Contents
The fire propagation apparatus (FPA) is a bench-scale fire test instrument that measures how easily a material ignites, how much heat and smoke it produces and whether flames spread over its surface. Archibald Tewarson developed it at FM Global.[1][2] A specimen about 100 mm across sits inside a quartz tube and is heated by infrared heaters, while a controlled mixture of oxygen and nitrogen flows past it. The combustion products are collected in a hood, and the heat release rate is found from the gas composition in the exhaust.[3]
Because the oxygen concentration can be set between 0 and 60 %, the apparatus can simulate fires that are short of air, normal fires and the strongly radiating flames of large fires.[3] Oxygen levels below ambient are used to simulate ventilation-controlled fires.[3] In the United States, the methods are standardised as ASTM E2058 (2000) and, for materials in cleanrooms, as NFPA 287 (2001).[2] Results include the critical heat flux, the thermal response parameter, the heat of gasification and the fire propagation index (FPI). The FPI is used to approve cables, conveyor belts and cleanroom materials for semiconductor plants.[4] It is also called the FMRC flammability apparatus or the Tewarson fire propagation test apparatus.[5][2]
Definition and development
In the 2008 Fire Protection Handbook, Madrzykowski and Stroup describe the fire propagation apparatus as a small-scale flammability apparatus for measuring heat release rate. Tewarson first described it at FM Global in 1975.[2] Hirschler describes the instrument as one used for assessing heat release and fire propagation, developed in parallel with the Ohio State University (OSU) heat release rate calorimeter.[1] In the SFPE Handbook, Janssens names it the Factory Mutual Research Corporation (FMRC) fire propagation apparatus. It was first developed to measure the convective heat release rate and the generation rates of smoke and combustion products.[3] The handbook and SFPE descriptions agree on its purpose and differ only in some dimensions, as noted below.
The early apparatus obtained the convective part of the heat release rate from the enthalpy rise of the exhaust gases. The total heat release rate came from the mass loss rate, the heat of combustion measured in an oxygen bomb and the generation of carbon dioxide and carbon monoxide.[2] Test results reported since the late 1970s also give total heat release rates from oxygen consumption or carbon dioxide generation.[3] In the 1980s the apparatus was built in industrial laboratories in France, Germany and the United States.[3] Tewarson used it to determine fire hazard indices and material properties for fire modelling. He also studied how the oxygen concentration of the combustion air affects heat release rate and burning behaviour.[3]
A larger version, the intermediate-scale flammability apparatus, could test samples up to 305 mm in diameter and 75 mm thick. It handled heat release rates up to 500 kW, whereas the small apparatus was limited to peaks of 10 kW.[2] Test methods based on the small apparatus were adopted in 2000 as ASTM E2058. In 2001 the same apparatus, with smoke and corrosion measurements added, was adopted as NFPA 287 for materials in commercial cleanrooms.[2] FM Global Research thus developed a small-scale calorimeter that has since been standardised for use with materials in clean rooms.[6]
Apparatus
The tests take place in a semi-open environment. The specimen sits inside a quartz tube 0.61 m long and 0.17 m in diameter. A mixture of oxygen and nitrogen is supplied at the bottom of the tube, and a stainless steel funnel and vertical exhaust duct are placed some distance above it.[3] Dilution air is drawn in between the tube and the exhaust. The total gas flow in the duct is measured from the pressure drop across a calibrated orifice.[3] The specimen rests on a load cell, so that its mass loss is recorded throughout the test.[7]
Four infrared heaters are placed outside the quartz tube, coaxially around the specimen.[3] Tewarson lists them as tungsten-quartz heaters giving an external heat flux of 0 to 65 kW/m², the same maximum as Madrzykowski and Stroup give. Janssens gives a maximum of 70 kW/m².[8][2][3] A small pilot flame about 10 mm above the centre of the specimen ignites the vapours.[3] The gas supply to the tube can be set between 0 and 8.3 L/s, with an oxygen content between 0 and 60 %.[3] Tewarson's table gives a co-flow gas velocity of 0 to 0.146 m/s and an exhaust flow of 0.035 to 0.364 m³/s.[8]
Horizontal specimens for ignition and combustion tests measure about 100 mm across. Janssens and Madrzykowski and Stroup describe a circular specimen 0.1 m in diameter and at most 50 mm thick.[3][2] Tewarson lists 100 × 100 mm horizontal and 100 × 600 mm vertical specimens, and a heat release rate capacity of 50 kW.[8] Besides heat release rate, the apparatus measures mass loss, smoke obscuration, smoke corrosivity and the concentrations of soot and gases in the exhaust.[3] Its smoke measurement uses a helium-neon laser beam across the exhaust duct.[9] Tewarson's comparison table shows that, unlike the OSU apparatus and the cone calorimeter, it also measures fire propagation rate, gas-phase corrosion, radiative heat release rate and flame extinction by water and halon.[8]
Test procedures
The protocol consists of three small-scale tests: ignition, combustion and fire propagation.[4] In the ignition test, a horizontal specimen is exposed to a known heat flux and the time to ignition is measured. From ignition times at several heat fluxes, the method yields the minimum heat flux for ignition. It also yields the product of the square root of the thermal inertia and the rise of the ignition temperature above ambient.[10]
The critical heat flux for ignition is the heat flux at or below which a material cannot generate a combustible mixture. The thermal response parameter (TRP) expresses the resistance of a material to generating such a mixture.[11] To measure the critical heat flux, a blackened horizontal sample is exposed to various heat fluxes until a value is found at which it does not ignite within about 15 minutes.[12]
The combustion test measures the chemical and convective heat release rate at an incident heat flux of 50 kW/m².[4] Chemical heat is the heat generated by the combustion reactions. It splits into a convective part carried by the hot gases and a radiative part emitted by the flame.[11] Hirschler notes that the apparatus assesses heat release by measuring carbon dioxide generation. Janssens states that total heat release rate is found by the oxygen consumption principle. He adds that the carbon dioxide method is less widely accepted, because the energy per unit of product generated depends more on the fuel.[4][3]
In a third mode, the oxygen is replaced by pure nitrogen, so that the specimen gasifies without flaming. The mass loss rate is then measured at several heat fluxes to obtain the heat of gasification, the energy needed to turn the solid into vapour.[13] These nonflaming tests use specimens about 100 mm square and up to 100 mm thick, in nitrogen or in air with an oxygen mass fraction of about 0.10.[14] Water evaporation tests are routinely used to calibrate the apparatus. Water has a heat of gasification of 2.58 kJ/g, and at 50 kW/m² its predicted evaporation rate of 19.0 g/m²s agrees closely with the measurement.[14]
The fire propagation test uses a vertical specimen. Hirschler gives its size as 300 mm high, 100 mm wide and no more than 13 mm thick.[4] Tewarson describes samples 300 to 600 mm long and up to about 100 mm wide and thick.[15] The bottom 120 to 200 mm is exposed to 50 kW/m² with a pilot flame. Above this ignition zone, upward flame spread continues by itself.[15] The co-flowing gas has an oxygen mass fraction of 0.40, which creates the strongly radiating flames typical of large fires. The chemical heat release rate is recorded over time.[15]
Governing equations and worked example
For a thermally thick material, the time to ignition depends on how quickly the surface heats up, which is set by the ignition temperature and the thermal properties. Tewarson combines these into the thermal response parameter.[12]
The ignition theory then predicts that the inverse square root of the ignition time rises in a straight line with the external heat flux, starting from the critical heat flux.[12] In the FPA, TRP is found by a linear regression of ignition data at several heat fluxes. TRP is the inverse of the slope of that line.[7]
For fire propagation, Tewarson defines the fire propagation index from the chemical heat release rate per unit width of the burning specimen and the TRP.[15]
| Symbol | Description | Unit |
|---|---|---|
| TRP | thermal response parameter | kW·s1/2/m² |
| ignition temperature minus ambient temperature | K | |
| , , | thermal conductivity, density and specific heat of the material | kW/(m·K), kg/m³, kJ/(kg·K) |
| time to ignition | s | |
| external heat flux | kW/m² | |
| CHF | critical heat flux for ignition | kW/m² |
| FPI | fire propagation index | – |
| chemical heat release rate per unit width of the specimen | kW/m |
The units of , and follow NUREG-1805.[16] The ignition equation applies to thermally thick materials, which most common materials are, and to the heating period before a steady state is reached.[7][5]
Worked example. Tewarson reports ignition times for a crosslinked polyolefin communications cable (XLPO/XLPO) in the FPA. They are 716 s at 40 kW/m², 318 s at 50 kW/m², 179 s at 60 kW/m² and 45 s at 100 kW/m².[17] The inverse square roots of these times are 0.0374, 0.0561, 0.0747 and 0.1491 s−1/2.
A linear regression against the heat flux gives a slope of 0.001861 and an intercept of −0.0370, with a coefficient of determination of 1.000. TRP is the inverse of the slope, 537 kW·s1/2/m², close to the 535 that Tewarson reports.[17] The line reaches zero at a heat flux of 19.9 kW/m², which is the extrapolated critical heat flux.[7]
As a check, the equation with these values predicts (537/(50 − 19.9))² = 318 s at 50 kW/m², the measured value. The fire propagation index then follows from the upward propagation test. Tewarson's own example takes a 0.3 m wide vertical sheet with a TRP of 95 kW·s1/2/m² that releases 50 kW of chemical heat.[17] The heat release rate per unit width is 50/0.3 = 167 kW/m, whose cube root is 5.50. The index is 750 × 5.50/95 = 43, well above 20, so the material belongs to Group 3, with accelerating fire propagation.[15][17] For the XLPO/XLPO cable, Tewarson reports an FPI of 4, which places it in the nonpropagating group.[17]
Typical values
Tewarson tabulates critical heat fluxes and thermal response parameters measured in the ASTM E2058 apparatus for a wide range of materials.[18] A selection is given below. A higher critical heat flux and a higher TRP mean that a material takes longer to heat up and ignite, and that fire propagates more slowly over it.[11]
| Material | CHF (kW/m²) | TRP (kW·s1/2/m²) |
|---|---|---|
| Newspaper | 10 | 108 |
| Wood (red oak) | 10 | 134 |
| Wood (Douglas fir) | 10 | 138 |
| Wood (Douglas fir, fire-retardant) | 10 | 251 |
| Polystyrene | 13 | 162 |
| Polypropylene | 15 | 193 |
| Polymethylmethacrylate (PMMA) | 11 | 274 |
| Polyethylene (high density) | 15 | 321 |
| Polyvinylchloride (PVC) | 10 | 194 |
| Rigid PVC (sample 1) | 15 | 406 |
| Heavy corrugated paper, no coating | 10 | 189 |
| Heavy corrugated paper, 20 % coating by weight | 15 | 714 |
The values depend on the specimen and its preparation. For PMMA, TRP is 274 kW·s1/2/m² with a blackened surface and 383 without.[7] For heavy corrugated paper, TRP rises linearly with the weight percentage of a surface coating, following TRP = 24.5w + 181.[18] For polypropylene, cone calorimeter data give a TRP of 291, against 193 in the FPA. The cone specimens were not blackened, which Tewarson notes may give somewhat higher TRP values.[18][7]
Tewarson uses the FPI to divide materials into four groups. The table also shows the threshold given in the 2008 Fire Protection Handbook.[15][4]
| FPI | Group (Tewarson) | Fire propagation beyond the ignition zone |
|---|---|---|
| below 7 | N-1, nonpropagating | none; the flame is at critical extinction |
| 7 to 10 | D-1, decelerating | limited, slowing down |
| 10 to 20 | Group 2, nonaccelerating | slow |
| above 20 | Group 3, accelerating | rapid |
Hirschler (2008a) states that experience indicates nonpropagating materials have FPI values at or below 6.0.[4] Measured values for power cables range from 6 to 8 for some ethylene-propylene rubber cables to 11 to 28 for PVC/PVC cables. Communications cables range from 4 to 5 for FEP/FEP cables to 36 for PVC/PVC cables.[17]
Applications
The best-known application is the approval of materials for cleanrooms. In a semiconductor fabrication facility, devices are made in cleanrooms, and a large facility can cost $2–3 billion to build and equip.[19] In the United States, as described in the 2008 Fire Protection Handbook, fab cleanroom materials are approved by testing to UL 2360 or to NFPA 287 with FM 4910 criteria.[20] Research at FM Global has shown that materials with a low FPI in ASTM E2058 or NFPA 287 are likely to perform well in the larger parallel panel test.[4] That test exposes two panels 2.4 m high and 0.6 m wide to a 60 kW sand burner for 20 minutes. It now qualifies materials for semiconductor cleanrooms under the ANSI FM 4910 protocol.[21]
FM Global developed three approval standards that use the apparatus: FM 3972 for electric cables, FM 4910 for cleanroom materials and FM 4998 for coal mine conveyor belts.[4] For cable fire protection, the FMRC cable standard classifies cables by their upward fire propagation under strongly radiating conditions, for use in noncombustible occupancies. As described by Tewarson (2002), FMRC recommends no additional protection for Group 1 cables in noncombustible occupancies with noncombustible construction. Group 2 cables can be used there without it under certain conditions, and Group 3 cables need fire protection.[15]
The measured properties are also used as input for fire calculations. NUREG-1805 was written for the fire protection inspection programme of the U.S. Nuclear Regulatory Commission. Its method of Tewarson uses critical heat fluxes and TRP values derived from the FMRC flammability apparatus to estimate ignition times.[5] The International Fire Engineering Guidelines (2005) refer to the same method and its data table.[22] In Australia, the Fire Safety Verification Method data sheets (2020) list soot yields and chemical heats of combustion from ASTM E2058 tests, supplemented by cone calorimeter data.[23] The critical heat flux of cladding and roofing materials can be taken from the cone calorimeter or the fire propagation apparatus when exposure from a fire outside a building is assessed.[24]
The heat of gasification measured in nitrogen helps rank materials. The ratio of the heat of combustion to the heat of gasification, called the heat release parameter by Tewarson, controls the heat release rate. Materials with a very low flammability hazard have values near 1, while more hazardous materials exceed 10.[25] The apparatus also gives smoke yields, a mass of smoke per unit mass of material burned, from its smoke generation rate.[9]
Limitations and uncertainty
The use of 40 % oxygen in the fire propagation test is somewhat controversial. Some researchers question its relevance to fires in normal air with 20.9 % oxygen and to fires in vitiated air.[4] The electrical heaters run at high temperatures, so the spectrum of their radiation differs from that of most fires.[3]
Bench-scale heat release data are always apparatus dependent. Comparisons have shown good agreement between the FMRC apparatus and the cone calorimeter for aircraft panels. For communication cables, however, the FMRC apparatus gave shorter ignition times and lower heat release rates than the cone.[26] The Australian data sheets warn that correlations from small-scale tests to larger scales are limited, so small-scale data should be extrapolated with caution.[23]
The critical heat flux is not measured directly but extrapolated to the heat flux at which the ignition time becomes infinite. Its value therefore depends on the model used to correlate the data.[5] Hirschler considers extrapolated critical heat fluxes below 15 kW/m² unreliable and recommends reporting them as "≤ 15 kW/m²".[27]
The sources also differ in how the ignition equation is written. NUREG-1805 defines TRP without the factor π/4 and puts π/4 in the ignition-time equation instead. The version given by Hurley and Bukowski in the handbook also has π/4 in that equation.[16][5][28] Tewarson's own examples also match this form: newspaper at 50 kW/m² ignites in 6 s, whereas the ignition equation above gives 7.3 s.[8] The TRP values tabulated by Tewarson, and reproduced in NUREG-1805, come from the FPA regression, and with the ignition equation above they reproduce the measured times. Applying the π/4 form to the same values gives times about 21 % shorter. For the XLPO/XLPO cable at 50 kW/m², it gives 250 s instead of the measured 318 s.[17]
The sources also give different dimensions and capacities. Maximum heat fluxes of 65 and 70 kW/m² are quoted, and vertical specimens of 300 mm and of up to 600 mm. The heat release rate capacity is given as 10 kW for the small apparatus by Madrzykowski and Stroup and as 50 kW by Tewarson.[2][3][8][4] The nonpropagating limit is given as an FPI below 7 by Tewarson and as at or below 6.0 by Hirschler.[15][4]
See also
Notes
- ↑ 1.0 1.1 Hirschler (2008a), p. 6-46.
- ↑ 2.00 2.01 2.02 2.03 2.04 2.05 2.06 2.07 2.08 2.09 Madrzykowski and Stroup (2008), p. 2-38.
- ↑ 3.00 3.01 3.02 3.03 3.04 3.05 3.06 3.07 3.08 3.09 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.17 Janssens (2002), p. 3-54.
- ↑ 4.00 4.01 4.02 4.03 4.04 4.05 4.06 4.07 4.08 4.09 4.10 4.11 Hirschler (2008a), p. 6-47.
- ↑ 5.0 5.1 5.2 5.3 5.4 Iqbal and Salley (2004), p. 6-5.
- ↑ Benedetti and Colonna (2008), p. 6-228.
- ↑ 7.0 7.1 7.2 7.3 7.4 7.5 Tewarson (2002), p. 3-84.
- ↑ 8.0 8.1 8.2 8.3 8.4 8.5 Tewarson (2002), p. 3-85.
- ↑ 9.0 9.1 Madrzykowski and Stroup (2008), p. 2-44.
- ↑ Madrzykowski and Stroup (2008), p. 2-35.
- ↑ 11.0 11.1 11.2 Tewarson (2002), p. 3-82.
- ↑ 12.0 12.1 12.2 Tewarson (2002), p. 3-83.
- ↑ Madrzykowski and Stroup (2008), p. 2-32.
- ↑ 14.0 14.1 Tewarson (2002), p. 3-100.
- ↑ 15.0 15.1 15.2 15.3 15.4 15.5 15.6 15.7 Tewarson (2002), p. 3-92.
- ↑ 16.0 16.1 Iqbal and Salley (2004), p. 6-4.
- ↑ 17.0 17.1 17.2 17.3 17.4 17.5 17.6 Tewarson (2002), p. 3-93.
- ↑ 18.0 18.1 18.2 18.3 Tewarson (2002), p. 3-87.
- ↑ DeGiorgio and Peterkin (2008), p. 9-119.
- ↑ Hirschler (2008a), p. 6-48.
- ↑ Madrzykowski and Stroup (2008), p. 2-43.
- ↑ National Research Council of Canada et al. (2005), p. 2.4-5.
- ↑ 23.0 23.1 Australian Building Codes Board (2020a), Data Sheet B2, p. 19.
- ↑ Janssens (2008), p. 2-72.
- ↑ Madrzykowski and Stroup (2008), p. 2-36.
- ↑ Janssens (2002), p. 3-56.
- ↑ Hirschler (2008a), p. 6-41.
- ↑ Hurley and Bukowski (2008), p. 3-126.
References
- Australian Building Codes Board. (2020a). Data sheets: Fire Safety Verification Method (Handbook annex, Print version 1.1).
- Benedetti, R. P., & Colonna, G. R. (2008). Polymeric materials. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-209–6-231). National Fire Protection Association.
- DeGiorgio, V., & Peterkin, H. (2008). Semiconductor manufacturing. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 9-119–9-129). National Fire Protection Association.
- Hirschler, M. M. (2008a). Concepts and protocols of fire testing. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-35–6-60). National Fire Protection Association.
- Hurley, M. J., & Bukowski, R. W. (2008). Fire hazard analysis techniques. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-121–3-134). National Fire Protection Association.
- Iqbal, N., & Salley, M. H. (2004). Fire Dynamics Tools (FDTs): Quantitative fire hazard analysis methods for the U.S. Nuclear Regulatory Commission Fire Protection Inspection Program (NUREG-1805). U.S. Nuclear Regulatory Commission, Office of Nuclear Reactor Regulation.
- Janssens, M. (2002). Calorimetry. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-38–3-62). National Fire Protection Association; Society of Fire Protection Engineers.
- Janssens, M. L. (2008). Basics of fire containment. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 2-59–2-74). National Fire Protection Association.
- Madrzykowski, D., & Stroup, D. W. (2008). Flammability hazard of materials. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 2-31–2-48). National Fire Protection Association.
- National Research Council of Canada, International Code Council, Department of Building and Housing, & Australian Building Codes Board. (2005). International fire engineering guidelines (2005 ed.). Australian Building Codes Board.
- Tewarson, A. (2002). Generation of heat and chemical compounds in fires. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-82–3-161). National Fire Protection Association; Society of Fire Protection Engineers.