RATE OF HEAT RELEASE OF PLASTIC MATERIALS FROM CAR INTERIORS

Transcription

1 RATE OF HEAT RELEASE OF PLASTIC MATERIALS FROM CAR INTERIORS Marcelo M. Hirschler * and Donald J. Hoffmann **, John M. Hoffmann ** and Elizabeth C. Kroll ** * GBH International, 2 Friar s Lane, Mill Valley, CA, ** Safety Engineering Laboratories, College Park Dr., Warren, MI, Abstract Materials in car interiors need to meet a horizontal flame spread rate of # 4 in./min, when tested according to US Federal Motor Vehicle Safety Standard 302, but the results of this test have been shown not to predict fire hazard. In this work, several materials and combinations, from the interiors of passenger cars, were tested in the cone calorimeter, at 25 & 40 kw/m 2 incident heat flux, with the emphasis on application as: engine cover, HVAC ducting, headliner, and seating. Three full scale tests were also conducted using actual cars. Test results were compared with the fire performance of some typical commercial plastics and with those of some component fire tests conducted by Factory Mutual as part of the agreement to conduct fire safety between a car manufacturer and the federal government (General Motors and the National Highway Traffic Safety Administration). Test results indicate that most of the car interior materials tested, which were reasonably representative of the type of materials contained in the vehicles tested, represent a fire performance that is, at best, mediocre. The fire hazard consequence of the data analysis is that flashover in a passenger vehicle interior does not require large ignition sources., so that even very small ignition sources can generate a large enough fire to severely hinder passenger egress (whether aided or unaided). Introduction Fire statistics indicate that the fraction of fire fatalities that occur in passenger road vehicles (or highway vehicles) is very significant: 8.6% of all fire fatalities in the United States (between 1985 and 2000), 10.6% in the United Kingdom (between 1985 and 1999) and 5.9% in the province of Ontario (Canada, its largest province, between 1990 and 1999) (Table 1) [1-4]. This is of course much less than the proportions of fire fatalities in dwellings (residences), which are 78.2, 75.7 and 86.7% respectively, but is still a very large population. Thus, clearly, the problem of fire fatalities in highway vehicles is not negligible, even if it is a much smaller problem than fire fatalities in buildings. NFPA statistics also give information on the material and item first ignited and on the time of day. It is interesting to note that, while the largest proportion of fires occurs during the afternoon rush hour, the largest proportion of fire fatalities occurs around midnight. Among passenger road vehicles, the NFPA statistics show that automobiles represent 95% of the fires and 92% of the fire fatalities (Table 2) [2].

2 The fire performance requirements are the same for all automotive interior materials (present within 13 mm (0.5 inch) of the passenger compartment) in the United States, since it was proposed in 1969, becoming effective in September 1972: the flame spread rate should not exceed 102 mm/min (4 inches per minute) following the application of a very mild flame in US FMVSS 302 [5]. Interestingly, materials present under the dash do not even have to meet the fire test, irrespective of whether they are, or are not, visible from the passenger compartment. This fire test requirement has not changed since 1972, in spite of the significant advances in fire sciences. In particular, there are no requirements for ignitability, heat release or smoke obscuration. Interestingly, the requirements are virtually identical in other parts of the world, except that the referenced test method is ISO 3795 [6] instead of US FMVSS 302. In 1979, the National Materials Advisory Board (NMAB) [7] reviewed a variety of test methods used for determining the flammability of materials in ground transportation vehicles. The NMAB stated that the FMVSS 302 test was not representative of the fire exposure conditions potentially present in vehicles and that the FMVSS 302 standard lacked significance, was ineffective and was extremely lenient. Such conclusions are still valid today, and are particularly poignant since many new fire tests exist, several of which have been shown to be predictive of realistic conditions. Vehicle fires that can result in harm to drivers or passengers can originate inside the passenger compartment, but are more likely to originate outside and spread into the passenger compartment. Thus, for example, fires can start in the engine compartment or in the fuel tank area and then spread into the passenger compartment either by penetrating the engine cover (although this is supposed to be a fire wall) or by burning through some of the ducts. Fire propagation from the engine compartment clearly will be easier as the fire performance of the engine cover (or its sealants) becomes poorer and as the number and size of the openings in it increase. One project investigated 13 collision-related fires and showed that fire originating in the engine compartment reached the passenger compartment in less than eight minutes and occasionally in as low as 2-4 minutes [8]. The most common ignition factor for passenger road vehicles are mechanical or electrical failures (65.7% between 1994 and 1998 in the US, [2], with electrical shorts at 18.4%), followed by incendiary or suspicious (17.1%). While collisions are the cause of only 1.8% of fires, they cause 59% of fire fatalities. A significant amount of work is going into studying the causes and effects of collision related vehicle fires e.g. [8]. Such fires are affected by ventilation issues (e.g. as a result of broken windows or deformed frames). It is interesting to note that vehicles involved in major fires are almost always a complete loss, as it is rare to find partially burnt vehicles on the road, while vehicles with collision damage are a common sight. Furthermore, it would appear that, in the event of a vehicle fire, conditions inside the passenger compartment (where passenger mobility is often impaired by injury) become untenable quite rapidly: this will be investigated in the present work. Finally, the amount of combustible material present in passenger vehicles is permanently increasing (in fact, it has increased several fold since US FMVSS 302 went into effect), so a review of the fire performance that should be expected of such materials is needed. For this work fire tests were conducted using two standard tests and two ad-hoc tests, and over 60 materials, from seven different types of cars that have been involved in severe cabin material fires in the 1990's. The standard tests were the regulatory requirement (FMVSS 302) and the cone calorimeter (ASTM E 1354 [9], ISO 5660 [10]). The ad-hoc tests were a small-scale vertical test (used for several

3 materials) and a real scale tests on an entire passenger car, conducted on three cars. Some of the cone calorimeter test results have already been presented [11, 12]. The data is also compared with that from a study of 35 commercial materials available in the same period [13]. Experimental Federal Motor Vehicle Safety Standard 302 (FMVSS 302): The test specimens are 102 x 356 mm (4 x 14 in.) by use thickness (up to 13 mm (0.5 in.). If the material in the vehicle is used at a greater thickness, it must be shaved down to maximum thickness. Where it is not possible to obtain a flat specimen, because of the component configuration, the specimen is cut to not more than 13 mm (0.5 in.) thickness at any point, from the area with the least curvature in such a manner to include the face side. The specimen is placed in a frame inside a metal cabinet and held within a U-shaped frame, and a 38 mm (1.5 in.) Bunsen burner gas flame is applied to the edge of the specimen for 15 s. The specimen is allowed to burn horizontally until: 1) the flame goes out, 2) the flame-front reaches a mark at 38 mm (1.5 in.) from the unlit end of the specimen, or 3) 15 min have elapsed. A material passes the test if it does not burn, or transmit a flame front across its surface, at a rate of more than 102 mm/min (4 in./min). A material also passes the test if it stops burning within 60 s and has not burned more than 51 mm (2 in.) from the timing mark (located 38 mm, 1.5 in., from the lit end of the specimen). These fire-test-response requirements apply to all materials in the occupant compartment of passenger cars, multipurpose passenger vehicles, trucks and buses. Cone calorimeter rate of heat release instrument (ASTM E 1354, cone): The test specimens are 100 x 100 mm (almost 4 x 4 in.) by use thickness (up to 51 mm, 2 in.) and they are exposed to a constant incident heat flux (ranging up to 100 kw/m 2 ) in the presence of a spark igniter. All tests reported here were conducted in the horizontal orientation, without using edge frame or wire grid, at heat fluxes of 25 and 40 kw/m 2 (except that one seat assembly was tested at 35 kw/m 2 ), at the specimen thicknesses used in the actual automotive application. Parameters that can be reported include: peak rate of heat release (Pk RHR, in kw/m 2 ), average rate of heat release (for the entire test and for the 3 min following ignition) (Avg RHR, in kw/m 2 ), time to sustained combustion, or time to ignition (TTI, in s), total heat released (THR, in MJ/m 2 ), effective heat of combustion (Ht Comb, MJ/kg), specific extinction area (SEA, in m 2 /kg), smoke factor (SmkFct, in MW/m 2 ), peak rate of smoke release (Pk RSR, in 1/s), the total smoke released (TSR, non-dimensional), fire performance index (ratio of time to ignition to peak rate of heat release (FPI, in s m 2 /kw), mass loss (MsLs, g or %) and time to extinction (TTE, s). Some of the variables used will be explained briefly. * Rate of Heat Release (RHR): The instantaneous amount of heat released per nominal sample surface area (based on the oxygen consumption principle). For each experiment, the maximum RHR value is the most significant one and is recorded here. * Time to Ignition (TTI): The time, in s, until the entire surface of the sample burns with a sustained luminous flame.

4 * Total Heat Released (THR): The integral of RHR data as a function of time, per unit nominal sample surface area. * Rate of Smoke Release (RSR): The rate of smoke release is calculated by equation [I]: RSR = (V * OD * ln10)/ (Sample area * Light Path Length) (I) where V is the volumetric flow rate (in m 3 /s, corrected for the relative locations of the flow measurement device and the light measurement device and for the elevated temperature in the duct), OD is the optical density, the Light Path Length is m and the sample area is m 2. The units are 1/s. * Smoke Factor (SmkFct): An empirical smoke/fire hazard variable calculated as the product of total smoke released and peak RHR. The total smoke released is calculated as the time integral of the rate of smoke release. * Fire performance index (FPI): This parameter, which approximately follows trends of time to flashover, is an empirical variable, calculated as the ratio of the time to ignition and the peak rate of heat release. It is, thus, a good indicator of overall fire hazard. Vertical burn demonstrations: Some of the car interior materials were exposed to a small Bunsen burner flame (38 mm, 1.5 in.), while in a vertical orientation, to investigate qualitatively the extent of upward flame spread and the melting and dripping characteristics of the material. The rationale for conducting these demonstrations is that exposure to flame in a vertical orientation is a more severe fire test than exposure to a similar flame in a horizontal orientation. Real-Scale car tests: Three vans were purchased and exposed to realistic fire scenarios. In the first test, a van was tested attempting to replicate the conditions of an earlier test by another author. The van was not modified, but the passenger and driver door windows were rolled down threequarters of the way. The modified van was positioned on a concrete pad, at an exterior location, with wind speeds of km/hr (10-15 miles/hr). Twenty six K-type air thermocouples were positioned inside the van: 8 on the headliner, sheathed in a steel conduit which extended the length of the van and was fixed to the ceiling with wire (positioned centerline along the length of the van from the windshield to the rear of the van, with 4 located between the windshield and the driver and passenger seats), 2 affixed to the exposed surfaces of each bench seat, and 14 distributed across on the top and face of the dash, in the HVAC vents and on the engine cover. Thermocouple temperatures were recorded every 5 s. Four video cameras were used to record the demonstration: one mounted inside behind the drivers seat and directed toward the engine cover and passenger side of the van and three located outside the van. Still photographs and video were used to document the condition of the interior of the van and the location of instrumentation before and after the demonstration. A temperature baseline was established for 3 min prior to the demonstration,

5 for instrument stabilization. A shallow aluminum pan with gasoline (50 ml) was placed on the floor under the dash on the passenger side of the van. An additional 20 ml of gasoline were poured onto three sheets of crumpled newspaper. The newspapers were placed beneath the dash on the passenger side of the van. The gasoline pool was ignited with an ignitor. In the second test, a post-collision fire inside a van was analyzed, in order to show the propensity of materials in the passenger compartment of the conversion van to ignite, burn and propagate fire, and to investigate time available until conditions inside the vehicle became untenable under the conditions of the demonstration. The van was modified to simulate an arrangement similar to that in a specific van following a collision. The modified van was positioned on a concrete pad, at an exterior location, with wind speeds of < 8 km/hr (< 5 miles/hr). A small ignition source was placed below the dash area in the vicinity of the engine cover under the transverse HVAC duct. The exemplar van was modified by: 1) removing the front windshield, 2) removing the top portions of the rear side windows, 3) displacing the roof of the van forward so the front of the headliner was directly above the dash, 4) displacing the dash upward in the center, and 5) placing the engine cover approximately 15 mm (6 in.) back from the dash. Twenty four K-type air thermocouples were positioned inside the van: 9 on the headliner centerline and along the length of the van, 4 on a rear bench seat, 2 on the rear seat, and 9 distributed across on the top and face of the dash and in the HVAC vents. Thermocouple temperatures were recorded every 5 s. Five video cameras were used to record the demonstration: one placed to right of the rear passenger seat in the interior of the van with a view of the front passenger area and engine cover and four located outside the van. Still photographs and video were used to document the condition of the interior of the van and the location of instrumentation before and after the demonstration. A temperature baseline was established for 3 min prior to the demonstration, for instrument stabilization. The engine of the van was started and run for approximately 30 min before starting the demonstration. After stopping the engine, the fuel tank was filled with acetone and water to remove residual flammable gasoline and displace any vapors. A diffusion type burner was made from a 6 mm (0.25 in.) diameter flexible copper tube, extending outside the van, and mounted to the engine beneath the dash and engine cover. Propane gas was fed to the copper tube burner with Tygon tubing from a small cylinder. The propane gas flame was applied with a flame height of 25 mm (1 in.) from the burner surface. The propane supply was turned off once sustained burning was achieved. Eventually, the fire was manually extinguished. In the third test, a post-collision fire inside another van was analyzed, in order to investigate a postcollision fire inside the van. The van was modified to simulate an arrangement similar to that in a specific van following a collision. The modified van was positioned on a concrete pad, at an exterior location, with wind speeds of < 8 km/hr (< 5 miles/hr). A small ignition source was placed below the dash area in the vicinity of the engine cover. The exemplar van was modified by: 1) removing the front windshield, 2) removing the rear side windows, 3) displacing the roof of the van forward so the front of the headliner was directly above the dash, 4) displacing the dash upward in the center, and 5) placing the engine cover approximately 15 mm (6 in.) back from the dash. Twenty five K-type air thermocouples were positioned inside the van: 9 on the headliner centerline and along the length of the van, 2 on each bench seat (one on the seat back and one on the sitting area), 1 each on the front driver and passenger seats, and 10 distributed across on the top and face of the dash and in the HVAC vents. Thermocouple temperatures were recorded every 5 s. Six video cameras were used to record the demonstration: one was placed on

6 the rear bench seat and oriented to view the front of the van and the headliner, one was placed behind the driver s seat and to the right of it, with a view of the dash and front passenger area and four located outside the van. Still photographs and video were used to document the condition of the interior of the van and the location of instrumentation before and after the demonstration. A temperature baseline was established for 3 min prior to the demonstration, for instrument stabilization. The engine of the van was started and run for approximately 30 min before starting the demonstration. After stopping the engine, the fuel tank was filled with acetone and water to remove residual flammable gasoline and displace any vapors. A diffusion type burner was made from a 6 mm (0.25 in.) diameter flexible copper tube, extending outside the van, and mounted to the engine beneath the dash and engine cover. Propane gas was fed to the copper tube burner with Tygon tubing from a small cylinder. The propane gas flame was applied with a flame height of 25 mm (1 in.) from the burner surface. The propane supply was turned off once sustained burning was achieved. Eventually, the fire was manually extinguished. Results The major qualitative results of the real-scale car tests are indicated below, with the time lines of events shown following each description. Test 1: The temperature recorded at the headliner near the windshield rapidly increased to a maximum temperature of 782 C (1439 F) at 200 s after ignition while the back portion of the front bench seat reached a maximum temperature of 446 C (835 F) at 340 s after ignition. The temperature profiles of the thermocouples in the HVAC vents show that fire spread through the central HVAC ductwork traversing the passenger compartment. Examination of the interior of the van after fire extinguishment showed that all combustible materials, including plastic dash components, HVAC duct, carpeting, seat fabric, door panels and the headliner, were damaged in the fire. The fire damage on the passenger door panel and seat was more extensive than the damage on the driver door and seat. The fabric on the exposed surfaces of the bench seats was burned and the exposed foam decomposed. The plastic components of the dash on the passenger side were totally consumed in the fire. The driver side dash components, including the instrument panel, were consumed or exhibited severe melting and charring. Time line (min: s) Event 0:00 Ignition of gasoline inside the van. 0:42 Flames are visible inside the center HVAC duct. 0:50 Smoke begins to vent from the two HVAC vents on the top and in the center of the dash. 1:52 Passenger compartment fills with smoke. 2:00 Flames emerge from HVAC vent on the face of the dash on the passenger side. Underneath the passenger dash is fully involved. 2:50 Smoke begins to vent from the air supply vents directly in front of the windshield on the exterior of the van. 3:10 Flames emerge from passenger side window.

7 3:40 Front windshield compromised. 4:00 Passenger compartment fully involved. 5:30 Van fire extinguished manually. Test 2: The temperature on the headliner directly above the dash reached a temperature of 699 C (1291 F) at 230 s after ignition. The headliner thermocouple temperature data indicates that the fire spread from the front to the rear of the van in approximately 30 s, once the headliner became involved in the fire. The passenger compartment of the van was already fully involved approximately 160 s after the start of the demonstration. Time line (min: s) Event 0:00 Ignition. 1:56 Dash fire. 2:17 Fire from dash impinges on headliner. Headliner dripping. 2:40 Front portion of van fully involved. 2:54 Rear bench seat in flames. 3:03 Fire emerges from rear side windows. 3:20 Side window on driver s side compromised from heat. 3:27 Side door windows compromised from heat. 3:44 Van fire extinguished manually. Test 3: The temperature on the headliner directly above the dash reached a temperature of 862 C (1584 F) at 335 s after ignition while the back and seat portions of the front bench seat reached a maximum temperature of 866 C (1590 F) at 380 s after ignition. The temperature on the passenger side edge of the dash reached a maximum of 460 C (860 F), at 335 s, and that on the HVAC vent, under the dash on the driver side, reached a maximum of 537 C (999 F) at 360 s after ignition. The headliner thermocouple temperature data indicates that the fire spread from the front to the rear of the van in approximately s, once the headliner became involved in the fire. The passenger compartment of the van was already fully involved approximately 5 min after the start of the demonstration. Time line (min: s) Event 0:00 Ignition. 2:00 Smoke emerges from passenger side HVAC vent. 3:30 Fire grows under dash and emerges from passenger side HVAC vent and out of space between engine cover and dash. 3:50 Top of the dash in flames. 4:20 Fire from dash impinges on headliner. Headliner debris falls from roof. 4:30 Dash fully involved. 4:40 Headliner on fire. 5:00 Front passenger seat in flames. 5:10 Flames out the side rear window space. Van fully involved.

8 5:34 Side door windows break. 5:45 Van fire extinguished manually. FMVSS 302 Tests: Over 70 different materials, from 7 cars, were tested and all but 2 passed the test. Materials tested included the following: * air conditioning distribution HVAC duct * carpet (nap up and nap down) * carpet foam underlay * carpet underlay * ceiling insulation * dash panel * dashboard * door panel * door panel armrest * door panel foam * floor padding * foam padding at the base of windshield * glove box * glove box and radio housing * headliner * heater core and evaporator housing (including a fiber reinforced one) * HVAC flexible hose * HVAC foam * instrument panel housing above steering column * seat fabric * seat foam * weather strip * wiring harness Vertical demonstrations: All of the under-dash and dash components tested in the vertical orientation were shown to 1) melt, drip and burn in a pool fire in the bottom of the test chamber; 2) exhibit rapid flame spread to the top of the test sample; 3) produce large flame heights indicative of high rates of heat release and high heats of combustion; and 4) continue to burn after the removal of the small initiating flame. Similarly, the HVAC ducts, foams and housings, headliners and heater housings tested were all rapidly consumed. For comparison purposes, a properly fire retarded, glass-filled polypropylene material was tested under the same conditions and it ceased burning once the flame was removed, without generating flaming droplets. It needs to be emphasized that all the materials that burn vigorously in the vertical orientation. Cone calorimeter tests: Much of the cone calorimeter test results information (particularly on 4 of the cars) has been presented elsewhere [11, 12]. Appendices A through F present once more some information on cone calorimeter test results for cars 1 through 4, in context with those for cars 5 and 6.

9 Appendix G presents similar information for car 7, while appendix H presents cone calorimeter test data for the two components of the engine cover of car 7 (with Figure 1 showing the heat release rate of the engine cover molding at 25 and 40 kw/m 2 ). Finally, Appendix 8 presents data on cone calorimeter testing of a variety of fire retarded polypropylene materials, for comparison with the car materials, with Figure 2 showing a comparison between the polyolefin HVAC duct from car 3 and a fire retarded polypropylene material. Discussion The information presented expands on the analysis conducted earlier, that showed that car interior materials exhibit poorer fire performance than average plastics. The most interesting issue presented earlier [11, 12] was that the median fire test data from the tests on automotive materials was much poorer than that of commercial plastic materials of the same vintage, in virtually all aspects of fire performance. Furthermore, car seats perform as poorly (or worse) than domestic fabric-foam seat composites, using non fire retarded foams. In fact, there are countries where such padding materials would not be permitted for use even in homes, e.g. the United Kingdom [14]. Real-scale tests: In all 3 real-scale tests conducted, fires inside the passenger compartment consumed virtually all the combustible materials, leaving a rusted interior with seat frames and springs and the metal frame of the dash. They also burnt off the vehicle paint. Considering that human tenability ceases when temperatures reach 60 C (140 F), heat fluxes reach 20 kw/m 2 and smoke layers get to 1.2 m from the ceiling, this happened no later than 1 min 52 s in test 1 (passenger compartment filled with smoke), or than 2 min 40 s in test 2 (front portion of vehicle fully involved) or than 4 min 40 s in test 3 (after dash is fully involved in fire, the headliner catches fire), so that clearly the vehicle interior became rapidly untenable in all cases. Figures 3 and 4 show traces of temperatures in the headliner, duct and front car seat, illustrating how rapidly high temperatures were reached. Thus, a vehicle occupant who may still be conscious, but is likely to be stunned or otherwise injured, has very little time left to exit or be rescued before receiving fatal injuries as a result of the fire. Such time available for escape or rescue could clearly be increased if the fire performance of the materials in the passenger compartment were improved, for example by better fire retardance [15]. It has now been shown that the cone calorimeter is a useful tool to predict heat release rate data in real scale for a variety of products (furnishings and contents), including interior lining materials and upholstery [16-17]. The information obtained is then useful to investigate material fire safety in a variety of environments, including a passenger car compartment. Engine cover: The engine cover should offer a high degree of protection so that ignition, if it occurs at all, is delayed for very long periods and a fire does not penetrate from the engine compartment into the passenger compartment. Thus, it is interesting and unfortunate that the molded fiber reinforced plastic material studied in car 7 exhibited fairly poor fire performance. The molded plastic material ignited at approximately 2 min at an incident heat flux of 25 kw/m 2, and at approximately 1 min at an incident heat

10 flux of 40 kw/m 2, with a high peak rate of heat release, close to 300 kw/m 2 (Figure 1, Appendix H). The material, which should be offering a barrier to penetration of fire from the engine compartment (i.e. a fire wall), suffers extensive mass loss at relatively low heat exposures. This offers a simple passageway for flames from the engine compartment to enter the passenger compartment and result in a severe fire that traps the passengers, as they are often injured as a result of collision or perhaps just stunned, and have lower mobility. Interestingly, an analysis of the fire performance of the plastic molded material in the engine cover by using the Karlsson inequality calculation [16, 18] indicates that it is likely to cause flashover in the compartment (i.e. the passenger compartment). Flashover resulting from the engine cover molded plastic may not be assured, but the analysis clearly indicates that the fire performance of such a material is poor for use as a fire barrier in an engine cover. Moreover, observations on several cars have shown that the engine cover often contains several penetrations sealed with elastomeric materials which are neither suitable "firestop" materials for use in a fire wall nor even exhibit suitable fire performance. HVAC ducting: The fire performance of the HVAC duct materials has been tested in every one of the 7 vehicles investigated. The conclusions reached from earlier analyses [11-12] are clearly still valid: the HVAC duct materials provide very poor fire performance, and one that can easily be improved by the use of existing fire retarded polyolefin materials (see Appendix I). Such materials are also surrounded by a large mass of other combustibles, most of which are easily ignitable. Thus, they can cause an untenable situation within a very short time. An analysis by using the Karlsson inequality [18] indicates that most of those HVAC duct materials (with their very high rates of heat release at 40 kw/m 2 incident heat flux) are likely to cause a self-propagating fire when exposed in a compartment ceiling. There have been alternative materials on the market for years, made from other polymers, including those from the survey of 35 materials discussed earlier [13]. Headliners: Vehicle headliners in vehicles are typically coated fabrics, with a thin covering layer and a back coating (often a foam), perhaps mounted on plywood or fiberglass. This acts, of course, as the interior ceiling finish of the vehicle s passenger compartment. In many fire scenarios, a fire safety objective is to control he fire before it ignites the ceiling, and then prevent flaming drips from occurring due to the ceiling finish. Headliners from four cars were tested, and their times to ignition ranged from 9 to 62 s, at an incident heat flux of 25 kw/m 2 ; these headliners clearly offered little protective escape time! As a matter of interest, in each of the three real-scale fires conducted, headliner temperatures quickly reached values that correspond to well over 50 kw/m 2 incident heat fluxes (approximately 695 C), so clearly ignition of the headliner would have resulted. It is interesting to note the consistently poor average of the peak heat release rate values of the cars tested, at the 25 kw/m 2 incident heat flux, indicative that the problem is not one type of car. Car kw/m 2 Car kw/m 2 Car kw/m 2 Car kw/m 2 Car kw/m 2 Car kw/m 2

11 Car kw/m 2 The results found and the interpretation are consistent with the findings of Factory Mutual Research Corp. in a study of the flammability of combustible materials in a van [19], which was conducted for a joint project managed by a car manufacturer and the US federal government. Conclusions * The fire performance of combustible materials used in passenger vehicle interiors is poor. * Real-scale fire tests show that vehicle fires lead to untenable conditions very rapidly. * Any simple analysis confirms the finding of real-scale fire tests on untenability. * Passenger vehicle occupants have insufficient time available for escape or rescue in a fire. * Fire test requirements for passenger vehicle interior materials should be changed so that materials with improved fire performance are used. * Materials with lower heat release rates or better ignitability would result in car fires of lower intensity and with a greater probability of escape or rescue for trapped occupants. References 1. Home Office, United Kingdom, Fire Statistics - United Kingdom through 1999". 2. Ahrens, M., "US Vehicle Fire Trends and Patterns", Natl Fire Protection Assoc., Quincy, MA, August Fire Loss Statistics in the United States, September issue of NFPA Journal, Ontario Fire Losses by Property Class to 1999", Office of the Ontario Province Fire Marshal. 5. FMVSS 302, "Motor Vehicle Safety Standard No. 302, Flammability of Materials - Passenger Cars, Multipurpose Passenger Vehicles, Trucks and Buses", National Highway Traffic Safety Administration, Washington, DC. [Code of Federal Regulations , originally Federal Register 34, No. 229, pp (December 31, 1969)]. 6. ISO : Road vehicles, and tractors and machinery for agriculture and forestry Determination of burning behaviour of interior materials. 7. Committee on Fire Safety Aspects of Polymeric Materials, Fire Safety Aspects of Polymeric Materials, Volume 8 Land Transportation Vehicles, National Materials Advisory Board, National Academy of Sciences, Publication NMAB 318-8, Washington DC, 1979, pp 77-88,

12 8. Shields, L., Scheibe, R. and Angelos, T. (Washington State Transportation Center), Motor- Vehicle Collision-Fire Analysis: Methods and Results, Natl Fire Protection Assoc., Quincy, MA, Presentation at NFPA Fall Meeting, Atlanta, GA, November 17, ASTM E 1354, "Standard Test Method for Heat and Visible Smoke Release Rates for Materials and Products Using an Oxygen Consumption Calorimeter (cone calorimeter)", Vol , Amer. Soc. Testing & Mater., West Conshohocken., PA. 10. ISO 5660, "Fire Tests - Reaction to Fire - Rate of Heat Release from Building Products (Cone Calorimeter Method), Intern. Organization for Standardization, Geneva, Switzerland. 11. Hirschler, M.M., "Fire Hazard of Automotive Interiors", Fire Risk & Hazard Assessment Symposium, National Fire Protection Research Foundation, June 24-26, 1998, San Francisco, CA, pp Grayson, S.J. and Hirschler, M.M., Fire Performance of Plastics in Car Interiors, in Proc. Flame Retardants 2002, Interscience Communications, London, UK, Feb. 5-6, 2002, pp Hirschler, M.M., "Heat release from plastic materials", Chapter 12 a, in "Heat Release in Fires", Elsevier, London, UK, Eds. V. Babrauskas and S.J. Grayson, pp UK Government Consumer Safety Research, "Effectiveness of the Furniture and Furnishings (Fire) (Safety) Regulations 1988", Consumer Affairs Directorate, Dept. Trade and Industry, London, UK, June 2000 [Research conducted by Professor Gary Stevens, Univ. of Surrey, Guildford, UK]. 15. Hirschler, M.M., "Fire Retardance, Smoke Toxicity and Fire Hazard", in Proc. Flame Retardants '94, British Plastics Federation Editor, Interscience Communications, London, UK, Jan , 1994, pp (1994). 16. Hirschler, M.M., "Use of Heat Release Rate to Predict Whether Individual Furnishings Would Cause Self Propagating Fires", Fire Safety J., 32, (1999). 17. Janssens, M.L., Dillon, S.E., and Hirschler, M.M., "Using the Cone Calorimeter as a Screening Tool for the NFPA 265 and NFPA 286 Room Test Procedures", Fire and Materials Conf., San Francisco, CA, Jan , 2001, Interscience Communications, London, UK, pp Karlsson, B., "Models for Calculating Flame Spread on Wall Lining Materials and the Resulting Heat Release Rate in a Room", Fire Safety Journal, 23(4), (1994). 19. Tewarson, A., A Study of the Flammability of Plastics in Vehicle Components and Parts, Factory Mutual Research Corp. Technical Report FMRC J.I. 0B1R7.RC, on General Motors Corp. Research Project, October 1997.

13 Table 1 - Fire Fatalities in Various Regions: United States, United Kingdom and Ontario Province (Canada) US Fire Fatalities UK Fire Fatalities Ontario (Canada) Fire Fatalities Year Dwellings Road Vehicles All Dwellings Road Vehicles All Dwellings Road Vehicles All # % # % # # % # % # # % # % # , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,147 Average 3, , Totals 62,395 6,886 79,745 8,999 1,263 11,891 1, ,345 Table 2- US Passenger Road Vehicle Fire Problem ( Annual Averages) Vehicle Type Fires Fire Fatalities # % # % Automobile 280, Motor Home 2, All Terrain Vehicle 2, Bus or Trackless Trolley 1, Travel Trailer Mobile Home or Mobile Building Camping Trailer Unclassified or Unknown Passenger Vehicle 6, Total 295,

14 Appendix A. Testing of Car 1 Materials - Heat Fluxes of 25 and 40 kw/m 2 Units Seat Foam 1 Fabric 1 Floor covering Headliner Dash a Door cover Duct Dash b Dash c Dash 25 * Pk RHR kw/m TTI s THR MJ/m FPI s m 2 /kw Ht Comb MJ/kg Avg RHR kw/m Avg RHR 3 kw/m Pk RHR kw/m TTI s THR MJ/m FPI s m2/kw Ht Comb MJ/kg Avg RHR kw/m Avg RHR 3 kw/m * Seat assembly (fabric/foam system composite) was tested at 35 kw/m 2

15 Appendix B. Testing of Car 2 Materials - Heat Flux of 25 kw/m 2 Units Seat a Seat b Foam Fabric Door Composite (Fabric-Foam) Carpet Headliner Pk RHR kw/m TTI s THR MJ/m FPI s m 2 /kw SEA m 2 /kg MsLs % TTE s Ht Comb MJ/kg Avg RHR kw/m Avg RHR 3 kw/m SmkFct MW/m TSR Pk RSR 1/s Avg MLR g/s Units Gray Dash Foam Foam Insulation Plastic Molding Duct Dash Dash Backing Dash + Backing Pk RHR kw/m TTI s THR MJ/m FPI s m 2 /kw SEA m 2 /kg MsLs % TTE s Ht Comb MJ/kg Avg RHR kw/m Avg RHR 3 kw/m SmkFct MW/m TSR Pk RSR 1/s Avg MLR g/s * Seat a and seat b are different seat fabric/foam assembly composites, at 25 kw/m 2

16 Units Appendix C. Testing of Car 3 Materials - Heat Flux of 25 kw/m 2 HVAC Shroud Dash Speaker Foam Side Panel Sidewall Cover Door Cover Carpet Interior Insulat. Pk RHR kw/m TTI s THR MJ/m FPI s m 2 /kw SEA m 2 /kg MsLs % TTE s Ht Comb MJ/kg Avg RHR kw/m Avg RHR 3 kw/m SmkFct MW/m TSR (-) Pk RSR 1/s Avg MLR g/s Appendix D. Testing of Car 4 Materials - Heat Flux of 25 and 40 kw/m 2 Units Dash Up Dash Down Dash Frame Up Dash Frame Down Duct Vent Duct Vent * Foam Pk RHR kw/m TTI s THR MJ/m FPI s m 2 /kw SEA m 2 /kg MsLs % TTE s Ht Comb MJ/kg Avg RHR kw/m Avg RHR 3 kw/m SmkFct MW/m TSR (-) Pk RSR 1/s Avg MLR g/s * Column with Duct vent 40 kw/m 2

17 Appendix E. Testing of Car 5 Materials - Heat Fluxes of 25 and 40 kw/m 2 Units A/C Dist Duct to Vents Dashboard Core Housing 25 Pk RHR kw/m TTI s THR MJ/m FPI s m 2 /kw SEA m 2 /kg MsLs % TTE s Ht Comb MJ/kg Avg RHR 3 kw/m TSR (-) Pk RHR kw/m TTI s THR MJ/m FPI s m 2 /kw SEA m 2 /kg MsLs % TTE s Ht Comb MJ/kg Avg RHR 3 kw/m TSR (-) Appendix F. Testing of Car 6 Materials - Heat Flux of 25 kw/m2 Units Lower Heater Housing Heater Housing Fiber Reinforced HVAC Duct Rubber Cellulosic Pad Upper Heater Housing Pk RHR kw/m TTI s THR MJ/m FPI s m 2 /kw SEA m 2 /kg MsLs % TTE s Ht Comb MJ/kg Avg RHR 3 kw/m TSR (-)

18 Appendix G. Testing of Car 7 Materials - Heat Fluxes of 25 and 40 kw/m 2 CONE CALORIMETER DATA AT 25 kw/m 2 Tig(s) RHR Max (kw/m 2 ) THR (MJ/m 2 ) FPI (s m 2 /kw) RHR 3 min (kw/m 2 ) Mass loss (g/%) EHC (MJ/kg) Dash / Dashboard Back Wrap / Dashboard Back Foam / Door Panel / Floor Covering / Glove Box / Glove Box / Head Liner / HVAC Box / HVAC Duct / HVAC Duct / Seat Fabric / Upper Dash Cover / Median CONE CALORIMETER DATA AT 40 kw/m 2 Tig(s) RHR Max (kw/m 2 ) THR (MJ/m 2 ) FPI (s m 2 /kw) RHR 3 min (kw/m 2 ) Mass loss (g/%) EHC (MJ/kg) Dash / Dashboard Back Wrap / Dashboard Back Foam / Door Panel / Floor Covering / Glove Box / Glove Box / Head Liner / HVAC Box / HVAC Duct / HVAC Duct / Seat Fabric / Upper Dash Cover / Median

19 Appendix H. Testing of Car 7 Engine Cover - Heat Fluxes of 25 and 40 kw/m 2 Tig(s) RHR Max (kw/m 2 ) THR (MJ/m 2 ) FPI (s m 2 /kw) RHR 3 min (kw/m 2 ) Mass loss (g/%) EHC (MJ/kg) 25 kw/m 2 Engine cover insulation 10, , / Engine cover insulation 10, , / Engine cover insulation 10, , / Average 10, , / Engine cover molding / Engine cover molding / Engine cover molding / Engine cover molding 100 Average / kw/m 2 Engine cover insulation 10, , / Engine cover insulation 10, , / Engine cover insulation 10, / Average 10, , / Engine cover molding / Engine cover molding / Engine cover molding / Average / Notes: RHR 3 min, for Engine cover insulation, starts at beginning of test and not at ignition, since ignition did not occur. EHC average for Engine cover insulation, at 40 kw/m 2 uses only two values for the average, because one value (20.5 MJ/kg) is obviously wrong.

20 Appendix I. Cone Calorimeter Tests on Fire Retarded Polypropylene Specimens Part A: Tests at 20 kw/m 2 Material Time to Ignition Pk RHR FPI Eff. Ht Comb. Avg. RHR 3 Mass Loss % Unit s kw/m 2 s m 2 /kw MJ/kg kw/m 2 % Part B: Tests at 40 kw/m 2 Material Time to Ignition Pk RHR FPI Eff. Ht Comb. Avg. RHR 3 Mass Loss % Unit s kw/m 2 s m 2 /kw MJ/kg kw/m 2 %

21 Figure 1: Heat Release Rate of Engine Cover and 40 kw/m2) RHR (kw/m^2) time (s) 40 25

22 Figure 2: Cone Data for Duct (car 3) & FR Polyproylene time (s) Car 3 Duct FR Polypropylene

23 Figure 3: Temperatures in Real Scale Car Fire Demonstrations (Headliner and Duct) Temperature (deg C) time (s) Demo 1 Headliner Demo 1 Duct Demo 2 Headliner Demo 2 Duct Demo 3 Headliner Demo 3 Duct

24 Figure 4 - Front Seat Temperatures in Real Scale Car Fire Demonstrations Temperature (deg C) time (s) Demo 1 Demo 2 Demo 3