C. 1, Fernandez-Pello C. 1, Urban D. L. 2, Ruff G. 2 1 University of California Berkeley, Department of Mechanical Engineering, CA,USA 2 NASA John H. Glenn Research Center, Cleveland, OH, USA *Corresponding author maria.thomsen@berkeley.edu ABSTRACT The selection of fabric materials used in fire safety critical applications, such as firefighters, car racers and astronauts suits, is an important task based on minimizing the risk that they can represent. Fire resistant materials are clearly preferred however their flammability properties can change depending of the ambient conditions at which they are exposed. An experimental study was conducted to analyze the influence of environmental variables such as oxygen concentration, ambient pressure and external radiant heat flux on the flammability properties of two different fire resistant fabrics: Nomex HT90-40 and a blend made of Cotton/Nylon/Nomex. In this work ambient pressure was varied between 40 and 100 kpa and ambient oxygen concentrations were decreased until the limiting conditions for flame propagation were found. Experiments using an external heat source of 5 kw/m 2 as well as no external heating were also conducted to examine the influence of the presence of a nearby heat source. It was found that as ambient pressure is reduced the oxygen concentration required for the flame to propagate must be increased. The external radiant heat flux acts as an additional source of heat and allows propagation of the flame at lower oxygen concentrations. An analysis of the propagation limits in terms of the partial pressure of oxygen suggest a higher flamability risk in reduced pressure environments for both of the materials studied. The experimental data provided in this work helps identify increased flammability risk of the materials tested under different environmental conditions that can be further considered for future applications. KEYWORDS: Environmental conditions, fire resistant, flame spread. INTRODUCTION Understanding how flame spread behavior is altered by different environmental scenarios represents an important factor from a fire safety point of view. The flammability of materials depends on several factors such as their physical and chemical properties, but also of the ambient conditions at which they are exposed. There are some particular situations where the type of material used is particularly important since any mistake can have disastrous consequences involving the safety of human lives. An especial situation is given in spacecrafts environments, where everything needs to be carefully controlled to avoid any potential accident that can compromise the success of the expedition and the safety of the astronauts. In this kind of applications, the atmospheric conditions used must be optimized in order to support human habitation while minimizing launch weight and all risks involved [1]. The selection of each material used is also extremely important since the possibility of ignition can have disastrous consequences. Changes in ambient pressure, oxygen concentration, forced flow, an external heat flux or gravity are some of the relevant environmental variables in this kind of scenarios. Pressure lower than atmospheric is also of importance in aircraft and high altitude cities [2] and an external radiation source is a common setting when a fire is exposed to an adjacent fire or an overheated element. This work addresses the effect of some of these variables, specifically an external radiant heat flux, reduced pressure and elevated oxygen concentration on the flammability characteristics of fire resistance fabrics. Proceedings of the Eighth International Seminar on Fire and Explosion Hazards (ISFEH8), pp Edited by Chao J., Liu N. A., Molkov V., Sunderland P., Tamanini F. and Torero J. Published by USTC Press ISBN: DOI: /c.sklfs.8thISFEH

2 Proceedings of the Eighth International Seminar on Fire and Explosion Hazards (ISFEH8) Several researchers have shown interest in the study of the flammability properties of materials and their flame spread behavior in conditions like those in space exploration vehicles. Olson et al. [3] conducted experiments using Nomex HT90-40 and found that the limiting oxygen concentration required for flame spread in micro gravity is about 4% lower than the one in normal gravity. Kleinhenz and Tʼien [4] tested Nomex III burning in an upward configuration under different environmental conditions and found the spread to be cyclic, with a cycle consisting of flame growth, split, blow-off of the downstream flame and re-growth of the upstream flame. Hirsch et al. [5, 6] tested different materials to identify their flammability risk under different oxygen concentrations and ambient pressure. Their results showed that the maximum oxygen concentration resulting in selfextinguishment depended inversely on the total pressure; however, they also noticed an opposite linear trend when considering the partial pressure of oxygen. Nakamura and Aoki [7] studied the flammability of thin samples of cellulosic paper under the effects of sub-atmospheric ambient pressure and different oxygen concentrations. Their results showed a wider flammable range at lower pressures, presenting the same linear trend found by [5]. Of particular interest it has also been the presence of an external heat radiant source in flammability related studies. Quintiere et al. [8] measured downward and lateral flame spread using a decaying heat flux, from the study an ASTM testing method was established to measure specific parameters useful in the prediction of ignition and fire behavior in combustible materials. Osorio et al. [9] studied the limiting conditions for flame spread focusing in the effects an external heat flux and oxygen concentration using a forced flow velocity for a fabric sample burning on one side. Their results showed that for a given oxygen concentration there is a minimum external heat flux that allows flame spread to occur. In the present work experiments were developed to study the effect of an external heat flux, oxygen concentration and reduced ambient pressure on the limiting conditions for flame spread over thin flame retarding fabric samples burning on a downward configuration under the influence of a forced/mixed flow. The results extended from this work will provide a better understanding of flammability characteristics of some materials and their suitability for use in different environments. EXPERIMENTAL APPARATUS The experiments were conducted in an upgraded version of the existing FIST-SEA pressure chamber [10] in order to be able to have a simultaneous control of: oxygen concentration, ambient pressure, flow velocity and external radial heat flux. A new duct and flow system were used to provide constant supply of fresh gases to the chamber and avoid vitiation problems. The duct, shown in Fig. 1, has a 125 mm square cross section and 600 mm total length. The first 350 mm section of the duct serves as a flow conditioner, where inlet gases pass sequentially through perforated stainless steel plates, a 30 mm layer of 3-5 mm borosilicate beads and 40 mm thick, aluminum honeycomb with 6 mm cells before entering the test section. This configuration was tested with hot-wire anemometry, at atmospheric pressure and found to provide steady flow that was uniform to within ±3.5% at the duct outlet. The other 250 mm segment of the duct is used as the test section. Flat samples are located vertically at the midplane of the duct. During the present experiments two materials were tested: a single layer of Nomex HT90-40 and a single layer of a fabric blend made of 29% Cotton, 31% Nylon and 40% Nomex. These materials are characterized as fire resistant and have been studied previously [9] in the context of understanding their flammability characteristics in environments similar to those of future space exploration vehicles. 424

3 Part II Fire Before placing them inside the duct, fabric samples were held between two stainless steel support plates 200 mm by 125 mm and 0.4 mm thick. Each plate had an identical rectangular opening 150 mm in the flame spread direction and 75 mm wide, which was the test sample area. Each sample holder was fitted with electrically insulated posts to connect Kanthal wires for ignition. These wires were sewn into the fabric samples at approximately 5 mm from the top of the opening in the holder. The side walls of the duct (normal to the plane of the samples) were clear polycarbonate with low profile pins protruding slightly (~2-3 mm) into the duct to constrain the samples and keep them straight. The walls parallel to the sample were 0.56 mm thick alkali-aluminosilicate glass to permit transmission of radiant flux from outside the duct. Four quartz near-infrared halogen lamps (Ushio QIH T/S) were used as the source of radiant flux. Each lamp had a lit length of 127 mm and a total length of 220 mm and were half surrounded and supported by parabolic reflectors (Research Inc A). Two of these heaters were located immediately outside of the alkalialuminosilicate windows on either side of the duct. The heaters were mounted on rails and positioned to provide a constant heat flux of 5 kw/m 2 over the entire sample. The distribution of heat flux from one side was measured using a Schmidt-Boelter radiometer. Lateral flux uniformity was 5%. Power to the lamps was controlled by a variable transformer (2000 VA / TDGC-2KM) and monitored by measuring both the voltage delivered to and current drawn by the ensemble of lamps. Figure 1. Schematic of experimental apparatus. Gas composition and pressure were controlled by admitting oxidizer gases through critical nozzles while constantly evacuating exhaust gases to maintain constant pressure. Dry compressed air and 425

4 Proceedings of the Eighth International Seminar on Fire and Explosion Hazards (ISFEH8) oxygen (Praxair 99.6% pure) were delivered to precision control orifices from OʼKeefe Controls, with pressure measured immediately ahead of the nozzles by mechanical gauges from Omega Engineering and Ashcroft Inc., respectively. In some cases, when the oxygen concentration requiered was low the dry compressed air was replaced with pure nitrogen to achieve the proper conditions. After metering, the two gas streams were mixed in the line, passed through a bulkhead in the pressure chamber and delivered directly to the bettom of the test duct. This process ensured that flow through the duct was continually fresh and maintained at a constant mass flow rate. The primary vacuum supply used was a high-capacity vacuum generator (Vaccon JS-300). This was connected to the pressure chamber by a mechanical vacuum regulator. Correct operation of this configuration was verified monitoring constantly the chamber pressure with and electronic pressure transducer (Omega Engineering, Inc.). Minor inconsistencies caused small deviations in the ambient pressure (~2 kpa), however the actual pressures achieved were recorded and are presented here. EXPERIMENTAL PROCEDURE The experimental procedure followed was the same for each of the experiments and was derived from the Oxygen Index tests described in NASA-STD-6001B [11]. Once the sample was prepared and placed in the holder, this was secured inside of the duct and the FIST chamber was sealed. The vacuum system was initiated and set to the ambient pressure required. After the operational pressure was achieved, the air/nitrogen and oxygen supply were activated and adjusted to provide the corresponding mass flow required to obtain the desired conditions. Next the heaters were energized, set to the operating voltage (40 V RMS for this work) and the actual voltage and current provided to the lamps was recorded. Once the heaters were on these conditions were maintained for 90 s before ignition in order to provide a consistent pre-heating for all the samples. After his period pressure was checked again to be at the expected values, the video recording was started and the igniter was energized using a controlled current power supply (BK Precision 1785) set to deliver 4 amps. The use of a constant current supply ensured power delivered per unit length of wire regardless of any changes in resistance in connectors. Videos of the experiments were recorded with a resolution of 1280 by 720 at 59 frames per second using a Nikon D3200 camera. A 532 nm dichoric filter was used to improve visibility by suppressing the substantially long wavelength light of the radiant flux. For each of the pressure values tested, when close to the limiting conditions, between three and five replicate experiments were done. RESULTS Downward flame spread experiments were conducted at four different ambient pressures that ranged from 40 to 100 kpa. Flow velocity was set at 10 cm/s and oxygen concentration was then varied between different decreasing values in order to find the limiting condition at which flame spread will occur. For the Nomex HT90-40 the oxygen concentration tested ranged between 23%-35% while for the Cotton/Nylon/Nomex fabric blend ranged between 14%-25%. Experiments were performed with and without the presence of the external radiant heat flux to evaluate the influence of this variable. From the data collected, flame propagation was identified as when the flame propagated distances longer than 10 mm away from the igniter, to eliminate the influence of this element in the spread of the flame. Based on this criterion the limiting oxygen concentration for flame spread was determined as the condition at which all of the tests resulted in no propagation of the flame and at least one at the following higher oxygen concentration resulted in full or partial propagation of the flame. 426

5 Part II Fire Fig. 2(a) shows the experimental data collected for the Nomex HT90-40 using an external radiant heat flux of 5 kw/m 2. The solid line plotted in the figure represents the flame spread boundary for these particular conditions. Each point represents a different experimental condition tested; red triangles show the conditions were the flame propagated over the sample longer than the 10 mm criteria previously defined and purple circles show the conditions in which the test performed, after ignition was achieved, resulted in no propagation. During the experiments, higher oxygen concentrations were initially tested resulting in a bigger flame that propagated uniformly over the entire fabric sample, leaving behind a char residue. As the oxygen concentration was reduced the flame spread started transitioning to an irregularly shaped pattern and in some cases extinction occurred before the flame could propagate over the total length of the sample. In some of these experiments the flame propagation front would spread in a diagonal way that overcame the effect of the flow velocity and prevented its extinction. When the oxygen concentration approached the limiting values for propagation, after ignition was achieved the fabric would only burn around the igniter and no propagation was observed. Figure 2. Flame spread boundary for Nomex HT90-40 as a function of oxygen concentration and ambient pressure at an external radiant flux of (a) 5 kw/m 2 and (b) 0 kw/m 2. Similar set of experiments were performed using the fabric blend of 29% Cotton, 31% Nylon and 40% Nomex. Fig. 3 shows the experimental data collected, presenting also the boundary conditions for flame and no flame spread with and without the external radiant heat flux. In this case, the experimental results seem to follow a similar trend, however the lower content of Nomex (fire resistant material [12]) is clearly reflected in a lower oxygen concentration required for the flame to propagate at the same pressures. Overall, the trend is consistent and independent of the fuel that is being considered; as pressure is reduced an increase in the oxygen concentration is required for the flame to propagate over the material. Similar analysis can be done for the data obtained without the external heat flux, see Fig. 2(b) and Fig. 3(b). The results in this case follow the same trend; however, for both types of fabric, the flame spread boundary is higher than the boundary obtained using the external radiation. The external heat contributes to the heat provided by the flame to pyrolyze the fuel and compensates for 427

6 Proceedings of the Eighth International Seminar on Fire and Explosion Hazards (ISFEH8) the lower flame temperatures obtained at lower values of oxygen concentrations, allowing the flame to propagate at lower oxygen concentrations. Figure 3. Flame spread boundary for the fabric blend as a function of oxygen concentration and ambient pressure at an external radiant flux of (a) 5 kw/m 2 and (b) 0 kw/m 2. A similar analysis of the limiting conditions for flame spread can also be performed in terms of the ambient pressure and the O 2 partial pressure. Oxygen partial pressures are relevant in spacecraft applications because they relate to the normoxic cabin atmosphere and the flammability of the material [13]. Fig. 4 shows the boundary for flame and no flame spread determined for the Nomex and the fabric blend for the two different radiant heat fluxes: 5 kw/m 2 and 0 kw/m 2. When the experimental data is presented in terms of the O 2 partial pressure the results show a linear trend. This trend is consistent with data presented in previous works [5, 7, 14]. In the pressure range studied ( kpa), the required partial pressure of O 2 for ignition decreased almost linearly as the total pressure decreased, indicating that higher fire risk is expected. As it can be seen, at ambient pressures lower than 80 kpa, Nomex becomes flammable in normoxic equivalent environment since the boundary is below the normoxic line. The fabric blend studied in this work has proven to be flammable in normoxic equivalent environment over the entire range of pressures tested. From Fig. 4 it can also be noted that the limiting oxygen concentration for flame spread is altered when considering the external heat flux. Independent of the type of fabric or the ambient conditions, the change in the flame spread boundary limits when considering the external heat flux seems to be proportional to a change in ~1 kpa in the O 2 partial pressure. The linear character of the flame spread boundary when presented in terms of partial oxygen pressure is somewhat puzzling. As it is shown in the discussion section it is speculated that this behavior is related to the reactivity of the flame at its leading edge. A decrease in ambient pressure will require a proportional change in the partial pressure of oxygen in order for the mixture generated to be reactive. 428

7 Part II Fire Figure 4. Flame/ No Flame spread boundary for fire resistant fabrics as a function of O2 partial pressure and total ambient pressure. IDT simulations for experimental conditions tested (s 1 ). DISCUSSION The flame spread process over a thin solid fuel can be analysed in terms of the interactions between the studied environmental variables. When the flame is spreading over the fuel different physical, heat transfer and chemical mechanisms interact at the same time making the understanding of the flame behavior complicated enough to be modelled or fully understood. Several researchers have developed theoretical models of the flame spread process, the reader is referred to the works of Wichman [15], Delichatsios [16], Bhattacharjee [17] or Takahashi [18], for references on the subject. Here analysis developed by Fernandez-Pello [19] is used to understand part of the general phenomenology behind the process. This model describes flame spread rate as: ( ) ρscst s p To Cx V = l t ( + ) f h chem qf qe qrs U 1, (1) where q f, represents the heat flux provided by the flame, q e is the external radiant flux, q rs the reradiation from the solid, U the flow velocity, t chem the chemical time, ρ s and c s are the solid density and specific heat and s is the solid thickness. T p and T o represent the pyrolysis and initial temperatures of the solid. From Eq. (1) it can be noted that the addition of an external heat flux will have an impact in the heating process of the solid. Oxygen concentration will influence the heat provided by the flame to pyrolyze the fuel as it affects the flame temperature, allowing the flame to propagate faster at higher oxygen concentrations. As oxygen concentration is decreased the external radiant flux supplies the additional heat required for the flame to propagate. These effects will remain the same independent of the fuel considered. However, different fuels will have different absorptivity and that can result in a bigger or lower impact of the external heat flux applied. For downward flame spread over a thin fabric sample the main heat transfer mechanism to the solid is defined by the gas-phase heat conduction ahead of the flame. In this type of configuration the flame is located at a small distance from the solid acting as the main source of heat to pyrolyze the fuel. As the ambient pressure is reduced the heat transfer from the flame to the solid is affected since the 429

8 Proceedings of the Eighth International Seminar on Fire and Explosion Hazards (ISFEH8) thickness of the boundary layer that faces the flame is increased and the flame moves away from the solid. In mixed/forced flows, this stand-off distance described by the boundary layer thickness is proportional to the Reynolds number (Re) and the Grashof number (Gr). At the same time, when buoyancy is reduced the convective heat losses from the surface to the surroundings are also affected. This heat losses are proportional to the convective heat transfer coefficient which can be scaled by the same non-dimensional numbers (Re, Gr). Then their relation to changes in ambient pressure can be described by: ρu L Re = Re P, (2) µ ρ β gl T Gr Gr P µ =, (3) where ρβµ,,, glp,, and T are the flow density, volumetric thermal expansion coefficient, viscosity, gravity, characteristic length, pressure and the temperature difference between the solid and the surroundings. The interaction between the heat provided and the heat losses suggests that the heat provided by the flame to pyrolyze the fuel is partially compensated by a simultaneous reduction in the heat losses from the surface, making the total effect of pressure less significant when compared to other variables. When oxygen concentration and ambient pressure are simultaneously changed, if the ambient pressure is significantly reduced an increase in the oxygen concentration is required in order to sustain pyrolysis and avoid extinction of the flame. The higher flame temperatures that result from increasing the oxygen concentration compensate for the reduction in the heat transfer process derived from the reduction in pressure. Additional to the changes in the heat processes, the environmental variables also influence the chemical kinetics in the gas phase. Osorio [14] discussed the theory of flame propagation limits defined by the rate of chemical reactions in the gas phase. Starting from a two-component global reaction, he demonstrates that a critical characteristic chemical time, t chem can be maintained by an approximately linear relationship between pressure and oxygen partial pressure. This very fundamental analysis begins to explain the experimental observations, but does not provide exceptional agreement. Most notably, linear fits to the observed limits require a non-zero y-intercept. A conceptually similar, but more accurate approach to this analysis is to use a detailed chemical mechanism to describe the gas-phase reactions. To this end 0-D simulations of ethane-oxidizer mixtures were conducted using Cantera [20], with the AramcoMech 1.3 [21] chemical kinetic mechanism. Stoichiometric ethane-oxidizer mixtures initially at 1500 K were allowed to react and ignition delay time, IDT, was determined as the point at which fuel concentration reached 90% of its initial value. Total pressure and oxidizer oxygen concentration were varied to span the conditions used in the experiments. The criteria used to determine the IDT was chosen based on the exponential behaviour of the reaction rate and the concentration level at which the reaction was initiated. Taking IDT as the primary measure of reactivity (an analogue of t chem) we have graphed lines of constant IDT in terms of P and PO2 in Fig. 4 to compare the trend of the reactivity of the fuel under the different scenarios tested with the ones obtained experimentally. As with the flame spread boundaries, isolines of IDT are linear in terms of these variables but do not exhibit the problematic zero y-intercepts mentioned earlier. This agreement strongly supports the premise that limiting behavior is controlled by t chem. CONCLUSION Experimental data regarding the flammability limits of two thin fabric samples is provided under different environmental conditions. A single layer of Nomex HT90-40 and a fabric blend of Cotton/Nylon/Nomex are used as different types of fuel. 430

9 Part II Fire The relations between the flame spread process over the sample and the different variables of interest are presented in terms of their influence on the different heat transfer and chemical mechanisms. As ambient pressure is reduced a consequent increase in the oxygen concentration is required in order for the flame to be able to provide enough heat to pyrolyze the fuel and avoid extinction. The addition of an external heat source will provide the additional heat that is required to compensate the reduction in pressure, allowing the flame to propagate at lower oxygen concentrations. The simplified analysis used in this work supports an understanding of how some of the characteristics of the fire change in response to variations in environmental conditions. The data presented together with these relations provide a clearer picture of flammability of materials and what is to be expected under untested conditions. ACKNOWLEDGMENTS This paper is based upon work supported by NASA Grant NNX12AN67A. The authors would like to thank Sruthi Davuluri, Andrew Dimauro, Andrew Mikhaíl, Mitchell Heschke and Shiyu Jin for their assistance in conducting these experiments and constructing the apparatus used. REFERENCES 1. Lange, K. E., Perka, A. T., Duffield, B. E., and Jeng, F. F. Bounding the Spacecraft Design Space for Future Exploration Missions, NASA/CR , NASA Johnson Spaceflight Center, Sforza, P. M. Commercial Airplane Design Principles, Butterworth-Heinemann, Olson, S. L., Ruff, G. A., and Miller, F. J. Microgravity Flame Spread in Exploration Atmospheres: Pressure, Oxygen, and Velocity Effects on Opposed and Concurrent Flame Spread, SAE International Journal of Aerospace, 1(1): , Kleinhenz, J. E., and Tʼien, J. S. Combustion of Nomex III Fabric in Potential Space Habitat Atmospheres: Cyclic Flame Spread Phenomenon, Combustion Science and Technology, 179(10): , Hirsch, D. B., Williams, J. H., and Beeson, H. Pressure Effects on Oxygen Concentration Flammability Thresholds of Polymeric Materials for Aerospace Applications, Journal of Testing and Evaluation, 36(1): 69-72, Hirsch, D. B., Williams, J. H., Harper, S. A., Beeson, H., and Pedley, M. D. Oxygen Concentration Flammability Thresholds of Selected Aerospace Materials Considered for the Constellation Program, Second IAASS Conference: Space Safety in a Global Wolrd, Nakamura, Y., and Aoki, A. Irradiated Ignition of Solid Materials in Reduced Pressure Atmosphere with Various Oxygen Concentrations for Fire Safety in Space Habitats, Advances in Space Research, 41(5): , Quintiere, J., Harkleroad, M., and Hasemi, Y. Wall Flames and Implications for Upward Flame Spread, Combustion Science and Technology, 48(3-4): , Osorio, A. F., Fernandez-Pello, C., Urban, D. L., and Ruff, G. A. Limiting Conditions for Flame Spread in Fire Resistant Fabrics, Proceedings of the Combustion Institute, 34(2): , McAllister, S., Fernandez-Pello, C., Urban, D., and Ruff, G. Piloted Ignition Delay of PMMA in Space Exploration Atmospheres, Proceedings of the Combustion Institute, 32(2): , Anon. Flammability, Odor, Offgassing, and Compatibility Requirements and Test Procedures for Materials in Environments that Support Combustion, NASA STD 6001, Anon. Technical Guide for Nomex Brand Fiber, DuPont de Nemours & Company Technical Report H-52720, Olson, S. L., Beeson, H., and Fernandez-Pello, A. C. Applying Flammability Limit Probabilities and the Normoxic Upward Limiting Pressure Concept to NASA STD-6001 Test 1, Submitted to 44th International Conference on Environmental Systems, Tucson, Osorio, A. F. Effect of Environmental Variables on the Flammability of Fire Resistant Fabrics. Ph.D. thesis, Univeresity of California, Berkeley, Wichman, I. S. Theory of Opposed-Flow Flame Spread, Progress in Energy and Combustion Science, 18(6): ,

10 Proceedings of the Eighth International Seminar on Fire and Explosion Hazards (ISFEH8) 16. Delichatsios, M. M., and Delichatsios, M. A. Effects of Transient Pyrolysis on Wind-Assisted and Upward Flame Spread, Combustion and Flame, 89(1): 5-16, Bhattacharjee, S., Altenkirch, R. A., Srikantaiah, N., and Vedhanayagam, M. A Theoretical Description of Flame Spreading over Solid Combustibles in a Quiescent Environment at Zero Gravity, Combustion Science and Technology, 69(1-3): 1-15, Takahashi, S., Ebisawa, T., Bhattacharjee, S., and Ihara, T. Simplified Model for Predicting Difference between Flammability Limits of a Thin Material in Normal Gravity and Microgravity Environments, Proceedings of the Combustion Institute, 35(3): , Fernandez-Pello, A. C. The Solid Phase, Combustion Fundamentals of Fire, Cox G. (ed.), Academic Press, pp , Goodwin, D. An Open-Source, Extensible Software Suite for CVD Process Simulation, Chemical Vapor Deposition XVI and EUROCVD, Metcalfe, W. K., Burke, S. M., Ahmed, S. S., and Curran, H. J. A Hierarchical and Comparative Kinetic Modeling Study of C1-C2 Hydrocarbon and Oxygenated Fuels, International Journal of Chemical Kinetics, 45(10): ,