Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Heat Exposure and Burning Behavior of Cabin Materials During an Aircraft Post-Crash Fuel Fire Constantine P. Sarkos* ABSTRACT The Federal Aviation Administration (FAA) has conducted numerous full-scale aircraft fire tests for the purpose of characterizing a post-crash cabin-fire environment and developing improved criteria for fire testing cabin materials. The tests subjected aircraft fuselages to an external fuel fire, usually adjacent to an opening in the fuselage. This paper reviews those tests. Emphasis is on the heating conditions experienced by the fuselage skin and cabin materials near an opening; the effect of wind and door openings on heat flux and cabin hazard development arising from the fuel fire; the important cabin phenomena related to survival, such as stratification of fire hazards and flashover; and hazard-time profiles and materials fire involvement. BACKGROUND Aircraft fire safety involves both in-flight and post-crash fire considerations. Fatal or uncontrollable in-flight fires are rare events; in fact, 30 years have passed since the last fatal accidental in-flight fire involving a U.S. commercial transport. Therefore, fire-test criteria for aircraft cabin materials, required by the FAA, are based primarily on post-crash fire conditions. Practically all post-crash aircraft fires are initiated by the ignition of jet fuel released from the damaged fuel system. Also, the relevancy of the fire performance of cabin materials to life _ ~ , ~ ~ ~ . · , ~ , · · · ~ · ~ , ~ i. r ~ · ~ ~ · . . __ safety Is most relevant In survivable accidents when the fuselage Is largely Intact. Hence, an intact fuselage subjected to an external fuel fire represents a post-crash fire scenario in which burning cabin materials may affect occupant escape. Past accidents, experiment studies, and a knowledge of fuselage design all show that a fuselage opening, perhaps a crash rupture or inadvertently opened emergency exit, provides the earliest opportunity for fire to enter the cabin. This mode of fire penetration contrasts with burnthrough of the fuselage shell. Burnthrough generally occurs later in time than when flame penetration occurs directly through an opening. In the past, the FAA has conducted many full- scale fire Bests, utilizing the aforementioned scenario, which has an external fuel fire adjacent to an opening in an otherwise intact fuselage, to characterize post-crash cabin fires and to develop improved fire-test criteria for cabin materials. This paper reviews those past FAA full- scale fire tests. *Fire Safety Branch, Federal Aviation Administration Technical Center, Atlantic City, New Jersey. 25
26 Imp roved Fire- and Smok;e-Resistant Materials FULI'SCALE FIRE TESTS Pool Fire The heat flux created by a pool of burning jet fuel is dominated by intense thermal radiation. For pool fires greater than about 3 in. diameter, the radiative heat flux is relatively invariant and approximately 14 Btu/ft2 · s (Eklund and Sarkos, 1980~. The convective heat flux is roughly 10-20 percent of the radiative component, depending of the size of the fire and other factors, and is, consequently, of less importance. Pool F~re/Fuselage Impact One of the earliest experiments to measure the fuselage skin heating conditions arising from an adjacent fuel fire utilized a titanium fuselage (Sarkos, 19711. The test configuration was a 28-ft long titanium fuselage, ~ ~ feet in diameter, abutting a 20-square-foot fire pit. The heat flux to the skin and flame temperature histories are shown in figures ~ and 2. During the test, a firewhirl developed adjacent to the fuselage at the aft end. At 80 seconds, the firewhir! unexpectedly moved to the forward end of the fuselage, where it lodged for 5-10 seconds, and then returned to the aft end, where it remained until the cessation of the test. The heat-flux measurements (total), which included the sum of the radiative and convective components, reflected the behavior of the firewhirl. When the effect of the firewhirl was minimal (center section), the heat flux fluctuated between 10 and 14 Btu/ft2 s, as expected. However, during the period of greatest intensification of the firewhirl, the heat flux attained peals values of IS Btu/ft2 s. The flame plume temperature measurements closely followed the trends exhibited by the heat flux. Flame temperatures ranged from 1400-~800 °F when the effect of the firewhir! was minimal but increased to as high as 2000 OF in the firewhirl. The swirling motion of the firewhirl causes increased air entrainment, raising the air-to-fuel ratio and combustion efficiency. Although the titanium skin was an effective fire barrier. it is noteworthy that a cabin flash fire _ , . . ~ . . . . .. . ... ~ . , ... ~ . . .. .. occurred at ~ minutes, caused by the ignition of combustible gases formed by the thermal degradation of the cabin pressure sealant elastomer, which was bonder! to the interior of the titanium skin, and of the thermal/acoustical fiberglass insulation. Radiant Heating of Cabin through Fuselage Opening As discussed earlier, the creation of a post-crash fuel fire at a fuselage rupture or door opening provides the most immediate fire threat to the cabin interior. A series of experiments was conducted in which fuselage models under zero wind conditions were subjected to a fuel fire to determine the amount of radiation to the interior through a door opening (Eklund, 19781. The models were essentially empty cylinders liner! with a ceramic fibrous insulation. Calorimeter measurements were taken at two locations: (~) on the fuselage symmetry plane, at an elevation of one-half of the doorway height, and (2) on the floor, midway between the doorway and
Constantine P. Sarkos 2n - - 12 X W 8 FIGURE 1 Titanium fi - stage incident heat flux. 2800~- Z400 2000 w 3 160a 54 1 20C g ages Penn 1 1 1 1 1 1 ~ ~ 1 1 1 1 1 i I I ~ 1 1 rr~-l~ Lot · 1 FUSELAG E . AFT CTR FWD 1 _ FIRE _ PIT _ 40 60 80 100 120 TIME AFTER IGNITION (SECONDS) TIME AFT`:R IGNITION tSISCONDIl' FIGURE 2 Titanium fuselage fuel-fire flame temperature. 27 ,......... - LEGEND: . _--FORWARD SECTION_ -AFT SECTION . - CENTER SECTION . . . . . . . . ~ ~ ~ ~ I ~ ~ FUSELAGE _ AFT CTR FWD _ _ . . . _ _ ._ FIRE _ PIT _ _ 166 zoo symmetry plane. For external fuel fires, the heat flux measured at the symmetry plane and at the floor was 1.8 Btu/ft2.s and 2.5 Btu/ft2 s, respectively. This study also developed the equations for predicting the internal heat flux. The tests showed that the doorway could be treated as a black body radiating heat at a temperature of 1874 OF. Figure 3 shows a representation of the heat flux to the interior calculated from the derived equations. It is evident that the high levels of radiative heat flux associated with the fuel fire will be confined to a small
28 Improved Fire- arm Smoke-Resistant Materials region near the doorway when there is no penetration of the fuel-fire flames through the opening for example, under a zero wind condition. Thus, during the early stages of a fire the thermal degradation and subsequent ignition of cabin interior materials will be very localized and limited to the doorway vicinity if the fuel-fire flames do not enter the cabin. / / Id/ ~.~, ~ J: Btu/ft - ·ec:_ _ _ _ __V: / In FUEL PAN 1 ~"~/ ~ FIGURE 3 Theoretical radiative heat-flux profiles through a fuselage opening. Effects of Wind on Cabin Heating H ~////////////////' A series of full-scale fire tests essentially confirmed the modeling experiments discussed above (Brown, 1979). A fire-hardened DC-7 fuselage, protected on the outside with steel sheeting and lined inside with a ceramic fibrous insulation, was subjected to a 20-square-foot fuel fire. A doorway was placed alongside the fire, and two doorways were also placed away from the fire on opposite sides of the fuselage. The variation of the symmetry plane heat flux during three full-scale tests is shown in Figure 4. Conditions during the test with calm wind and all doors closed more closely matched the modeling experiments. In this case, the agreement between the modeling and full-scale results was very good. Test results suggested that when all doors were closed the pressure rise in the heated cabin, however slight, was not relieved and tended to prevent flame entry. In the test with all doors open and calm winds, the heat flux increased steadily above the modeling value. Higher heat-flux readings and significant variations
Constantine P. Sarkos 29 were measured in the test with high winds (4-10 miles per hour) and all doors open. The fuel- fire flames appeared to periodically enter and withdraw from the cabin, almost doubling the symmetry plane heat flux dunng peak heating conditions yet returning to the modeling value during minimal heating. 4-10 MPH WIND ALL DOORS OPEN ;,`t 3 o 2 2 o .~ - a, x U. 'c- 1 o FIGURE 4 DC-7 symmetry plane heat flux. I ~ / I, "; I CALM WIND ALL At' / DOORS OPEN ~ ~ ,1 \ r-~ ~'_ · ~ _e~ .' --~r.~ CALF WIND ALL DOORS CLOSED MODELING PREDICTION ·. _--. DOOR FIRE DOOR AL ' X:I ~ l l ; l l UPWIND DOOR 30 40 50 60 TIME ~ SECONDS Another factor that has been shown to have a critical effect on fire penetration is the door opening configuration. For the full-scale test conditions illustrated in Figure 5, the worst case is when the fuel fire is upwind of the fuselage and door openings are on the downwind side of the fuselage. In this case, the DC-7 tests have shown that there is a rapid development of fuel- fire hazards inside the fuselage. Conversely, if the door openings are on the upwind side only, the wind entering the door opening away from the fire has a buffering effect on fire penetration, and the accumulation of fuel-fire hazards within the fuselage will be greatly retarded. Figure 5 compares the ceiling heat-flux level for those extreme conditions. With minimal flame penetration, the ceiling heat-flux ranged from ~ to 2 Btu/ff s. Current aircraft materials are virtually noncombustible under these heat-flux exposure values. Conversely, downwind door openings created maximum flame penetration, causing the ceiling heat flux to reach 10 Btu/ft2 s. At this heat flux, aircraft materials will ignite quickly and burn rapidly. However, since the
30 Imp roved Fire- aru] Smoke-Resistant Materials cabin became rapidly nonsurvivable from the fuel-fire hands alone, the role of material flammability in cabin-fire hazard would be of minimal consequence in this scenano. 10 6 en m X LL ~ 4 Ul 2 n FIGURE 5 DC-7 ceiling heat-flux histories. DOWNWIND DOOR / 1 / DOWNWARD ~r ONLY OPEN, I I / DOOR ~' 8.5 -13 mph | V ~ _ 1 X / UPWIND Lit / DOOR 1 JO 1 /N I ·,:. 1 ~. \ ~UPWIND DOOR ~, . .~. ~.' '., ONLY OPEN, :,. ' "" 'a ;, , ,., . ..~. \! , , , - , , 0 ~30 40 50 60 70 TIME - SECONDS Fuel-fire Hazards Accumulation within Cabin In terms of the post-crash fire-heat exposure of cabin matenals, the impact of fuel-fire penetration through a fuselage opening is twofold. First, matenals in the vicinity of the opening will experience higher radiant heat exposure from fuel flames than when the fire does not penetrate, as discussed previously. Second, cabin materials will also be subjected to the hot fuel- fire smoke accumulating in the cabin. The quantity of hot smoke will increase as the flame penetration becomes greater. A large number of experiments were conducted in a wide-body (surplus C-133) test article, devoid of cabin matenals, to examine the hazards associated with the fuel fire alone (Hill et al., 1979; Hill and Sarkos, 1980~. An important finding was the pronounced stratification of heat, smoke, and toxic gases that is, the hot smoke from the fuel fire accumulated at the ceiling. Results from a typical test are shown in Figure 6. The hot smoke layer near the ceiling was very distinct and did not thicken appreciably with time. Temperatures
Constarair~e P. Sarkos 31 in the lower region were much lower than in the hot smoke layer and decreased almost linearly to near ambient conditions at the floor. Thus, heat exposure of cabin materials by a fuel fire adjacent to a fuselage opening consists of intense radiant exposure of materials very close to the opening and convective/radiative heating of materials in the upper cabin, including ceiling and stowage bins, extending over some area away from the fire penetration opening. The size of this area is determined by heat losses to the smoke layer due to mixing with entrained air and surface heat conduction. The heat tosses are very sub social; for example, ceiling layer temperature gradients as high as 1000 OF along a 60-foot cabin length were measured during the C-133 tests. 100 80 t~ 60 z - - CD 40 20 1 /// 06 1 VALUES PRESENTED IN MINUTES 1 2 TESTES / | | 3 4 HEAT STRATIFICATION 50 ft. FROM FIRE DOOR AT THE CENTERLINE OF FUSELAGE 0 100 200 300 400 TEMPERATURE (OF) FIGURE 6 C-133 cabin air temperature stratification. Characteristics of Burning Cabin Materials 500 600 700 The FAA has conducted numerous full-scale fire tests on cabin materials subjected to the previously discussed post-crash fire scenario, that is, an external fuel fire adjacent to a fuselage opening (Sarkos et al., 1982; Sarkos and Hill, 1982, 1985, 1989; Hill et al., 1984, 1985~. The purpose of those Bests was to determine survivability gains from improved materials and to develop improved fire-test criteria. Test observation and data analysis provide the following description of the characteristics of a cabin fire. The seat closest to the fuel fire is the initial material ignited. As the seat burns, its combustion products accumulate and spread along the ceiling. Later, the fire spreads to portions of the seats fore and aft of the initial seat ignited;
32 Improved Fire- aM Smoke-Resistant Materials ceiling and stowage bins above the burning seats also ignite and burn. The fire remains localized and confined to the three outboard seats and overhead materials, which flash intermittently in the black smoke layer. A distinct partitioning of the cabin is created a hot overhead smoke layer clinging to the ceiling, approximately 2 or more feet thick and a clear region below the ~ ~ ~ ~ . ~ . . ~ . . . %~ ~ ~ ~ . ~ ~ . . ~ smoke layer, largely at or near ambient conditions. lne smoke layer spreads throughout the cabin. The observed fire below the smoke layer remains localized, and the two-zone effect persists until the occurrence of a flashover. Although definitions of flashover vary, it is basically a sudden, very rapid spread of fire within an enclosure. The extent of the spread of the fire is dependent on the availability of oxygen, which is consumed in great quantities because of the tremendous burning rates associated with the flashover. Cabin flashover is clearly the critical factor affecting occupant survivability during a post- crash fire that is dominated by burning cabin materials (as opposed to a post-crash fire in which survival is governed by the fuel-fire hazards). Cabin hazard history measurements taken during a full-scale fire test in a wide-body test article are shown in Figure 7 (Sarkos and Hill, 1985~. O O_ A) lo . VISIBILITY (a.) x 0.1 02 (%) X 0.2 . C-133 STATION 650 HEIGHT 5.5 FT. TEMPERATURE (°F^) x .01 //HCI (ppm) x .01 1/~- C02(%) 0.0 30.0 60.0 FIGURE 7 C-133 cabin hazard histones. //~ CO(%)x10 / //~ 90.0 120.0 150.0 180.0 210.0 TIME - SECONDS ~ HCN x .01 Before the flashover, which occurred at approximately 150 seconds, the cabin environment was clearly survivable; after flashover, the conditions deteriorated to such a degree that survival would have been highly unlikely. For example, as shown in Figure 7, within 30 seconds visibility was reduced from clear to 3 feet, temperature increased from ambient to 400 °F, carbon monoxide jumped from zero to 2,500 parts per million, and oxygen dropped from 21 percent to 16 percent. Obviously, improvements in post-crash fire survivability, when burning cabin materials predominate, can best be achieved by taking measures that delay the onset of flashover.
Constantine P. Sarkos 33 The primary areas of fire involvement during a post-crash cabin fire, as described above, are seats near the fire origin and ceiling materials. The most extensively furnished fire test in the C-133 wide-body test article corroborated this finding, which was observed in earlier experiments that employed a cluster of materials about the fire opening (Sarkos and Hill, 1989~. In the "fully furnished" C-133 test, the forward cabin was completely furnished over a length of 45 feet (see experimental arrangement in Figure 81. Fourteen rows of seats, in a double-triple- double seating arrangement, and a triple seat in front of the galley (totaling 101 seats) were used. Aircraft ceiling panels, stowage bins, sidewalls, and carpeting were installed throughout the furnished cabin length, as well. As in previous experiments, survivability was driven by cabin flashover and extreme fire-hazard gradients were documented. After the test, it was observed that the entire ceiling was consumed by fire, as were the outboard seats in the immediate vicinity of the fire door. On many of the remaining seats the dress cover of the seat back was largely burned away, but the seat fire-blocking material underneath was still present and intact. INSTALLED CEILING GALLEY PARITION (STATION 540) ~16 FT BIN.- ~ OFT ~ _ ~BFT INSTALLED FLOOR - -a I A _-_ CARGO FLOOR - 421N. ~ - 14FT 101N. ~ _ CEILING AREA I I - FIRE DOOR 761N. 1 ' ~ GALLEY PARITION _ .. ala I. .,. 1 1 ' ~ ~ AIR EXHAUST DOOR 761N. i 1 / o 140 A~ 540 650 880 920 INTERIOR: INSTRUMENTATION: FURNISHED UNFURNISHED TEMPERATURE/HEAT FLUX GASES/SMOKE/TEMP FILTER PERFORMANCE FIGURE 8 C-144 fully furnished full-scale test arrangement. - An account of the fire exposure/behavior of seats located away from the fire door is shown in figures 9 and 10. Thermocouples and calorimeter (facing the ceiling) measurements were taken near the top of the center seat at the noted seat rows (row 4 was at the fire door).
34 Improved Fire- arm Smoke-Resistant Materials As shown in Figure 9, it appears that the onset of flashover occurred at 210 seconds. Before flashover, the data indicate that the fire had not spread to the center seats nor were these seats undergoing thermal degradation. After flashover, the seat top temperatures peaked at 1600 to 1900 °F. Based on the separation between the rising portions of the temperature profiles, the flashover propagated at a speed of approximately 60 feet/minute, or at a rate of one seat row about every 3 seconds. The trailing edge of the temperature profile shows the fire self- ex~cinguished and the cabin cooled down. Oxygen concentration measurements indicated that the fire became oxygen-starved; oxygen concentration readings at the seat top level decreased to less than 5 percent throughout the test article. The calorimeter profiles in Figure 10 provide an indication of the duration and intensity of flaming combustion in the upper cabin caused by flashover. Tom fire involvement of the upper cabin lasted for about ~ minute; the burning rate (heat release) was greater near the fire door and decreased toward the rear of the furnished cabin. 2000 1600 - IL a - u' 1200 In Ul Al . 800 400 o . . ROWS . a_ 11 \~7\ I t' 1''t: ~--~---~-~--~-~. ~ 0 60 120 180 240 300 360 420 TIME (SEC.) FIGURE 9 C-133 fully furnished test seat top temperatures.
Constantine P. Sarkos 15 FIGURE 10 C-133 fully furnished test seat top heat flux. 20 - - ~ 10 IU S ROW 4 ~ i: ROW 1 ~ 1; ROW 13 ~ '.~, 0 60 120 180 240 300 360 420 480 TIME (~EC.) DEVELOPMENT OF IMPROVED FIRE TEST CRITERIA FOR CABIN MATERIALS 35 In recent years, the FAA has implemented two important standards to minimize the hazards of burning cabin materials during a post-crash fire (Sarkos, 1989~. Full-scale fire tests have demonstrated that both material fire performance advances seat cushion fire-blocHng layers and low-heat-release panels improve passenger survivability by extending the time of cabin flashover (Sarkos and Hill, 1982; Hill et al., 19851. Future development of advanced fire- resistant materials should aim to further delay or possibly eliminate the occurrence of flashover. The design scenano~s) should be based on a fuel fire that penetrates into the cabin through a fuselage opening but is not so severe as to cause nonsurvivable conditions from the fuel fire itself (without any involvement of cabin materials). Materials targeted for improvement should be initially focused on the upper cabin area (ceiling panels and stowage bins) and seat cushions and should be based on the measured cabin heating conditions and fire growth and the degree of material fire damage observed in full-scale fire tests. REFERENCES Brown, L.~. 1979. Cabin Hazards From a Large External Fuel Fire Adjacent to an Aircraft Fuselage. FAA final report FAA-NA-79-65. Atiantic City, New jersey: Federal Aviation Administration Technical Center.
36 Improved Fire- arm Smoke-Resistant Materials Eklund, T.~. 1978. Pool Fire Radiation Through a Door in a Simulated Aircraft Fuselage. FAA final report FAA-RD-78-135. Atlantic City, New Jersey: Federal Aviation Administration Technical Center. Eklund, T.~., and C.P. Sarkos. 1980. The thermal impact of external pool fires on aircraft fuselages. Journal of Fire and Flammability Il(3~:231-240. Hill, R.G., and C.P. Sarkos. 1980. Post-crash fuel fire hazard measurements in a wide body aircraft cabin. Journal of Fire and Flammability Il(2~:151-163. Hill, R.G., G.R. John son and C.P. Sarkos. 1979. Postcrash Fuel Fire Hazard Measurements in a Wide-Body Aircraft Cabin. FAA final report FAA-NA-79-42. Atlantic City, New Jersey: Federal Aviation Administration Technical Center. Hill, R.G., L.~. Brown, L. Speitel, G.R. Johnson, and C.P. Sarkos. 1984. Aircraft Seat Fire Blocking Layers: Effectiveness and Benefits Under Various Scenarios. FAA final report DOT/FAA/CT-83/43. Atlantic City, New Jersey: Federal Aviation Administration Technical Center. Hill, R.G., T.~. Eklund, and C.P. Sarkos. 1985. Aircraft Interior Pane! Test Criteria Derived From Full-Scale Fire Tests. FAA final report DOT/FAA/CT-85/23. Atiantic City, New Jersey: Federal Aviation Administration Technical Center. Sarkos, C.P. 1971. Titanium Fuselage Environmental Conditions in Postcrash Fires. FAA final report FAA-RD-71-3. Atlantic City, New Jersey: Federal Aviation Administration Technical Center. Sarkos, C.P. 1989. Development of improved fire safety standards adopted by the Federal Aviation Administration. Pp. 5-l to 5-13 in AGARD Conference Proceedings No. 467 on Aircraft Fire Safety, Sintra, Portugal, May 22-26, 1989. Neuilly-Sur-Seine, France: North Adantic Treaty Organization, Advisory Group for Aerospace Research and Development. Sarkos, C.P., and R.G. Hill. 1982. Effectiveness of Seat Cushion Fire Blocking Layer Materials Against Cabin Fires. SAE paper No. 821484. Presented at Aerospace Congress and Exposition, Anaheim, California, October 25-2S, 1982. Warrindale, Pennsylvania: Society of Automotive Engineers, Inc. Sarkos, C.P., and R.G. Hill. 1985. Evaluation of Aircraft Interior Panels Under Full-Scale Cabin Fire Test Conditions. ATAA-85-0393. AlAA 23rd Aerospace Sciences Meeting, Reno, Nevada, January 14-17, 1985. New York: American Institute of Aeronautics and Astronautics. Sarkos, C.P., and R.G. Hill. 1989. Characteristics of aircraft fires measured in full-scale tests. Pp. It-! to Il-17 in AGARD Conference Proceedings No. 467 on Aircraft Fire Safety, Sintra, Portugal, May 22-26, 1989. Neuilly-Sur-Seine, France: North Atiantic Treaty Organization, Advisory Group for Aerospace Research and Development. Sarkos. C.P.' R.G. Hill, and W.D. Howell. 1982. The development and application of a full- scale wide body test article to study the behavior of interior materials during a postcrash fuel fire. Pp. 6-l to 6-21 in AGARD Lecture Series No. 123 on Aircraft Fire Safety, OsIo, Norway, June 7-8, 1982. Neuilly-Sur-Seine, France: North Atiantic Treaty Organization, Advisory Group for Aerospace Research and Development. - - - - r - ~-at- r - --- - - -
