Skip to content
From Fire Safety Academy Wiki
(Redirected from Steiner tunnel)
Contents
  1. 1Origin and designations
  2. 2Apparatus and procedure
  3. 3Flame spread in the tunnel
    1. 3.1Worked example: the red oak calibration
  4. 4Results and classification
  5. 5Other applications
  6. 6Factors affecting results
  7. 7Limitations and comparison with room-corner tests
  8. 8See also
  9. 9Notes
  10. 10References

The Steiner tunnel test is a North American fire test that measures how far and how fast flame spreads along the underside of a long, horizontal strip of building material.[1][2] It was developed by A. J. Steiner at Underwriters Laboratories and is standardised as NFPA 255, ASTM E84 and UL 723.[1] The specimen, about 7.3 m long, forms the ceiling of a 7.62 m (25 ft) tunnel. A gas burner of about 89 kW ignites one end while air is drawn through the tunnel at a controlled speed.[1] Observers record the position of the flame front for 10 minutes, and a light beam across the exhaust duct records the smoke.[2]

The results are two relative numbers, the flame spread index and the smoke-developed index, with red oak flooring at 100 and a cement board at 0.[2] In the United States, NFPA 101 and NFPA 5000 (2006 editions) use these numbers to place wall and ceiling finishes in Classes A, B and C.[3][4] The test is simple and widely specified, but it gives no results in engineering units.[5] It can also rate thin, melting and foamed materials too favourably, so room-corner tests are now accepted as an alternative.[6][3][1]

Origin and designations

The test grew out of the need to control the interior finish of buildings. After a series of fatal fires, a method for controlling wall and ceiling finishes was sought, and A. J. Steiner of Underwriters Laboratories proposed the tunnel test.[1] Among the early descriptions, the handbook cites Bulletin of Research No. 32 of Underwriters Laboratories and a 1951 article by Steiner in Fire Engineering.[7]

The American Society for Testing and Materials, now ASTM International, adopted the method as a tentative standard in 1950 and as an official standard, ASTM E84, in 1958.[1] The National Fire Protection Association adopted it tentatively in 1953 and officially in 1958 as NFPA 255. Both organisations have revised it periodically since.[1]

NFPA 255 carries the title Standard Method of Test of Surface Burning Characteristics of Building Materials.[8] A comparison table in the 2008 Fire Protection Handbook lists NFPA 255, ASTM E84 and UL 723 as equivalent tunnel standards. The same row contains NFPA 262 and the withdrawn UL 910, which apply the tunnel to cables, and no ISO counterpart.[9] The handbook also calls it simply the tunnel test or the 25 ft Steiner tunnel test.[1]

Hirschler describes the tunnel as perhaps the most widely accepted test for surface flammability, but only in North America.[5] Quintiere calls ASTM E84 the primary flammability test used in the United States.[10] The method was developed for building materials such as "fireproof paint" coatings, wood and gypsum board.[5] It is considered adequate for traditional building materials and for materials that are not very thin and stay in place during the test.[1] Its purpose is comparative: NFPA 255 provides a basis for comparing one product with another under one standardised set of conditions.[2]

Apparatus and procedure

The apparatus is a long, narrow, insulated box. The test chamber is 7.62 m (25 ft) long, with an internal width of 448 mm (17.625 in.) and a height of 305 mm (12 in.).[1] Its walls are lined with fire brick and carry double-pane observation windows along one side.[2]

Side view of a long horizontal tunnel: burner at the left end, specimen forming the ceiling, air flowing right towards an exhaust duct with a light beam.
Layout of the Steiner tunnel: the specimen forms the ceiling of a 7.62 m chamber, a gas burner ignites it at the fire end, and air is drawn towards the vent end, where a light beam across the exhaust duct measures the smoke.

The specimen is 508 to 610 mm (20 to 24 in.) wide and 7.3 m (24 ft) long, and normally up to 0.15 m thick.[1] It is laid face down on a ledge at the top of the furnace, so that it forms the ceiling of the chamber. It can be one unbroken length or several sections joined end to end.[1][11] A removable top panel of metal and mineral composite closes the chamber above the specimen.[2]

Tunnel cross-section: fire-brick walls, a 448 by 305 mm chamber, the specimen face down on ledges as the ceiling, a lid and a side window.
Cross-section of the Steiner tunnel: the specimen rests face down on ledges at the top of the fire-brick chamber, about 448 mm wide and 305 mm high, under a removable lid.

At the fire end, a gas burner projects a flame about 1.4 m (4½ ft) long onto the underside of the specimen. Hirschler gives its intensity as 89 kW in the interior finish chapter of the handbook.[1] The fire testing chapter describes two gas burners giving 88 kW,[11] and Benedetti and Colonna give 5000 Btu/min (88 kW).[12] Air enters through an adjustable shutter at the fire end. An induced-draft system with an automatically controlled damper draws it along the tunnel at a linear velocity of 1.22 m/s (240 ft/min).[1][2]

Two measurements are made during the 10-minute test. Observers follow the flame front through the side windows and record its position as a function of time.[2] Downstream of the vent end, a lamp and a photoelectric cell mounted across the 406 mm (16 in.) exhaust duct measure how much light the passing smoke absorbs.[1][2]

The apparatus is calibrated with two reference materials. According to the test standard, red oak flooring spreads flame over the full 7.3 m (24 ft) length of the specimen in 5½ minutes ± 15 seconds.[2] A cementitious board serves as the second reference, with a value of zero for both indices.[2]

Key test conditions of the Steiner tunnel test (NFPA 255, as described in the 2008 edition of the Fire Protection Handbook)[1][2][11]
Quantity Value
Length of test chamber 7.62 m (25 ft)
Specimen size 508 to 610 mm wide, 7.3 m long, normally up to 0.15 m thick
Specimen position face down, forming the ceiling of the chamber
Ignition source gas burner, flame about 1.4 m long, 88 to 89 kW
Air velocity 1.22 m/s (240 ft/min)
Test duration 10 minutes
Reference materials red oak flooring (100) and cementitious board (0)
Calibration red oak: flame reaches the end of the specimen in 5½ min ± 15 s

Flame spread in the tunnel

Flame spread over a solid surface can run against or with the surrounding gas flow. Quintiere calls spread against the flow opposed-flow spread and spread with the flow wind-aided spread.[13] In the tunnel, the flame travel and the forced airflow have the same direction.[14] Quintiere therefore describes the test as a complex form of wind-aided spread on a ceiling-mounted material in a duct under forced flow.[10] The radiant panel test, ASTM E162, works the other way: its flame travels downward against the air, as in opposed-flow flame spread.[15]

The direction matters, because wind-aided spread is the case that Madrzykowski and Stroup regard as the most dangerous.[15] The flame then lies over the unburnt surface ahead of the pyrolysis zone and heats it. For upward turbulent spread on a wall, Quintiere gives a flame heat flux ahead of the pyrolysis zone of about 25 kW/m².[10] Van Rossum explains the effect: the flame transfers much heat to the material beyond the burning part, preheats it and so lets the fire spread faster.[16] With some materials, the visible flame front in the tunnel can run several feet ahead of the zone where the surface is actually burning.[14] Quintiere notes that the test cannot measure the position of the pyrolysis front, which is the physical measure of surface flame spread.[10][17]

Two sketches: top, air and flame moving right together under a ceiling specimen; bottom, a flame moving down a sloping panel against rising air.
In the tunnel (top), air and flame move in the same direction, so the flame heats the surface ahead of the burning zone (wind-aided spread). In the radiant panel test (bottom), the flame travels against the air flow (opposed-flow spread).

A correlation reported by Quintiere links tunnel ratings to the energy release rate per unit area of the burning material. Ratings stayed low and insensitive below about 70 kW/m², and rose steeply above that value.[10] This suggests that a critical rate of energy release is needed before spread becomes self-sustaining. Quintiere considers it likely that the tunnel can be modelled with wind-aided spread theory, once tunnel-specific relationships for its input quantities are developed.[10]

The standard test does not measure heat release rate.[18] Parker did measure it in the ASTM E84 tunnel by analysing the oxygen consumed in the exhaust gases.[19] Janssens describes this as perhaps the first application of the oxygen consumption principle in fire research.[20]

Worked example: the red oak calibration

The calibration values give an idea of how fast flame moves in the tunnel. They also show how slow the flame is compared with the air that drives it. The mean speed of the flame front is the distance it covers divided by the time it takes:

v¯=Lt

Symbol Description Unit
v¯ mean speed of the flame front over the specimen m/s
L distance covered by the flame front m
t time taken to cover that distance s

For red oak flooring, the flame covers L = 7.32 m (24 ft) in t = 330 s ± 15 s.[2] The equation gives a mean speed between 7.32/345 = 0.0212 m/s and 7.32/315 = 0.0232 m/s, or about 2.1 to 2.3 cm/s. The air moves at 1.22 m/s, roughly 50 to 60 times faster than the flame front.[1] The result is of the same order as the 1 to 6 cm/s that Quintiere reports for wind-aided spread over PMMA samples 2.4 to 9.8 m long, at air speeds of 0.8 to 3.8 m/s.[10] The calculation gives only an average. The index itself is based on the whole distance-time record, not on a single speed.[11]

Results and classification

The flame spread record is turned into the flame spread index. This is a relative number based on the area under the curve of flame spread distance against time. Cement board is assigned 0 and red oak flooring approximately 100.[11] The lower the index, the better the presumed performance.[2] The light attenuation record is turned into the smoke-developed index in a similar way, except that the area for red oak is exactly 100 by definition.[11] A product with a smoke-developed index of 450 produces 4.5 times the area under the light attenuation curve of red oak flooring.[2]

In the United States, the building, fire and life safety codes group the results into classes, as summarised by Hirschler (2008a).[21] The flame spread limits differ between the classes, while a single smoke-developed limit of 450 is common to all three.[3] Plenum materials have stricter limits for both indices.[21]

Flame spread index and smoke-developed index classes in United States codes (Fire Protection Handbook, 2008)[3][21]
Class Flame spread index Smoke-developed index
Class A 0 to 25 0 to 450
Class B 26 to 75 0 to 450
Class C 76 to 200 0 to 450
Plenum materials 25 or less 50 or less

Tunnel data vary considerably, and replicate tests are not likely to give precise values. Differences of 10 to 20 points are not likely to be discernible in actual fire behaviour, so flame spread indices are usually rounded to the nearest whole number divisible by 5.[3] The two indices measure different things and are not directly related. In one study by Underwriters Laboratories, a material with a flame spread index of 490 had a smoke-developed index of 57, while another with 44 had 1387.[3]

In the United States, NFPA 101, Life Safety Code, and NFPA 5000 (2006 editions), as described by Hirschler (2008b), assign a class to each location and occupancy. Exit enclosures receive the strictest flame spread limits, because occupants there have little room to avoid the heat of a flame.[3] Occupancies with less capable occupants, such as health care facilities, are also stricter.[4] In general, NFPA 101 limits the flame spread index of wall and ceiling finishes to a maximum of 200, with lower values in exits and exit access corridors.[22]

The annex of NFPA 101 permits a finish one class lower where a complete automatic sprinkler system is installed, except in new detention and correctional occupancies.[4][23] In the International Building Code, as quoted in the International Fire Engineering Guidelines (2005), wall and ceiling finishes must not exceed a tabulated flame spread index. Finishes other than textiles that pass NFPA 286 may be used where Class A to ASTM E 84 is required.[24]

Other applications

Several product tests use the tunnel apparatus. NFPA 262 tests electrical cables intended for plenums, the spaces that carry environmental air. Other application tests derived from the tunnel cover plastic sprinkler pipe, pneumatic tubing and optical fibre raceways; the handbook names UL 1887, UL 1820 and UL 2024 among them.[11] In the United States, NFPA 70, the National Electrical Code, requires plenum cables to pass NFPA 262, as described by Hirschler (2008a).[25] The cables are laid horizontally in a grouped arrangement and exposed to an 88 kW flame. They may show a flame travel of at most 1.52 m, a peak optical density of at most 0.5 and an average optical density of at most 0.15.[25]

The tunnel is also the reference test for fire-retardant-treated wood. Shaw quotes the definition used with NFPA 255 (ASTM E84): the test is extended to 30 minutes, the flame spread must not exceed 25 and no progressive combustion may occur.[26] In addition, the flame front may never progress more than 3.2 m (10½ ft) beyond the centreline of the burner.[26] Underwriters Laboratories requires listed fire-retardant coatings to reach a flame spread rating of 50 or less, or a 50 percent reduction compared with the untreated surface, whichever is lower.[27] Shaw's table for listed coatings on unprimed Douglas fir gives three numbers: flame spread, fuel contributed and smoke developed. Untreated Douglas fir scored 100 on all three, while one coating reduced them to 10, 15 and 5.[28]

Other fields have adopted tunnel ratings as well. The original United States nuclear power plants, designed to 1970s code requirements, required ASTM E84 Class A flame spread for the carpet of the main control room. Because carpet rated by ASTM E84 is no longer manufactured, replacement carpet must be tested by the licensee or justified by an engineering analysis.[29] In a stochastic model of fire spread along a corridor, Ramachandran takes the probability of spread before flashover as governed, to a first approximation, by the flame spread classification of the corridor finishes.[30]

Factors affecting results

A tunnel result applies only to the product as it was mounted. Products should be tested as they are proposed for use.[2] A thin wall covering fixed with high-temperature mortar to a cement board may perform artificially well, but such an assembly does not represent real installations.[2] Vinyl wall covering that will be glued to gypsum board must be tested on gypsum board, because the paper facing and the adhesive affect the result. Adhesives can significantly change both the flame spread index and the smoke-developed index.[2]

In the United States, NFPA 101 (2006 edition) requires classification from tests that simulate actual installations.[2] NFPA 255 (2006 edition) itself states that the specimen shall be representative of the material for which results are desired.[3] ASTM has issued standard practices for mounting pipe and duct insulation (ASTM E2231), paper and vinyl wall and ceiling coverings (ASTM E2404) and site-fabricated stretch systems (ASTM E2573).[21]

Thickness and backing also matter. Thinner sections of a material often spread flame faster than thicker ones. For acrylic sheet (PMMA), the spread rate stays constant only above a thickness of about 9.5 mm (3/8 in.).[3] Thermally thin wood panelling, 6.4 mm (1/4 in.) thick or less, spreads flame faster when installed over furring strips with an air space than when fixed directly to cement board or gypsum board.[3][31] A single layer of wall covering may perform well, while several layers of the same product can spread flame. Multiple layers of wall covering on gypsum board have contributed to flame spread in fires with multiple deaths.[3]

Support methods can distort the result. Single-layer membranes and thin laminates are often supported on poultry netting over steel rods. The metal absorbs much of the heat from the burning material and lowers the flame spread artificially.[21][11]

Effect of poultry netting on the tunnel flame spread index[21]
Material Flame spread index with poultry netting Flame spread index without poultry netting
Sample 1 20 340
Sample 2 25 310
Sample 3 55 105
Sample 4 15 45

Limitations and comparison with room-corner tests

The tunnel is unsuitable for fire safety engineering calculations, mainly because its indices are non-dimensional and carry no engineering units.[5] There is not necessarily a correlation between the flame spread index and the behaviour of a product in a real fire.[2] Madrzykowski and Stroup add that the relationship of tunnel values to real-world applications has not been established. They list the ceiling mounting, the limited ability to test materials that melt and drip, and the limited ventilation for some polymer foams as limitations.[15]

The errors are mostly in one direction. Hirschler states that the tunnel is very likely to give falsely favourable results, for example for materials that melt and drip and for thin films. It rarely gives falsely unfavourable results, so a high index almost always indicates mediocre or poor performance.[6] Thermoplastic specimens that fall from the ceiling and keep burning on the tunnel floor give results that are often misleading.[11] Results should also be treated with suspicion for lightweight materials of low thermal inertia.[3]

Foamed plastics show the problem clearly. In a test series at Underwriters Laboratories, 31 materials were tested in the tunnel and also exposed to a 9.1 kg (20 lb) wood crib in a full-size corner. The crib was judged about as severe as the tunnel burner.[12] Four materials, including a glass fibre blanket and a sprayed polyurethane foam, all achieved a flame spread classification of 18. In the corner test, however, the share of the wall and ceiling area affected by burning ranged from 24 to 85 percent.[12]

Bar chart: four materials all rated 18 in the tunnel; wall and ceiling area affected in the corner test was 24, 37, 25 and 85 percent.
Four materials with the same tunnel flame spread classification of 18 and the share of wall and ceiling area they affected in a full-scale corner test with a 9.1 kg wood crib.

A 51 mm (2 in.) extruded polystyrene board obtained a flame spread classification of 3, yet burning adversely affected 76 percent of the wall and ceiling area of the corner structure.[12] Building and fire codes have therefore identified foam plastic insulation as a product for which the tunnel protocol is not appropriate.[1] Wall coverings behave in a similar way: in the 1980s, coverings with tunnel ratings as low as 15 sometimes produced flashover in room-corner tests.[23]

Floor coverings were once tested in the tunnel as well. In 1965 the United States Public Health Service set flame spread limits for floor coverings in federally funded hospitals, citing the tunnel test.[32] The covering then had to be mounted upside down on the ceiling of the apparatus, a position unlike any real installation. The tunnel cannot assess ignitability, its fuel source does not suit slow spread, and it failed with carpets that melt, drip or delaminate.[32] The test was judged inadequate for floor coverings, and most codes now use the flooring radiant panel test instead.[32][33]

The room corner test addresses these weaknesses. It assesses interior finish in a realistic room and measures heat release, the key fire property, so its results are more reliable than tunnel results.[1] A survey of 25 materials tested in both methods found five materials with a flame spread index of 200 or less that led to flashover in the room-corner test.[6] Two materials rated Class C in the tunnel released heat very close to, but below, the flashover level in the room-corner test, and the codes consider them equivalent to Class A materials.[6]

The 2006 editions of the United States codes accept NFPA 286 as an alternative. A material passes if flames do not reach the ceiling during a 40 kW exposure, and no flashover occurs during the following 160 kW exposure. Peak heat release must stay at or below 800 kW, and total smoke release at or below 1000 m².[6]

Smaller versions of the tunnel have not lasted. The Forest Products Laboratory developed a 2.4 m (8 ft) tunnel, which ASTM adopted as ASTM E286 in 1968 and withdrew in 1991.[14] A floor-mounted 2.44 m version, UL Subject 992, never gained wide acceptance and is no longer used.[32] Hirschler concludes that present fire testing of interior finish is probably adequate to eliminate the poorest performers. The dominance of the tunnel test, however, means that the full capabilities of room-corner tests, including measured heat release rates, are not being used.[34]

See also

Notes

  1. ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.11 1.12 1.13 1.14 1.15 1.16 1.17 1.18 Hirschler (2008b), p. 18-25.
  2. ↑ 2.00 2.01 2.02 2.03 2.04 2.05 2.06 2.07 2.08 2.09 2.10 2.11 2.12 2.13 2.14 2.15 2.16 2.17 2.18 2.19 Hirschler (2008b), p. 18-26.
  3. ↑ 3.00 3.01 3.02 3.03 3.04 3.05 3.06 3.07 3.08 3.09 3.10 3.11 Hirschler (2008b), p. 18-27.
  4. ↑ 4.0 4.1 4.2 Hirschler (2008b), p. 18-28.
  5. ↑ 5.0 5.1 5.2 5.3 Hirschler (2008a), p. 6-42.
  6. ↑ 6.0 6.1 6.2 6.3 6.4 Hirschler (2008b), p. 18-31.
  7. ↑ Hirschler (2008b), p. 18-40.
  8. ↑ Hirschler (2008b), pp. 18-24–18-25.
  9. ↑ Hirschler (2008a), p. 6-38.
  10. ↑ 10.0 10.1 10.2 10.3 10.4 10.5 10.6 Quintiere (2002b), p. 2-253.
  11. ↑ 11.0 11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8 Hirschler (2008a), p. 6-43.
  12. ↑ 12.0 12.1 12.2 12.3 Benedetti and Colonna (2008), p. 6-226.
  13. ↑ Quintiere (2002b), p. 2-247.
  14. ↑ 14.0 14.1 14.2 Hirschler (2008b), p. 18-32.
  15. ↑ 15.0 15.1 15.2 Madrzykowski and Stroup (2008), p. 2-42.
  16. ↑ Van Rossum (2027), §4.3.
  17. ↑ Quintiere (2002b), p. 2-246.
  18. ↑ Madrzykowski and Stroup (2008), p. 2-44.
  19. ↑ Madrzykowski and Stroup (2008), p. 2-37.
  20. ↑ Janssens (2002), p. 3-46.
  21. ↑ 21.0 21.1 21.2 21.3 21.4 21.5 Hirschler (2008a), p. 6-44.
  22. ↑ Lathrop (2008a), p. 4-75.
  23. ↑ 23.0 23.1 Hirschler (2008b), p. 18-29.
  24. ↑ National Research Council of Canada et al. (2005), p. 0.2-10 (Part 0, United States).
  25. ↑ 25.0 25.1 Hirschler (2008a), p. 6-45.
  26. ↑ 26.0 26.1 Shaw (2008), p. 6-133.
  27. ↑ Shaw (2008), p. 6-136.
  28. ↑ Shaw (2008), p. 6-135.
  29. ↑ Iqbal and Salley (2004), p. B-34.
  30. ↑ Ramachandran (2002c), p. 3-393.
  31. ↑ Hirschler (2008b), p. 18-38.
  32. ↑ 32.0 32.1 32.2 32.3 Hirschler (2008b), p. 18-33.
  33. ↑ Hirschler (2008b), p. 18-35.
  34. ↑ Hirschler (2008b), p. 18-39.

References

  • 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.
  • 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.
  • Hirschler, M. M. (2008b). Interior finish. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 18-23–18-42). 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.
  • Lathrop, J. K. (2008a). Concepts of egress design. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 4-69–4-91). 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.
  • Quintiere, J. G. (2002b). Surface flame spread. 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. 2-246–2-257). National Fire Protection Association; Society of Fire Protection Engineers.
  • Ramachandran, G. (2002c). Stochastic models of fire growth. 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. 3-381–3-401). National Fire Protection Association; Society of Fire Protection Engineers.
  • Shaw, J. R. (2008). Fire-Retardant and flame-resistant treatment of cellulosic materials. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-129–6-139). National Fire Protection Association.
  • Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.