Solid combustion research in microgravity as a basis of fire safety in space

Available online at www.sciencedirect.com

Proceedings

ScienceDirect

of the

Combustion Institute

Proceedings of the Combustion Institute 35 (2015) 2487–2502

www.elsevier.com/locate/proci

Solid combustion research in microgravity as a basis of fire safety in space Osamu Fujita Division of Mechanical and Space Engineering, Hokkaido University, Kita 13 Nishi 8, Kita-ku, Sapporo 060-8628, Japan Available online 12 November 2014

Abstract This paper introduces fire safety standards for flammability evaluation of solid material intended for use in a spacecraft habitat. Two types of existing standards include material evaluation by pass/fail criteria corresponding to Test 1 of NASA STD 6001B and evaluation by a flammability index such as maximum oxygen concentration (MOC) corresponding to the improved Test 1. The advantage of the latter is the wide applicability of the MOC index to different atmospheres in spacecraft. Additionally, the limiting oxygen index (LOI) method is introduced as a potential alternative index for the evaluation using the improved Test 1 method. When criteria based on an index such as MOC or LOI are applied for material screening, the discrepancy of the index to the actual flammability limit in microgravity such as minimum limiting oxygen concentration (MLOC) is essential information for guaranteeing fire safety in space because material flammability can be higher in microgravity. In this paper, the existing research on the effects of significant parameters on material flammability in microgravity are introduced, and the difference between the limiting value in microgravity and the indices given by the standard test methods on the ground is discussed. Finally, on-going efforts to develop estimation methods of material flammability in microgravity according to normal gravity tests are summarized. Ó 2014 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

Keywords: Fire safety in space; Flame spread; Extinction limit; International standard; Electric wire

Abbreviations: LOC, Limiting oxygen concentration, Minimum oxygen concentration in which a spreading flame is sustained for a given condition (sample thickness, flow velocity, pressure, and so on) other than oxygen concentration; MLOC, Minimum limiting oxygen concentration, Minimum LOC in a wide range of external flow velocity including both opposed and concurrent flow fields under microgravity. This value is determined for a given material and sample thickness under the pressure condition in a spacecraft; MOC, Maximum oxygen concentration, Oxygen concentration in which all tested samples pass Test 1 [14] defined for upward flame spread; ULOI, Upward limiting oxygen index, Oxygen concentration in which 50% of the tested samples pass Test 1 [14] defined for upward flame spread; LOI, Limiting oxygen index, LOC under the condition specified by ISO 4589-2 [20] defined for downward flame spread. E-mail address: [email protected] http://dx.doi.org/10.1016/j.proci.2014.08.010 1540-7489/Ó 2014 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

2488

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

1. Introduction Fire safety is one of the most sensitive issues in manned space missions. At the very early stages of space flight development, oxygen concentration in spacecrafts was almost 100%, and combustible materials present in the spacecraft were flammable under such oxygen concentration, which resulted in high risk of fire. At present, the oxygen concentration in spacecraft is much less, such as 21% in the International Space Station (ISS), and materials intended for use in spacecraft are approved by NASA’s fire safety standard STD-6001B (originally NHB 8060.1C) [1,2]. The test method is essentially a pass/fail test in which upward flame spread over a material is observed under ambient conditions to which the material is exposed in the spacecraft. If the material passes the test, it is allowed into the spacecraft cabin. A drawback of such pass/fail tests is that their results are applicable only to the same conditions as those in the test, and it does not provide further information on the actual flammability of the material under the spacecraft conditions. If the test provides information on limiting flammability values of the material, the test results can be used as design data for other space activities. Even if the oxygen concentration in the spacecraft changes, the user does not have to repeat the same material evaluation tests for those conditions. Actually, oxygen concentration in pre-extravehicular activity (EVA) or in future spacecraft can be higher than that in the present ISS. Under the present fire safety protocol based on NASA STD 6001B, all tests for spacecraft material need to be retested under the worst conditions if the cabin air condition changes. Considering such disadvantages, NASA recently introduced the concepts of upward limiting oxygen index (ULOI) and maximum oxygen concentration (MOC) based on the test method of NASA STD 6001B. ULOI and MOC, which will be explained in detail in the next section, refer to the oxygen concentrations at which almost half of the samples of a material pass the test and all samples pass the test, respectively. Although ULOI and MOC are useful concepts it is unclear how these indices work in microgravity. In such an environment, natural convection does not occur, and very low flow velocity can be present due to the HVAC system or movement by the astronauts. Moreover, combustion products remain around the area where combustion took place. Such differences cause changes in material flammability. Therefore, the difference between the maximum flammability limit in microgravity and the indices such as ULOI or MOC must be determined so that the indices taken in 1G can be effective for evaluating the risk of fire in microgravity even when the designed air conditions in spacecraft change. Moreover, the

effects of many experimental parameters such as external flow velocity, material thickness, type of material, external heating and ambient pressure should also be considered. However, obtaining such results can be extremely difficult because quantitative data on the parameters are necessary for the use of indices as criteria to evaluate material fire safety in space. The author believes the combustion community is the appropriate group to address such issues, because a scientific understanding on material flammability is the crucial element to address them. In this paper, the latest research on the individual parameters is reviewed, and the necessary information to be added to the body of knowledge in this field is discussed. Although some excellent review papers already exist on microgravity combustion [3–5], fire safety technology in space [6–10] and fire research [11], those focusing on the fire safety standards for space applications under the aspect of combustion fundamentals are limited. The present study summarizes the standard test methods for material flammability evaluation and existing research relating to material flammability in microgravity. 2. Fire safety standard 2.1. Test 1 and Test 4 in NASA STD 6001B In this section, two standard test methods of material flammability evaluation, NASA-STD6001B and the LOI method, are introduced. As previously mentioned, differences in the indices

Fig. 1. Experimental configuration for the upward flame propagation test (Test 1) of NASA-STD-6001B [12]. The active area of the test sample is typically 5 cm  33 cm. A chemical igniter is placed at the bottom of the sample. (From“ A Research Plan for Fire Prevention, Detection, and Suppression in Crewed Exploration Systems” by A.G. Ruff, L.D. Urban, K.M. King [9]; reprinted by permission of the American Institute of Aeronautics and Astronautics, Inc.”).

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

given by the test methods in 1G and the actual flammability limit in microgravity are key parameters for guaranteeing fire safety in space. Before discussing these differences, however, we will explain the test methods. The most fundamental material flammability test in NASA STD 6001B is Test 1, which is also registered with the International Organization for Standardization (ISO) as ISO 14624-1 [9,12]. Figure 1 shows an example of Test 1. In this test, a vertical strip of a material 5 cm wide and 33 cm long, typically performed with the worst-case thickness, is set in a holder. The material is then ignited at the bottom end (resulting in concurrent upward flame propagation) by a chemical igniter in ambient conditions set to equal that of the worst-case environment to which the material will be exposed. The material fails the test if the flame spread exceeds 6 in (15 cm) or if the flaming debris drips onto a piece of K-10 paper placed 20 cm below the sample and ignites the paper. This test is conducted in 1G and is essentially a pass/fail test to evaluate selection or rejection of the material for spacecraft usage. For electric wires, a similar test method, known as Test 4, is detailed in NASA STD 6001B. This method is used for evaluating wire flammability under the worst-case environment potentially realized in spacecraft. This test method is also registered as ISO 14624-2 [13]. Figure 2 shows a schematic description of a sample holder for Test 4. American wire gauge (AWG) 20 wire of approximately 120 cm long with an active area of 30 cm is typically used as a test sample. The

Fig. 2. Experimental configuration for electric wire flammability (Test 4) of NASA-STD-6001B [13]. The standard sample gauge is American wire gauge (AWG) 20, and the length of the inclined part is typically 31 cm with an active length of 30 cm, which is set at 15° inclined to the vertical line. A chemical igniter is placed at the bottom of the sample. The internal wire temperature is set at 125 °C or at the maximum operating temperature of the wire.

2489

wire is inclined at 15° to the vertical line and is ignited at the bottom end by a chemical igniter in ambient condition equal to that of the worstcase environment to which the material will be exposed. The material fails the test if the flame spread exceeds 6 in (15 cm) at an internal wire temperature of 125 °C or at the maximum operating temperature of the wire or if flaming debris drips onto a piece of K-10 paper placed 20 cm below the sample and ignites the paper. Similar to Test 1, this test is also conducted in 1G and is essentially a pass/fail test used to evaluate selection or rejection of the material. 2.2. MOC and ULOI As previously mentioned, a drawback of Tests 1 and 4 is that they are evaluated merely as pass or fail, and their results are applicable only to the same conditions as those tested. Broader application of the results in determining the limiting values of flammability would yield a more meaningful outcome. Information on these limiting values can assign a margin of ambient conditions in spacecraft to the material flammability for assessing the fire risk. Such parameters would also allow better material choices when designing spacecraft for specific atmospheres. Under such considerations, NASA introduced the concept of MOC and ULOI [14] corresponding to Test 1, which is known as the improved test method. The MOC is defined [18] as the level of oxygen concentration where at least five samples passed the burning criteria and where at least one sample failed in the environment that contained 1 percent more oxygen by volume. ULOI is defined as the oxygen concentration in which approximately

Fig. 3. Definition of MOC and ULOI with an example of cumulative probability of material burning, failing Test 1 (From “Evaluating Material Flammability in Microgravity and Martian Gravity Compared to the NASA Standard Normal Gravity Test” by S.L. Olson, P.V. Ferkul; reprinted by permission of the American Institute of Aeronautics and Astronautics, Inc. [15]).

2490

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

50% of the group samples fail the test criteria for Test 1, as schematically shown in Fig. 3 [15]. Under these definitions, many trials for determining these indices have been conducted by NASA Johnson Space Center White Sands Test Facilities (JSC WSTF). MOC and ULOI for plastics and elastomers [16], commercially available fabrics and plastics, and composite material [17] are reported in these references. Furthermore, these researchers have investigated the effects of other experimental parameters such as pressure [18] and other selected parametric effects [19]. Also NASA Glenn Research Center (GRC) researchers have compared values obtained in both 1G and reduced gravity [15], which will be described in the following section. 2.3. LOI method The limiting oxygen concentration (LOC) for a given material corresponding to Test 1 or 4 in NASA STD 6001B varies considerably, which explains the need of the ULOI. Such variation is attributed to the following characteristics of the test method: (1) the judgment method of pass/fail is based on a 6 in (15 cm) flame spread in combination with the use of a chemical igniter. Heat input by the chemical igniter fluctuates depending on the individual run, which causes differences in flame spread. (2) Upward propagation leads to fluctuation in heat input from the flame to the unburned material because natural convection causes flickering of the flame. (3) When sample thickness is large enough Test 1 or 4 can often be an ignition test due to the type of igniter used rather than flammability test. The results of an ignition test do not mean the same thing as the results of a flammability test, when the difference in normal gravity and microgravity is considered. In comparison with MOC or ULOI by Test 1, the LOI method has advantages in its ease of determining limiting conditions by the nature of its test method. LOI is the minimum oxygen concentration for self-sustained flame spread over a solid material strip under the test method specified in ISO 4589-2 [20] or ASTM D2863 [21]. The method employs downward flame spread (i.e., opposed flame spread) followed by ignition to the top end of the fuel strip to determine the LOC for self-sustained flame spread. Here, sustained flame means flame spread for 180 s or 50 mm long over the fuel strip after the ignition. Ignition of the sample is achieved with a propane diffusion-type burner that supplies a sufficient amount of heat to the top end of the specimen, which reduces the chance of data dispersion caused by igniter differences. In the LOI method, the specimen, typically 80–150 mm long, 10 mm wide, and 4 mm thick, independent of its use in the spacecraft. A major concern with this evaluation method of material for space usage is that the

downward flame spread is generally not conservative in comparison with upward spread such as that used in Test 1, although the LOI method has an advantage of precise determination of the limiting value. However, if the discrepancy in the actual limiting oxygen concentration in microgravity to the LOI value is properly formulated, such as minimum limiting oxygen concentration (MLOC), the LOI method can be a promising method for evaluating the acceptability of material for space usage. The author believes the evaluation based on LOI method is more acceptable as a future standard because of its popularization as a test method and its ease for determination of limiting values. Now, a project attempting to prove the applicability of the LOI method for such a purpose is currently under way by an international research team and will be introduced in the final section of this paper.

3. Available data for material flammability in microgravity As previously mentioned, indices such as MOC and ULOI provide broader applications and additional information over that obtained from pass/ fail tests. However, challenges remain in verifying whether such indices are applicable to individual microgravity environments. In this section, several microgravity experiments are introduced, and differences in material flammability in 1G and microgravity are described, which are helpful for discussing the validity of the indices given by tests in 1G. Some experimental data on the flame extinction limit of a spreading flame over solid material in microgravity are available. Olson et al. [22] investigated flame spread over a laboratory cellulosic wipe and its extinction limit in a quiescent environment by using NASA 2.2s and 5.18s drop towers, and the results were compared with a downward propagating flame in 1G. The flame spread rate was determined as a function of surrounding O2 concentration and finally the oxygen concentration in which the flame spread terminated was obtained. They examined single- and double-thickness samples, both of which were thermally thin. In their experiments, the flame spread rate showed little difference in 1G and microgravity in the high oxygen concentration region, whereas at the low oxygen concentration near the extinction limit, the spread rate differed depending on gravity; the spread rate was lower in microgravity than that in 1G. Furthermore, the LOC was approximately 16.5% in 1G for both thicknesses. This value is much lower than that in quiescent microgravity, which was 21% for single thickness and 26% for double thickness. They also examined the LOC change as a function of the characteristic opposing flow velocity and

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

demonstrated that a MLOC of sustained flame appeared [23]. Altenkirch and Bhattacharjee, and their research groups [24–28] reported results of Space Shuttle experiments with thermally thin filter paper [24–26] and a thermally thick poly methylmethacrylate (PMMA) slab [27,28] in quiescent but elevated oxygen atmosphere under microgravity. The results showed that steady flame spread was realized over thin fuel, whereas unsteady spreading leading to extinction was always observed with thick fuel. Although a comparison of flammability in microgravity with that in 1G was not made, the data imply that a flame with thick fuel was not sustained in quiescent ambient but can be sustained under 1G even without forced flow because of the presence of buoyancy-induced flow. Flame extinction experiments onboard Mir Russian Space Station have also been conducted by an experimental apparatus known as Skorost [29,30]. Three types of material including PMMA, low-density polyethylene (PE), and Delrin were tested in very low concurrent flow with elevated oxygen concentration from 22.5% to 25.4%. Individual samples were 4.5 mm in diameter. When external flow velocity decreased, the flame spread rate decreased and flame extinction finally occurred in all tested material at 0.5 cm/s or less. In normal gravity, the flame was sustained under the same oxygen concentration even without forced convection. In microgravity, the experiments described above were conducted in quiescent or very low flow velocity. The material flammability in microgravity appeared to be lower than that in normal gravity because oxygen supply to the combustion zone is the controlling process in low or quiescent atmosphere, and a microgravity environment prevents further supply of oxidizer. However, other

Fig. 4. Extinction limits of polyethylene-insulated NiCr and Cu core wires as a function of external flow velocity in 1 g and lg [31]. The wire is set horizontally in 1G, and oxygen/nitrogen mixture is supplied opposing the flame spread direction.

2491

experiments show that depending on the conditions, the flammability limit could be extended even in microgravity. Figure 4 shows plots of LOC of a spreading flame over PE-insulated electric wire in opposed flow under normal gravity and microgravity [31]. The outer diameter and inner core diameter of the sample were 0.8 mm and 0.5 mm, respectively, and the core material of the sample wire was Cu or NiCr. The microgravity environment was provided by an airplane parabolic flight. The dotted and solid lines in Fig. 4 show the results in 1G and microgravity, respectively. In 1G, the sample was set horizontally. The open circle in the graph shows the case in which the flame did not go to extinction, which indicates that the limiting value of oxygen concentration in microgravity at 60 mm/s flow velocity is at a lower oxygen concentration than that at 1G. From Fig. 4, LOC in microgravity decreased monotonically with a decrease in the external flow velocity with both NiCr and Cu wires. Furthermore, the LOC of NiCr wire was lower than that of the Cu wire. A comparison of 1G and microgravity data reveals that LOC in microgravity is 2% lower than that in 1G at each external flow velocity. Recently, Osorio et al. [32] conducted similar experiments with ethylene-tetrafluoroethylene (ETFE)-insulated copper core wire in 12 cm/s opposed flow under microgravity, which revealed that the LOC became approximately 6% less than that in 1G. This result indicates that a material can be more flammable in microgravity than that in 1G. Zhang et al. [33] and Honda et al. [34] have investigated the effect of dilution gas in an oxidizer without forced flow on the flame spread rate over a thin cellulose sheet in different O2 concentrations; LOC was also determined from the test. In cases of CO2 and SF6 being used as dilution gases, LOC decreased in microgravity approximately 3% and 9% less, respectively, than that in normal gravity, whereas He, N2, and Ar dilution led to higher LOC in microgravity. From the perspective of fire safety, it is important the LOC may be lower in microgravity depending on the dilution gas because inert gasses such as CO2 may be used as a fire extinguishing medium. Recently, Olson et al. [15,35] determined the LOC of several thin materials under lunar gravity [35] and Mars gravity conditions [15] utilizing a centrifuge system installed to the capsule of NASA 5.18 s drop tower following the NASASTD-6001B Test 1, and compared the results with those of MOC and ULOI obtained by Test 1 in 1G. Furthermore, they obtained the LOC in concurrent forced flow of 30 cm/s under microgravity, assuming that the condition is comparable with the flow velocity induced in 1G. The results showed that LOC in reduced gravity is much less than MOC or ULOI after the tests of three commercially available material, Normex HT90-40,

2492

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

Ultem 1000, and Mylar G [15,35]. For example, the results with Normex HT90-40 of 12-mm-thick fire retardant aromatic nylon showed that LOC in both lunar gravity and Mars gravity was 2.35% less than the average value of MOC and ULOI based on Test 1 in normal gravity. Interestingly, microgravity tests with 30 cm/s concurrent flow gave only a 0.3% difference, which is still less than that in 1G, suggesting partial gravity results in higher material flammability than microgravity. For the case with Mylar G, a 5-mm-thick plastic film created from polyethylene terephthalate gave the larger difference of 5.75% less in lunar gravity than that in 1G and 4.1% less than that in microgravity. Feier et al. [36] studied the upward and downward flame spread in micro and partial gravity by experiments accompanied with numerical calculation. They determined that the total limiting pressure under 21 vol% O2 in reduced gravity for downward spreading is lower than that in 1G. Interestingly, a gravity level exists in which the minimum limiting pressure appears, such as MLOC. The limiting pressure was determined to be lower for upward spread (concurrent) than that for downward (opposed) spread. Apparently, the material flammability range is extended in microgravity or in reduced gravity. As discussed in the second section of this paper, oxygen indices such as those given by Test 1 have more advantages over those given by conventional test methods such as a pass/fail test. However, as shown in this section, the indices are not always on the conservative side in reduced gravity; therefore, a clarification of the differences in flammability between 1G and microgravity is required. Ideally, it is desired that a quantitative formulation used to estimate MLOC in microgravity is given as a function of the index given by the standard test method on the ground. This presents a high challenge because many parameters affect the LOC in microgravity, such as external flow velocity, material thickness, external heating, flow direction, and material characteristics including reaction parameters and thermo– physical properties. In the following section, therefore, the effect of individual parameters on the LOC for a spreading flame over solid material will be introduced, and necessary information to be added to the existing knowledge base will be presented.

4. Effective parameters on extinction limit of a spreading flame over solid material 4.1. Effect of external flow velocity Here the microgravity environment is characterized by the flow field surrounding the solid fuel. The flow field is not accelerated by buoyancy and

is instead controlled by only an artificially produced flow such as the spacecraft HVAC. The flow field in normal gravity, however, is a result of superposition of forced flow and buoyancy-induced flow. Thus, understanding the effect of flow field on the flammability limit is essential for quantifying the differences between microgravity and normal gravity. It is widely accepted that the extinction of a spreading flame over solid material is caused by two mechanisms: blowoff extinction in high flow velocity conditions and quenching extinction in low flow velocity conditions. T’ien and his coworkers [37-40] studied the extinction phenomenon of stagnation flame above condensed fuel based on numerical calculations. They demonstrated the importance of kinetics in a high stretch regime and the importance of radiative heat loss from the condensed phase surface in the low stretch region, suggesting the existence of two extinction branches, radiative extinction and blowoff extinction. Thus, MLOC, which they referred to as the fundamental limit, can exist at the flow condition in which two branch curves cross each other. Their research provided important insight into the extinction of a spreading flame over solid material, although the experimental proof of the presence of MLOC was not available at that time because of the difficulty in attaining very low stretch conditions under the gravitational field. The importance of radiative heat loss in low stretch flame, including heat loss from condensed surface, has also been discussed by other researchers [41–43]. The blowoff extinction of a spreading flame over solid fuel was extensively studied by Fernandez-Pello et al. [44–46] for both thermally thick PMMA sheets and thermally thin filter paper. Blowoff extinction is explained on the basis of Da, the ratio of residence time (sp ) and characteristic reaction time (sc ). When Da (=sp /sc ) < 1 (or a critical value), the flame existing above the solid surface is blown off. They provided the definition of Da and experimentally showed the relationship between Da and the non-dimensional flame spread rate. Moreover, they demonstrated that the extinction limit in a high stretch region is controlled by the limit of Da. The effect of hypergravity on the downward flame spreading over index cards in elevated pressure and that spreading with force-opposed flow mixed with natural convection and their extinction limit was evaluated by Altenkirch et al. [47,48], and a discussion based on Da was conducted, as was done for the high flow velocity experiments introduced above. As previously mentioned, Olson et al. [22,23] determined the extinction limit of a spreading flame over thin cellulose paper opposed to an external flow by short-term microgravity experiments, and plotted the results with the data in high flow velocity conditions obtained in previous

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

Fig. 5. Map of controlling mechanism of flame spread as a function of oxygen percentage and characteristic relative velocity for thermally thin fuel [23].

Fig. 6. Flame spread rate over thin cellulose paper in opposed flow [23].

research [44,47]. As suggested by T’ien, et al. [33–36] a stagnation flame shows two branches, blowoff branch and quenching branch, and MLOC (noted as low oxygen limit in Fig. 5) exists in a certain characteristic flow velocity, as shown in Fig. 5. According to the flame spread rate as a function of characteristic relative velocity, as shown in Fig. 6, the sustained flame spread region is separated into three regions (Fig. 5). These areas include Region I: Gas-phase-conductionlimited flame spread region, Region II: Residence-time-limited flame spread region, and Region III: Oxygen-transport-limited flame spread region. Region I is a heat-transfer-controlled region [49,50], resulting in a constant spread rate for thin fuel independent of external flow velocity. The flame spread rate decreases with an increase in flow velocity in Region II, and increases with an increase in flow velocity in Region III. In the discussion on fire safety standards, the basic concept includes choosing the material in which MLOC is higher than that of the spacecraft environment. For such a purpose, determining

2493

the relation between the indices given by the standard test method in 1G and MLOC, defined as the minimum value in Fig. 5, is essential. The quenching branch has been extensively studied by Altenkirch and Bhattacharjee, and their group [24–28,51,52] for thick and thin fuel in combination with Space Shuttle experiments in a quiescent environment under different pressure and oxygen concentrations. They reported the importance of radiative heat loss under a quiescent environment even when the oxygen concentration was quite high, up to 70% in 1 atm. Furthermore, Bhattacharjee and his coworkers [53,54] introduced the “radiation number” derived from the scale analysis for a spreading flame over solid material in opposed flow. Moreover, they showed that the estimated limiting condition corresponding to the quenching extinction based on the radiation number agreed well with the trend given by microgravity experiments with the 4.5 s drop tower. Takahashi et al. [55–57] attained further microgravity experiments for thin PMMA sheets in low external flow velocity and showed results for different fuel thicknesses of thin films [55] and for different diluent gases [56,57]. Then, they demonstrated that the results are effectively explained by the radiation number. From the perspective of fire safety standards, the quantitative estimation of MLOC is desired. Takahashi et al. [58] made a further attempt to estimate MLOC for thermally thin PMMA films in opposed flow based on Da and the radiation number. According to the scale analysis, Da and radiation number (Rrad) are given as Eqs. (1) and (2), respectively. The physical meaning of Eq. (2) is the ratio of radiation heat loss to the heat input from the gas phase in the preheat length given by ag/Vr. Da  tres =tchem
 Da ¼ B1 ag =V 2r qg Y O Aexp E=ðRT f Þ Rrad ¼ tgsc =tser ¼ Q_ rad =Q_ gs Rrad ¼ B2

es ð1aabs ÞrðT 4m T 41 Þ qg cg V r ðT f T m Þ

ð1Þ


ð2Þ

;

where tres is the residence time in the gas phase (ag =V 2r ), tchem is the characteristic reaction time, ag is the thermal diffusivity of gas at Tm, Vr is the absolute spread rate (=Vg + Vf), Vg is the external flow velocity, Vf is the flame spread rate, qg is the gas density at Tm,Yo is the oxygen mass fraction, A is the pre-exponential factor, E is the activation energy, R is the gas constant, Tf is the flame temperature, tgsc is the characteristic time of heat conduction from gas to solid surface, tser is the characteristic time of radiation heat transfer from the solid surface to the environment, Q_ rad is the radiation heat loss from the solid surface of preheat length, Q_ gs is the conductive heat transfer from gas phase to solid surface of preheat length,

2494

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

es is the emissivity of the solid surface, aabs is the absorption coefficient of the gas, r is the Stefan– Boltzmann constant, Tm is the pyrolysis temperature of the solid fuel, T1 is the surrounding temperature, and cg is the specific heat of surrounding gas. As shown in Eqs. (1) and (2), the dimensional relation is clear, although the proportional factors B1 and B2 are unknown. These values are given by assuming that Da is unity at the oxygen concentration in which blowoff extinction occurs and that Rrad is unity at the oxygen concentration in which radiative extinction occurs [58]. The oxygen concentrations were determined by downward flame spread in 1G and opposed flame spread in microgravity; the resulting values of B1 and B2 were 0.131 and 2.67, respectively. Then, the stable flame region was determined by substituting both values into Eqs. (1) and (2) for different gas balances, as shown in Fig. 7. This example uses methods for quantitative estimation of LOC of sustained flame spread by combination of the scale analysis and microgravity experiments, which are essential information for fire safety in space. In Fig. 7, the effect of dilution gas is also considered. The radiation effect is important through the absorption coefficient aabs, and as a result, both quenching extinction limit and blowoff extinction limit extend toward lower flow velocity with CO2 dilution. MLOC for a spreading flame in opposed flow is also given by numerical calculations based on the mathematical model developed for a spreading flame [59,60] for thermally thin fuel [36,61– 65] and thermally thick fuel [66,67]. These studies include the effects of side edge, downward spreading in 1G corresponding to the LOI method, and comparison with concurrent flow (or with upward propagation in 1G). Numerical calculation also gives the value of MLOC in terms of external flow velocity. These approaches provide MLOC quantitatively in microgravity according to the value

given for tests on the ground, such as the LOI method, if the verification of the numerical calculation is performed by well-defined microgravity experiments. Two branches appear again, one in the low flow velocity region (quenching branch) and other in the high flow velocity region (blowoff branch), and MLOC appears in the point where the two curves merge, as suggested in a previous study for stagnation flame above a solid fuel plate [38]. 4.2. Effect of external flow direction The effect of flow direction is also important because concurrent flow usually gives higher flammability values than opposed flow, and MLOC given in concurrent flow should provide conservative information for judging material acceptability. Therefore, it is important to understand how material flammability differs in opposed and concurrent flows in microgravity. Experimental studies on the comparison of flame spread phenomena in concurrent flow (upward spreading in 1G) and opposed flow (downward spreading in 1G) are limited because research on the flame structure is generally conducted in only one flow direction. A comparison of LOC of downward and upward spreading over thin paper sheets in normal gravity has been previously conducted [45,68]. As expected, LOC in concurrent flow is lower than that in opposed spreading, as shown in Fig. 8. The difference is attributed to the preheating effect of the unburned fuel by combustion gas flowing over the sample. Unique experiments on the comparison of concurrent and opposed spreading in microgravity include ignition at the center of the thin filter paper and subsequent flame spreading phenomena [69–73], which were conducted by the Japan Microgravity Centre (JAMIC) at the 10 s drop tower in Japan and on the Space Shuttle. An important finding was that the flame spread rate

Fig. 7. Regime map with various diluent gases. The mark of GC denotes ground condition. Note that with CO2 gas, the ground condition is out of the sustained flame region [58].

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

2495

Fig. 8. Flammability (no propagation) limits of upward and downward spread of flames over thin paper sheets. Sample width is 1 cm [45,68].

of the downstream flame is less than that of the upstream flame because downstream fuel is in the shadow of the combustion gas produced by the upstream-opposed spreading flame. In such cases, the limiting condition of the upstream flame spread is more important than that of the downstream flame spread. Takahashi et al. [55] compared the spread rates for thin PMMA sheets between low-velocityopposed and concurrent cases. They demonstrated that the spread rate can be larger in the opposed case in the low flow velocity condition. Furthermore, at O2 = 21%, extinction occurred at flow velocity close to zero, and the absolute flow velocity of flame extinction was larger in the concurrent case. These results indicate that extinction occurs more easily in the concurrent case. Olson and Miller [74] further studied the flow direction effect on thin cellulose paper and demonstrated that the spread rate in the concurrent case linearly increases with flow velocity, whereas for opposed flow, spread rate increases rapidly with gas flow velocity in a nonlinear manner before reaching a constant spread rate of the thermal regime. As a result, low flow velocity of 5–20 cm/s, similar to ventilation flow velocities in spacecraft, gives a faster spread rate in the opposed flame front in microgravity. In a recent study [67], the case in which the extinction limit of downward spreading is lower than that of upward spreading was introduced. This behavior is caused by heat loss from the charred region, which is longer in the concurrent case because of preheating caused by combustion gas. These experiments indicate that opposed flame can be more flammable, although the typical situation is higher flammability with the concurrent

mode. Therefore, the flammability results are nonconclusive. T’ien’s group conducted extensive numerical calculation for concurrent flow [36,59,61,75–79] as well as opposed flow. Figure 9 shows an example of the extinction limit curve for both opposed and concurrent flow. MLOC, noted as fundamental limit in their articles, appeared in both flow direction cases. In the figure, Xe denotes the distance from the front edge of the sample to the flame position, showing the effect of boundary layer growth. Also, in the figure, the LOC is lower in the concurrent case, whereas that in the opposed flow is lower in very low flow velocity condition, at less than 2 cm/s [61]. According to the figure, actual MLOC

Fig. 9. Comparison of extinction boundary of opposed flow flame spread with concurrent flow flame spread [61].

2496

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

appeared in the concurrent case. They also studied the difference of forced concurrent flow in microgravity and upward spreading in normal gravity [77] and the chemical kinetic effect on the spreading and extinction limit [78,79]. Tizon et al. [80] studied the effect of angle of external flow relative to the axis of the rod-like sample and demonstrated the importance of radiation heat loss even with the existence of the angle . 4.3. Effect of sample thickness Fuel thickness is an additional important factor for controlling the extinction limit. When the fuel is thermally thick, heat conduction into solid fuel takes a major role in the heat balance of the preheat region [49], as shown by the detailed temperature measurement reported previously [81], and the heat loss term is to be added in the numerator in Eq. (2). Thus, thicker fuel relates to higher LOC. Therefore, the thickness effect should be considered when estimating MLOC of material intended for use in spacecraft. Test results of thick material in microgravity are limited because its combustion time scale is fairly long, and tests are possible only in long-term microgravity experiments. Representative microgravity tests include Space Shuttle experiments with thick PMMA samples in a quiescent environment [27,28,54,55], as introduced in Section 3. In these experiments, pressures of 1–2 atm and oxygen concentration up to 50% were examined, and numerical calculations were conducted. The results showed that the thick sample in quiescent atmosphere went into extinction in 50% oxygen, while numerical calculation showed extinction even in pure oxygen when the thickness was sufficient. Thus, they demonstrated the extinction limit with different thicknesses in opposed airflow with a strong emphasis on the radiation heat loss effect. Olson et al. [82] conducted sounding rocket experiments with thick PMMA samples to investigate the effect of very low external flow velocity of less than 5 cm/s, and observed the effect of external radiation over 0–2 W/cm2. They determined that the flame in very low flow velocity is more robust than that estimated by the two-dimensional model because of the three-dimensional effect. Kumar and T’ien [64,66] studied the flammability limit of thick fuel identical to the LOI standard sample, and determined that the LOC did not differ significantly in such thicknesses in opposed flow with the thin sample because the wake flame mode, that is the mode flame is stabilized in the wake of downstream end of a thick fuel, gave the most flammable condition. Recently, Hosogai et al. [83,84] reported the effect of sample thickness on LOC following the LOI test method, and demonstrated the increasing trend of the LOC with an increase in the sample thickness. The above information indicates that

the inclusion of the sample thickness effect on the extinction limit is required to give proper estimation of MLOC based on the tests in 1G. 4.4. Effect of sample geometry Typical case, where flame spread is strongly affected by sample geometry, is wire insulation combustion. Screening wire insulation material for fire safety in spacecraft is also an important issue for safety in space because a main source of fire in spacecraft is the electric circuit or wire harness as a result of overloading or short circuiting. Therefore, as indicated in Section 2, Test 4, specialized for wire flammability evaluation, is described in NASA STD-6001B. Again, if we introduce the concept of limiting value, rather than just a pass/fail score, and compare it with the spacecraft atmosphere, we need to determine the differences in limiting values given by the standard test in 1G and MLOC. Nonetheless, the determination of wire insulation flammability in microgravity and testing of wire insulation combustion in microgravity is limited. Greenberg et al. [85,86] conducted flammability tests of PE-coated NiCr core wire in opposed and concurrent flows in a Space Shuttle experiment in a project known as WIF. However, the number of test runs was limited and insufficient for discussing the mechanism of flame spread over wire insulation. Other microgravity tests were conducted by Hirata et al. [87] with ETFE wire flame spreading in opposed flow in a 4.5 s drop tower and by the author’s group for ETFE- and PE-insulated wire flame spread in quiescent environment [88,89] in a 10 s drop tower, flame spread in opposed flow [90-93] in 4.5 and 10 s drop towers, extinction experiments in opposed flow with parabolic flight [31,32,94], and wire ignition by overloading in 2.7 and 4.5 s drop towers [95–98]. Wire combustion is characterized by its geometry and the presence of a conductor in the wire. It is cylindrical and its surface curvature enhances combustion in the low flow velocity region [89,90,99]. Heat conduction through the inner core also has a significant effect on flame spread and extinction phenomena, as discussed in previous research [31,93,94,100–105]. We can apply a similar concept to Eqs. (1) and (2) for wire insulation combustion to estimate the LOC for sustained combustion, considering the effects of geometry and conductor presence. The equations are given as (3) and (4) for the blowoff extinction limit and quenching extinction limit [31]:
 Da  tres =tchem  ag =V 2r qg Y O Aexp E=ðRT f Þ ð3Þ
Rloss ¼ Q_ rad þ Q_ sc =Q_ gs rs es rðT 4 T 4 Þþks ðT m T c Þ= lnðrs =rc Þ  k ðTm T1Þ= ln 1þa =ðr V Þ ; f g s rg g f m

ð4Þ

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

where tres is the residence time ð¼ ag =V 2r Þ, tchem is the characteristic reaction time, qg is the density of ambient gas, Yo is the mass fraction of the ambient gas, A is the pre-exponential factor, E is the activation energy, R is the gas constant, rc is the wire outer diameter, rc is the wire core diameter, es is the emissivity of the wire insulation, r is the Stefan–Boltzmann constant, Tm is the insulation gasification temperature, T1 is the ambient temperature, ks is the thermal conductivity of the insulation, Tc is the wire core temperature, kg is the thermal conductivity of the ambient gas, and Tf is the flame temperature. In Eq. (4), the term of heat loss from the preheat zone insulation to the inner core, Q_ sc appears in addition to the radiation heat loss term, Q_ rad , whereas Eq. (3) is the same as Eq. (1). If the fuel is thermally thick, heat conduction into the fuel depth exists, whereas heat loss to the inner conductor exists for the case of wire combustion. The effect of geometry also appears as a logarithmic effect [89,90,99] in the equation. If we assume that quenching extinction occurs at Rloss = 1, as in the previous discussion, we would find a factor in Eq. (4), such as B2 in Eq. (2), once experimental data in microgravity are available, because local flow velocity at the preheat zone is well defined in microgravity. Then, we can draw a quenching extinction curve corresponding to Eq. (4) for wire insulation combustion. In the same way, we can determine a factor for Eq. (3) for the blowoff extinction limit in the higher flow velocity case and draw the curve. The curves corresponding to Eqs. (3) and (4) merge with each other, and at the cross point of the curves, MLOC appears. Then, we can compare the value with the index given in 1G following the standard test method. The important information is the difference between LOI or MOC and MLOC, which provides a margin to be considered for space usage. The method above based on Eqs. (3) and (4) is a practical approach for estimating the difference. When Eq. (4) is considered, the effect of curvature and core insulation’s presence is clear. When the outer diameter of the wire, , decreases, the logarithmic part of the denominator in Eq. (4) decreases, as does Rloss. When the conductor temperature is low, the second term in the numerator of Eq. (4) becomes negative and is counted as heat loss. However, when the inner core temperature is higher than the insulation pyrolysis temperature, the value is positive and contributes to supplying heat to the preheat zone. Because opposed or concurrent flow is given to the wire in microgravity, the flame extends along the wire and heats the inner core, which results in extension of the flammable limit. When external flow is supplied with inclination to the wire axis, it cools down the core at the burnout region, which results in lower flammability. Therefore, maximum flammability appears when the external flow direction is

2497

parallel to the wire, as suggested in the recent study [100]. When discussing wire combustion, the electric and magnetic field effects must be considered. When electric potential is applied to the wire, an electric field is formed around the wire. On the contrary, an electric current going through the wire causes a magnetic field. Although combustion phenomena can be affected by the electric field [106] and the magnetic field [107], such field effects around the electrical wire are relatively small in comparison with gravitational field under ordinary usage conditions. In microgravity, however, buoyancy-induced flow disappears, and other forces such as the electric field effect [108] or the magnetic field effect [109,110] can be significant, and the resulting flammability be affected. Furthermore, Kim et al. [111] demonstrated that the electric field effect is significant in its flame spreading phenomena depending on the voltage applied to the wire. 4.5. Effect of ambient pressure Pressure is another dominant factor for control material flammability, which may vary in the spacecraft habitat [112]. As introduced in the previous section, several reports include the effect of pressure on solid material flammability relative to fire safety in space. Frey et al. [113] studied the effect of pressure on the extinction limit of horizontal and downward spreading flame over thin paper sheets, and showed that the limiting oxygen mole fraction decreases with an increase in pressure for both spreading directions in 1G. Ramachandra et al. [25] and Bhattacharjee et al. [26] showed experimental results attained in the Space Shuttle under quiescent atmosphere, including an increasing trend in the flame spread rate with increased pressure. West et al. [27] showed the results on thick PMMA plates in which the spread rate was unsteady and the pressure effect was unclear for the tested pressure range. Kikuchi et al. [88] included the pressure effect in their experimental work in the 10 s drop tower, which showed that the extinction limit under microgravity for flame spreading over ETFE-coated wire in quiescent atmosphere is larger than that in normal gravity. Olson et al. [75,114] gave the dependency of the spread rate on pressure with charring (root square dependency on pressure) and non-charring (weak dependency on pressure) of thin sheet material in microgravity. As previously mentioned, Hirsch et al. included the pressure effect in their determination of ULOI and MOC under NASA STD 6001B. Nakamura et al. [115–117] studied the effect of pressure on flame spreading over polyethylene wire insulation with and without external flow velocity, and determined that the spread rate increases with a decrease in pressure in the tested range where flame spreading is

2498

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

controlled by heat supply to the preheat length, while extinction limit shown in Fig. 8 is affected by decrease in chemical reaction rate with decrease in pressure. They also reported that the decreased pressure tests can be used to simulate reduced gravity tests as long as the pressure effect on the chemical reaction is negligible. McAllister et al. [118,119] and Fereres et al. [120] also studied piloted ignition in reduced pressure considering similarity to a reduced gravity environment, and showed higher flammability under reduced pressure conditions. 4.6. Effect of external radiation As emphasized in the previous section, surface radiation heat loss from the preheat zone to ambient air has a dominant effect on the extinction phenomena in low flow velocity. Therefore, processes affecting the surface radiation heat loss alter the quenching extinction limit. Re-absorption of radiation heat in ambient gas has a significant effect on the extinction limit [34,56–58,121], which is typically re-absorption by CO2, H2O, and soot The effect can be considered in Eq. (2) through the absorption coefficient of ambient gas [58]. External radiation directly affects the heat balance of the preheat zone and extends the extinction limit [32,122,123]. When a flame is formed along the solid material, heat supply from the flame to the preheat zone by radiation is significant when the flame dimension increases, to some extent. It is widely accepted that soot formation increases in microgravity [124–131] and enhanced radiation heat transfer from the flame to the preheat zone increases the flame spread rate and extends the flammability limit. External radiation effect in material flammability test is also found in Ref.[132,133]. 5. Challenges to fire safety standard improvements and other substitute methods 5.1. Potential method to substitute microgravity environment in 1G tests Determination of MLOC of the material intended for use in space is essential to ensure fire safety in space. Once the value is known, we can judge the acceptance of the material for space use by comparing the value with oxygen concentration in the spacecraft. A potential approach is to develop a method for simulating a microgravity environment under 1G. Olson et al. [134,135] proposed the use of stagnation flow to determine the flame structure and extinction characteristics in a low-stretch condition, which is a promising method for determining MLOC in microgravity. However, further research is required for application to different types of practical material. The

use of a reduced pressure environment, as introduced in Section 4.5, is an additional potential method for simulating microgravity or partial gravity environments. Pressure decrease results in density decrease and reduction of buoyancy force in proportion to the square of density. If the effect of pressure on a chemical reaction is negligible in the overall process, such a method is a promising substitute for microgravity experiments. The narrow channel method [136,137] is a novel method used to simulate microgravity phenomena by reducing the characteristic height of the flow duct. Buoyancy force is proportional to the cube of the characteristic length; its reduction is highly effective for suppressing the buoyancy force. Similar phenomena have successfully reproduced the fingering structure unique in low-speed oxidizer flow under microgravity, which was reported in previous Space Shuttle experiments [138]. The results also indicate that the narrow channel method quantitatively captures the essential features of the microgravity tests for thin fuels in opposed flow. The method is affected by heat loss to the top and bottom walls because the distance between the walls is very small. Thus, improvement in compensating for the heat loss effect is a challenge for this method. 5.2. Large-Scale Spacecraft Fire Safety Experiments (Saffire) Two international projects on fire safety in space are currently under way. The first is “Large-Scale Spacecraft Fire Safety Experiments (Saffire)” [139,140]. The scale effect on material flammability is very important because the dominant heat transfer mechanism changes depending on the scale of fire, such as a change from convective heat transfer to radiative heat transfer to the unburned region. However, such large-scale experiments in microgravity have not been conducted previously because of the restriction in size and safety of standard microgravity experiments. In this project, a series of flight experiments, known as Saffire 1–3, will be conducted in the unmanned Orbital Science Corporation Cygnus vehicle after it has undocked from the ISS. These experiments will enable combustion of large-scale samples. Saffire 1 and 3 will burn large samples, while Saffire 2 will burn nine small samples. This project is led by NASA GRC and the European Space Agency (ESA) international topical team and involves 14 international members. The final objective is to obtain data to validate the modeling of spacecraft fire response scenarios and to demonstrate the validity of NASA’s normal gravity material flammability screening test (NASA STD 6001 Test 1) for low-gravity conditions. Further details have been reported elsewhere [139,140].

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

5.3. Solid material flammability tests in Japanese Experiment Module in ISS Another project is “Flammability Limits at Reduced Gravity (FLARE),” which aims to formulate the relation between MLOC and LOC obtained by the 1G standard method. In particular, the oxygen concentration determined by the LOI method (LOI itself) will be used as base data to estimate MLOC in microgravity. Furthermore, the recommendation of an alternative fire safety standard for screening material intended for usage in spacecraft will be proposed on the basis of the formula developed in the project. As discussed in the second section, the present NASA STD 6001 includes some uncertainties in determining ULOI or MOC resulting from the nature of the test method. For example, upward propagation causes data scattering for the extinction limit in comparison with downward spreading adopted in the LOI method. Additionally, the ignition method of NASA STD 6001 causes a difference in the initial condition of the sample preheating or a relative difference in heat capacity of the sample with the same igniter, which may affect differences in ULOI or MOC even with the same material. Such situations introduce difficulties for estimating MLOC in microgravity to be compared with ambient oxygen concentration in a spacecraft. This is because scattering of 1G data to be included in the developed MLOC formula causes difficulty in obtaining a fixed MLOC. In this regard, the LOI method as a basis for MLOC has an advantage, which is the reason for development of the formula based on the LOI in the FLARE project. On the contrary, the LOI method may not give a conservative information because opposed flame spreading is generally less flammable than the concurrent flame spread. Therefore, determining MLOC in microgravity based on LOI is a difficult challenge because MLOC appears in concurrent flow and the LOI is given in opposed flow. To achieve the goal of the project, an understanding of the science of combustion, such as that explained in the present paper, is crucial. Moreover, continuous discussion in the combustion community will take an essential role, particularly in creating the formula mentioned above. This international project is led by the Japan Aerospace Exploration Agency (JAXA) in cooperation with NASA; ESA; Centre national d’e´tudes spatiales (CNES); and university researchers from Japan, the United States, and Europe. This project includes experiments in the Japanese Experimental Module in the ISS. The hardware used for the space experiments is being developed for the JAXA project, known as “Solid Combustion,” which investigates the fundamentals of ignition and flame spreading of laboratory solid material [141].

2499

6. Conclusions This paper discussed improvements in fire safety standards for space and the applicable scientific evidence behind such improvements. The advantages of the material screening methods based on a flammability index, such as MOC or LOI, were emphasized in comparison with a pass/fail test based on NASA STD 6001B. A major advantage is the wide applicability of such limiting data to the varied cabin air conditions. Material flammability in microgravity, however, can be higher than that in normal gravity according to the existing data; therefore, quantification of the difference between microgravity and normal gravity is required to utilize the flammability index as the material screening method. For such purposes, MLOC in microgravity is very important because a comparison between MLOC and a flammability index such as MOC or LOI can give quantitative differences. Many parameters affect material flammability, such as external flow velocity, flow direction, material thickness, presence of a conductor in the material, external heating and ambient pressure; thus, a scientific understanding of the effects of such parameters is required for estimating MLOC. Finally, on-going efforts to develop estimation methods of material flammability in microgravity based on normal gravity tests were summarized in this study. Acknowledgments The author appreciates Japan Aerospace Exploration Agency (JAXA) for its support in conducting a part of this research under the second phase utilization of JEM/ISS, titled “Quantitative Description of Gravity Impact on Solid Material Flammability as a basis of Fire Safety in Space” and the third phase utilization of JEM/ISS, titled “Evaluation of gravity impact on combustion phenomenon of solid material towards higher fire safety.” The author greatly appreciates all team members of Solid Combustion and FLARE for their helpful suggestions. References [1] NASA NHB 8060.1C, Flammability, Odor, Offgassing, and Compatibility Requirements and Test Procedures for Materials in Environments that Support Combustion, 1991. [2] NASA STD 6001B, Flammability, Offgassing, and Compatibility Requirements and Test Procedures, 2011. [3] C.K. Law, G.M. Faeth, Prog. Energy Combust. Sci. 20 (1994) 65–113. [4] M. Kono, K. Ito, T. Niioka, T. Kadota, J. Sato, Twenty-Sixth Symposium (International)

2500

[5] [6] [7] [8] [9] [10] [11] [12] [13]

[14] [15] [16]

[17] [18] [19] [20] [21]

[22] [23] [24] [25] [26] [27] [28] [29] [30]

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502 on Combustion, The Combustion Institute, Pittsburgh, 1996, pp. 1189–1199. P.D. Ronny, Proc. Combust. Inst. 27 (1998) 2485– 2506. R. Friedman, Fire Mater. 20 (1996) 235–243. R. Friedman, NASA TM-2998-207417, 1998. R. Friedman, NASA TM 1999–209285, 1999. A.G. Ruff, L.D. Urban, K.M. King, AIAA 2005– 0341. G. A. Ruff, D.L. Urban, AIAA-2007–350, 2007. J. Torero, Proc. Combust. Inst. 34 (2013) 99–124. ISO 14624-1, Space Systems – Safety and compatibility of materials-Part1 Determination of Upward Flammability of Materials, 2003. ISO 14624-2, Space Systems – Safety and Compatibility of Materials – Part2 Determination of Electrical-Wire Insulation and Accessory Materials, 2003. ISO/TS 16697, Space Systems — Safety and Compatibility of Materials — Method to Determine the Flammability Thresholds of Materials, 2012. S.L. Olson, P.V. Ferkul, in: Proceedings of the 42nd International Conference on Environmental Systems, AIAA 2012–3492, 2012. D.B. Hirsch, H.D. Beeson, Halon Options Technical Working Conference, 2001, pp. 7–12./D.B. Hirsch, H.D. Beeson, J. Testing Eval., 30(2) (2002) 156–159. D.B. Hirsch, J.H. Williams, S.A. Harper, H.D. Beeson, M.D. Pedley, in 2nd IAASS Conference Space Safety in a Global World, 2007. D.B. Hirsch, J.H. Williams, H.D. Beeson, J. Test. Eval. 1–6 (2006). D.B. Hirsch, A. Juarez, G.J. Peyton, S.A. Harper, S.L. Olson, AIAA2011-5067, International Conference on Environmental System, 2011. ISO 4589-2, Plastic – Determination of Burning Behaviour by Oxygen Index – Part2: Ambient – Temperature Test. ASTM D2863, Standard Test Method for Measuring the Minimum Oxygen Concentration to Support Candle-Like Combustion of Plastics (Oxygen Index). S.L. Olson, P.V. Ferkul, J.S. T’ien, Proc. Combust. 22 (1988) 1213–1222. S.L. Olson, Combust. Sci. Technol. 76 (1991) 233– 249. S. Bhattacharjee, R.A. Altenkirch, Proc. Combust. 24 (1992) 1669–1676. A.P. Ramachandra, R.A. Altenkirch, S. Bhattacharjee, B.L. Tang, K. Sacksteder, M.K. Wolverton, Combust. Flame 100 (1995) 71–84. S. Bhattacharjee, R.A. Altenkirch, K. Sacksteder, J. Heat Transfer 118 (1996) 181–190. J. West, L. Tang, R.A. Altenkirch, S. Bhattacharjee, K. Sacksteder, M.A. Delichatsios, Proc. Combust. 26 (1996) 1335–1343. R.A. Altenkirch, B.L. Tang, K. Sacksteder, S. Bhattacharjee, M.A. Delichatsios, Proc. Combust. 27 (1998) 2515–2524. A.V. Ivanov, Ye.V. Balashov, T.V. Andreeva, A.S. Melikhov, Experimental Verification of Material Flammability in Space, NASA/CR-1999-209405, 1999. D.B. Hirsch, H.D. Beeson, R. Friedman, Microgravity Effects on Combustion Polymers, NASA/ TM-2000-209900.

[31] S. Takahashi, H. Ito, Y. Nakamura, O. Fujita, Combust. Flame 160 (9) (2013) 1900–1902. [32] A.F. Osorio, K. Mizutani, C. Fernandez-Pello, O. Fujita, Proc. Combust. Inst. 35 (3) (2015) 2683– 2689. [33] Y. Zhang, P.D. Ronney, E. Roegner, J.B. Greenberg, Combust. Flame 90 (1992) 71. [34] L. Honda, P.D. Ronney, Combust. Sci. Technol. 133 (1998) 267–291. [35] S.L. Olson, D.B. Hirsch, P.V. Ferkul, Improving materials flammability screening for spacecraft fire safety, in: Proceedings of the 2010 Technical Meeting of the Central States Section of the Combustion Institute, 2010. [36] O.O. Feier, H.Y. Shih, K.R. Sacksteder, J.S. T’ien, Proc. Combust. Inst. 29 (2002) 2569–2577 (Upward flame spread over thin solids in partial gravity). [37] J.S. T’ien, S.N. Sinhhal, D.P. Harrold, J.M. Prahl, Combust. Flame 33 (1978) 55–68. [38] J.S. T’ien, Combust. Flame 33 (1986) 31–34. [39] J.S. T’ien, Combust. Flame 80 (1990) 355–357. [40] J.L. Rhatigan, H. Bedir, J.S. T’ien, Combust. Flame 112 (1998) 231–241. [41] U. Bonne, Combust. Flame 16 (1971) 147–159. [42] M. Sibulkin, A.K. Kulkarni, K. Annamalai, Proc. Combust. Inst. 18 (1981) 611–617. [43] M. Sibulkin, A.K. Kulkarni, K. Annamalai, Combust. Flame 44 (1982) 187–199. [44] A.C. Fernandez-Pello, S.R. Ray, I. Glassman, Proc. Combust. Inst. 18 (1981) 579–589. [45] A.C. Fernandez-Pello, T. Hirano, Combust. Sci. Technol. 32 (1983) 1–31. [46] A.C. Fernandez-Pello, “The Solid Phase”, chapter 2 of Combustion Fundamentals of Fire, Academic Press Ltd., 31–100, 1996. [47] R.A. Altenkirch, R. Eichhorn, P.C. Shang, Combust. Flame 37 (1980) 77–83. [48] R.A. Altenkirch, M. Vedha-Nayagam, Proc. Combust. Inst. 22 (1988) 1495–1500. [49] J.N. de Ris, Twelfth Symposium(International) on Combustion, The Combustion Institute, Pittsburgh, 1968, pp. 241–252. [50] M.A. Delichatsios, Combust. Sci. Technol. 44 (1986) 257–267. [51] S. Bhattacharjee, R.A. Altenkirch, Proc. Combust. Inst. 23 (1990) 1627–1633. [52] S. Bhattacharjee, R.A. Altenkirch, Combust. Flame 84 (1991) 160–169. [53] S. Bhattacharjee, K. Wakai, S. Takahashi, Combust. Flame 132 (2003) 523–532. [54] S. Bhattacharjee, R. Ayala, K. Wakai, S. Takahashi, Proc. Combust. Inst. 30 (2005) 2279–2286. [55] S. Takahashi, M. Kondou, K. Wakai, S. Bhattacharjee, Proc. Combust. Inst. 29 (2002) 2579–2586. [56] S. Takahashi, S. Bhattacharjee, T. Ihara, K. Wakai, J. Jpn. Soc. Microgravity Appl. 24 (3) (2007) 225–230. [57] S. Takahashi, K. Wakai, T. Ihara, S. Bhattacharjee, J. Jpn. Soc. Microgravity Appl. 25 (2008) 29–36. [58] S. Takahashi, M. Hotta, S. Bhattacharjee, T. Ihara, K. Wakai, J. Jpn. Soc. Microgravity Appl. 29 (1) (2012) 23–31. [59] P.V. Ferkul, J.S. T’ien, Combust. Sci. Technol. 99 (1994) 345–370. [60] J.S. T’ien, H.Y. Shih, C.B. Jiang, H.D. Ross, J. Miller, A.C. Fernandez-Pello, J.L. Torero, D. Wal-

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

[61] [62] [63] [64] [65] [66] [67] [68] [69]

[70]

[71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83]

[84]

[85]

ther, in: H. Ross (Ed.), Microgravity Combustion: Fire in Free Fall, Academic Press, New York, 2001. A. Kumar, H.Y. Shih, J.S. T’ien, Combust. Flame 132 (2003) 667–677. A. Kumar, K. Tolejko, J.S. T’ien, J. Heat Transfer 126 (2004) 611–620. C. Kumar, A. Kumar, Combust. Sci. Tech. 182 (9) (2010) 1321–1340. A. Kumar, J.S. T’ien, Int. J. Spray Combust. Dyn. 4 (4) (2012) 299–322. C. Kumar, A. Kumar, Combust. Theory Model. 16 (3) (2012) 537–569. A. Kumar, J.S. T’ien, Combust. Flame 164 (2006) 366–378. J.S. T’ien, M. Endo, Proc. Eng. 62 (2013) 120–129. L. Chu, J.S. T’ien, Upward Spreading Flames over Paper Samples, Case Western Reserve University Reports, FTAS/TR-78-139. T. Kashiwagi, K.B. McGrattan, S.L. Olson, O. Fujita, M. Kikuchi, K. Ito, in: 26th Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 1996, pp. 1345–1352. S.L. Olson, H.R. Baum, T. Kashiwagi, in: 27th Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, 1998, pp. 2525– 2533. W.E. Mell, S.L. Olson, T. Kashiwagi, Proc. Combust. Inst. 28 (2000) 2843–2849. S.L. Olson, T. Kashiwagi, O. Fujita, M. Kikuchi, K. Ito, Combust. Flame 125 (1–2) (2001) 852–864. K. Prasad, Y. Nakamura, S.L. Olson, et al., Proc. Combust. Inst. 29 (2002) 2553–2560. S.L. Olson, F.J. Miller, Proc. Combust. Inst. 32 (2009) 2445–2452. C.-B. Jiang, J.S. T’ien, H.-Y. Shih, Proc. Combust. Inst. 26 (1) (1996) 1353–1360. H.-Y. Shih, J.S. T’ien, Proc. Combust. Inst. 28 (2) (2000) 2777–2784. S.Y. Hsu, J.S. T’ien, Combust. Sci. Technol. 183 (2011) 390–407. S.Y. Hsu, J.S. T’ien, Proc. Combust. Inst. 33 (2011) 2433–2440 (Buoyant flow and Forced flow (concurrent). Y.T. Tseng, J.S. T’ien, Journal of Combustion, Hindawi Publishing Co., Article ID250391, 2011, p. 9. J.M. Tizon, J.J. Salva, A. Linan, Combust. Flame 119 (1999) 41–55. Y. Kudo, A. Ito, Proc. Combust. Inst. 29 (2002) 237–243. S.L. Olson, U. Hegde, S. Bhattacharjee, J.L. Reering, L. Tang, R.A. Alternkirch, Combust. Sci. Technol. 176 (4) (2004) 557–584. A. Hosogai, Y. Nakamura, K. Wakatsuki, Y. Kimoto, in: Proc. 8th International Symposium on Advanced Science and Technology in Experimental Mechanics (8th ISEM’13-Sendai), #089 (on CD-ROM) (2013). A. Hosogai, Y. Nakamura, K. Wakatsuki, Y. Kimoto, in: Proc. 29th International Symposium on Space Technology and Science, 2013-c-17 (on USB), 2013. P.S. Greenberg, K.R. Sacksteder, T. Kashiwagi, in: Proceedings of the Joint Launch Plus One Year Science Review of USML- and USMP-1 with the Microgravity Measurement Group, NASA CP 3272 2:631-655, 1994.

2501

[86] P.S. Greenberg, K.R. Sacksteder, T. Kashiwagi, in: 3rd International Microgravity Combustion Workshop, NASA CP 10174, 25-30, 1995. [87] T. Hirata, J. Sato, T. Iwanami, S. Iwasaki, T. Imagawa, K. Ito, in: Proceedings of the ThirtySixth Japanese Symposium on Combustion, 1998, pp. 480–482. [88] M. Kikuchi, O. Fujita, K. Ito, A. Sato, T. Sakuraya, Proc. Combust. Inst. 27 (1998) 2507– 2514. [89] O. Fujita, M. Kikuchi, K. Ito, K. Nishizawa, Proc. Combust. Inst. 28 (2000) 2905–2911. [90] O. Fujita, K. Nishizawa, K. Ito, Proc. Combust. Inst. 29 (2002) 2545–2552. [91] Y. Onishi, O. Fujita, K. Agata, et al., JSASS Aerospace Tech. Japan, 8:Ph_19–Ph_24, 2010. [92] S. Takahashi, H. Takeuchi, H. Ito, Y. Nakamura, O. Fujita, Proc. Combust. Inst. 34 (2013) 2657– 2664. [93] K. Nishizawa, O. Fujita, K. Ito, Trans. JSME (B) 70 (694) (2004) 1555–1562. [94] S. Takahashi, H. Izumo, H. Ito, Y. Nakamura, O. Fujita, in: Proc. 43rd International Conference on Environmental Systems, AIAA 2013–3388, 2013. [95] K. Agata, O. Fujita, Y. Ichimura, T. Fujii, H. Ito, Y. Nakamura, J. Jpn. Soc. Microgravity Appl. 25 (1) (2008) 11–16. [96] O. Fujita, T. Kyono, Y. Kido, H. Ito, Y. Nakamura, Proc. Combust. Inst. 33 (2) (2011) 2617–2623. [97] Y. Takano, O. Fujita, N. Shigeta, Y. Nakamura, H. Ito, Proc. Combust. Inst. 34 (2) (2013) 2665– 2673. [98] N. Shigeta, H. Ito, O. Fujita, in: Proc. 43rd International Conference on Environmental Systems, (2013.7), AIAA 2013–3387. [99] M.A. Delichatsios, R.A. Altenkirch, M.F. Bundy, S. Bhattacharjee, L. Tang, K. Sacksteder, Proc. Combust. Inst. 28 (2000) 2835–2842. [100] L. Hu, Y. Zhang, K. Yoshioka, H. Izumo, O. Fujita, Proc. Combust. Inst. 35 (3) (2015) 2607– 2614. [101] N.N. Bakhman, L.I. Aldavaev, B.N. Kondrikov, V.A. Filippov, Combust. Flame 41 (1981) 17–34. [102] N.N. Bakhman, L.I. Aldavaev, B.N. Kondrikov, V.A. Filippov, Combust. Flame 41 (1981) 35–43. [103] A. Tewarson, M.M. Khan, in: Twenty-Second, Symposium (International) on Combustion, The Combustion, Institute, Pittsburgh, 1988, pp. 1231– 1240. [104] A. Umemura, M. Uchida, T. Hirata, J. Sato, Proc. Combust. Inst. 29 (2002) 2535–2543. [105] X. Huang, Y. Nakamura, F.A. Williams, Proc. Combust. Inst. 34 (2) (2013) 2506–2512. [106] J. Lawton, F.J. Weinberg, Electrical Aspects of Combustion, Clarendon Press, Oxford, 1969. [107] I.N. Wakayama, Combust. Flame 93 (1993) 207– 214. [108] I.N. Wakayama, K. Ito, O. Fujita, H. Ito, Y. Kuroda, Combust. Flame, 107(1/2) (1996–10), pp. 187–192. [109] O. Fujita, K. Ito, T. Chida, S. Nagai, Y. Takeshita, in: 27th Symp. (Int.) on Combustion, 1998, pp. 2573–2578. [110] T. Ueda, O. Imamura, K. Okai, M. Tsue, M. Kono, J. Sato, Proc. Combust. Inst. 29 (2) (2002) 2595–2601.

2502

O. Fujita / Proceedings of the Combustion Institute 35 (2015) 2487–2502

[111] M.K. Kim, S.H. Chung, O. Fujita, Proc. Combust. Inst. 33 (1) (2011) 1145–1151. [112] P.D. Campbell, D.L. Henninger, C.J. Camarda, J.R. David, S.C. Doering, K.V. Rominger, JSC63309, 2006. [113] A.E. Frey Jr., J.S. T’ien, Combust. Flame 26 (1976) 257–267. [114] S.L. Olson, G.A. Ruff, Microgravity Flame over Non-charring Materials in Exploration Atmospheres: Pressure, Oxygen, and Velocity Effects on Concurrent Flame Spread, AIAA 2009-012489. [115] Y. Nakamura, N. Yoshimura, T. Matsumura, H. Ito, O. Fujita, J. Therm. Sci. Technol. 3 (3) (2008) 430–441. [116] Y. Nakamura, N. Yoshimura, H. Ito, K. Azumaya, O. Fujita, Proc. Combust. Inst. 32 (2009) 2559–2566. [117] Y. Nakamura, J. Therm. Sci. Technol. 3 (1) (2008) 58–67. [118] S. McAllister, C.A. Fernandez-Pello, D. Urban, G. Ruff, Proc. Combust. Inst. 32 (2009) 2453–2459. [119] S. McAllister, C.A. Fernandez-Pello, D. Urban, G. Ruff, Combust. Flame 157 (2010) 1753–1759. [120] S. Fereres, C. Lautenberger, C.A. FernandezPello, D. Urban, G. Ruff, Combust. Flame 158 (7) (2011) 1301–1306. [121] Y. Son, P.D. Ronny, Proc. Combust. Inst. 29 (2002) 2587–2594. [122] T. Kashiwagi, “Polymer Combustion and Flammability – Role of the Condensed Phase”, Invited Topical Review, in: Proceedings of the 25th Symposium (International) on Combustion, 1423–1437, 1994 (external radiation effect). [123] A. Osorio, C. Fernandez-Pello, D.L. Urban, G.A. Ruff, Proc. Combust. Inst. 34 (2) (2013) 2691–2697. [124] P.B. Sunderland, S. Mortazavi, G.M. Faeth, Combust. Flame 96 (1994) 97–103. [125] F.J. Diez, C. Aalburg, P.B. Sunderland, D.L. Urban, Z.G. Yuan, G.M. Faeth, Combust. Flame 156 (2009) 1514–1524.

[126] G. Legros, A. Fuentes, S. Rouvreau, P. Joulain, B. Porterie, J.L. Torero, Proc. Combust. Inst. 32 (2009) 2461–2470. [127] H. Ito, O. Fujita, K. Ito, Combust. Flame 99 (1994) 363–370. [128] O. Fujita, K. Ito, The Experimental Thermal and Fluid Science (The special issue), Elsevier Science 26(2–4), 305–311 (2002–6). [129] J.H. Choi, O. Fujita, T. Tsuiki, J. Kim, S. H. Chung, JSME Int. J. Ser. B, 48(4) (2005–11) 839–848. [130] J.H. Choi, O. Fujita, T. Tsuiki, J. Kim, S.H. Chung, JSME Int. J. Ser. B 49(1) (2006–2) 167–175. [131] B.H. Jeon, O. Fujita, Y. Nakamura, H. Ito, J. Therm. Sci. Technol. 2 (2) (2007) 281–290. [132] T.J. Ohlemiller, K.M. Villa, “Material Flammability Test Assessment for Space Station Freedom”, NISTIR 4591, NASA CR-187115, 1991. [133] T.J. Ohlemiller, in: An Assessment of the NASA Flammability Screening Test and Related Aspects of Material Flammability, NISTIR 4882, NASA CR-189226, 1992. [134] S.L. Olson, J.S. T’ien, Combust. Flame 121 (2000) 439–452. [135] S.L. Olson, H.D. Beeson, J.P. Haas, J.S. Baas, Proc. Combust. Inst. 30 (2005) 2335–2343. [136] S.L. Olson, F.J. Miller, S. Jahangirian, I.S. Wichman, Combust. Flame 156 (2009) 1214–1226. [137] I.S. Wichman, S.L. Olson, F.J. Miller, S.A. Tanaya, Fire Mater. 37 (2013) 503–519. [138] S.L. Olson, H.R. Baum, T. Kashiwagi, TwentySeventh Symposium (International) on Combustion, The Combustion Institute, 1998, 2525–2533. [139] D.L. Urban, G.A. Ruff, C.A. Fernandez-Pello, et al., IAC-12. A2.2.2., 2012. [140] G.A. Ruff, D.L. Urban, A.C.A. Fernandez-Pello, et al., in: 43rd International Conference on Environmental Systems, AIAA 2013–34140, 2013. [141] M. Kikuchi, T. Mizushima, O. Fujita, et al., in: Proc. 5th International Symposium for Physical Sciences in Space (ISPS), 2013.