Skip to content
From Fire Safety Academy Wiki
Lisa (talk | contribs)
New FSE wiki article: NFPA 921 (written by Claude Opus 5.5)
 
Lisa (talk | contribs)
FSE wiki: Fact-check: scope historical US and Canadian investigation account; label derived charring relation (written by Claude Opus 5.5)
Line 88: Line 88:
Custer and Sheppard (2008) note that research related to fire investigation was ongoing and that past techniques were being refined. As the state of the art develops, previous determinations may need to be re-evaluated.<ref name="CS-3-39" /> The burning rate of wood, measured as char depth over time, is an example. It was traditionally taken as about 0.6 mm/min (1.4 in. in 60 minutes). Research showed that it varies with radiant heat flux and ventilation and can be as high as 4.3 mm/min.<ref name="CS-3-35" />
Custer and Sheppard (2008) note that research related to fire investigation was ongoing and that past techniques were being refined. As the state of the art develops, previous determinations may need to be re-evaluated.<ref name="CS-3-39" /> The burning rate of wood, measured as char depth over time, is an example. It was traditionally taken as about 0.6 mm/min (1.4 in. in 60 minutes). Research showed that it varies with radiant heat flux and ventilation and can be as high as 4.3 mm/min.<ref name="CS-3-35" />


The traditional value was used to estimate the time of burning and to detect flammable liquids.<ref name="CS-3-35" /> In that use, the burning time implied by a char depth is the depth divided by the charring rate, so an uncertain rate gives an uncertain time. The handbook therefore states that indicators such as depth of char should be used with care.<ref name="CS-3-35" />
The traditional value was used to estimate the time of burning and to detect flammable liquids.<ref name="CS-3-35" /> In that use, the burning time implied by a char depth is the depth divided by the charring rate, so an uncertain rate gives an uncertain time. The handbook therefore states that indicators such as depth of char should be used with care.<ref name="CS-3-35" /> To illustrate the sensitivity to the assumed rate, this relation can be written as:


<math display="block">t = \frac{x_\mathrm{c}}{\beta}</math>
<math display="block">t = \frac{x_\mathrm{c}}{\beta}</math>
Line 104: Line 104:
As an illustration, an assumed char depth of 20 mm gives about 33 minutes at the traditional rate of 0.6 mm/min. At the rate of 4.3 mm/min reported by research, the same depth gives only about 4.7 minutes.<ref name="CS-3-35" /> The depth is an invented example, but the sevenfold spread follows directly from the two rates in the handbook.
As an illustration, an assumed char depth of 20 mm gives about 33 minutes at the traditional rate of 0.6 mm/min. At the rate of 4.3 mm/min reported by research, the same depth gives only about 4.7 minutes.<ref name="CS-3-35" /> The depth is an invented example, but the sevenfold spread follows directly from the two rates in the handbook.


Fire models were increasingly used in forensic work, but Custer et al. (2008) caution that they must be applied with care to gain credibility there.<ref name="CSW-3-118">Custer et al. (2008), p. 3-118.</ref> Their guidance is to know the limitations of the models, to use more than one model, and to check the model against experiments. Users should also apply ranges of values for sensitive variables and consider ambient conditions.<ref name="CSW-3-117">Custer et al. (2008), pp. 3-117–3-118.</ref> Some real fires may be too complicated for models to analyse as a whole, although parts, such as early growth in the compartment of origin, may still be studied.<ref name="CSW-3-118" /> The organisations, legal mechanisms and qualification standards that the handbook links to NFPA 921 are those of the United States, as they stood in 2008.<ref name="CS-3-32" />
Fire models were increasingly used in forensic work, but Custer et al. (2008) caution that they must be applied with care to gain credibility there.<ref name="CSW-3-118">Custer et al. (2008), p. 3-118.</ref> Their guidance is to know the limitations of the models, to use more than one model, and to check the model against experiments. Users should also apply ranges of values for sensitive variables and consider ambient conditions.<ref name="CSW-3-117">Custer et al. (2008), pp. 3-117–3-118.</ref> Some real fires may be too complicated for models to analyse as a whole, although parts, such as early growth in the compartment of origin, may still be studied.<ref name="CSW-3-118" /> The organisational, legal and qualification details in this article follow the handbook's 2008 account of US practice, with a separate mention of Canadian provincial fire marshals.<ref name="CS-3-32" /><ref name="CS-3-40" />


== See also ==
== See also ==

Revision as of 03:33, 8 November 2026

NFPA 921, Guide for Fire and Explosion Investigations, is a guide of the National Fire Protection Association (NFPA) in the United States for the investigation of fire and explosion incidents. According to the 2008 Fire Protection Handbook, it assists people who investigate and analyse such incidents and give opinions on their origin, cause, responsibility or prevention.[1] The NFPA Standards Council started the project in 1984, and the first edition was issued in 1992.[2] The 2004 edition included 24 chapters, from basic methodology and fire science to origin and cause determination, explosions and the management of major investigations.[2]

In the United States the guide was widely used for state and local training and formed the basis of the performance criteria in NFPA 1033, the qualifications standard for fire investigators.[2] Handbook chapters and a 2004 report of the US Nuclear Regulatory Commission quote its definitions of terms such as pyrolysis, ignition temperature and explosion.[3][4][5] The guide supports fire investigation, whose results can lead to targeted inspection, public education or proposed code changes.[6][7]

Scope and purpose

Farr and Sawyer (2008) describe NFPA 921 as a guide for individuals charged with investigating and analysing fire and explosion incidents and rendering opinions on their origin, cause, responsibility or prevention.[1] Origin and cause are central, but an investigation can also analyse other factors. Examples are the performance of built-in fire protection systems, building construction assemblies and egress systems, and the response of building staff in occupancies such as health care.[1] The same authors divide fire investigation into two areas: determining the fire cause and investigating criminal actions that may have contributed to a fire.[6] Analysis of the causes of fires in a community is the basis for setting fire prevention priorities.[6]

Custer and Sheppard (2008) explain that in most fires the origin and cause are expressed in four elements. These are the area of fire origin, the heat of ignition, the materials involved, and the factors that brought them together, with a reference to NFPA 921.[7] Knowing the origin and cause can lead to targeted inspection, public education programmes or proposed code changes. The results of a failure analysis provide input to insurance loss adjustment and underwriting, design practice, and civil and criminal litigation.[7] Complete investigations were once reserved for fires with a large loss of life or property, often months or years after the event. By 2008, insurance claims, subrogation activities and litigation concerned ever smaller losses, so more detailed examination was needed of all fires that result in a loss.[7]

History and development

In 1984 the NFPA Standards Council initiated a project to develop guidelines for fire and explosion investigation under the consensus process used for NFPA standards. The Technical Committee on Fire and Explosion Investigations was established at that time.[2] The committee reviews public comments and documents its actions in a final proposal. That proposal is again open to public review, so that commenters can respond and raise questions during the final approval process.[2] Custer and Sheppard (2008) state that this open procedure gives a document representative of the investigative community, although some individuals may disagree on specific points.[2]

The first edition was issued in 1992, eight years after the project started, and the guide was revised on a three-year basis through 2004.[2] Even without a new edition, the guide would have to be opened to public comment and reviewed every five years to keep it from becoming outdated.[2] NUREG-1805, a 2004 report of the US Nuclear Regulatory Commission, notes more generally that NFPA codes and standards are revised on cycles of three to five years. The edition in effect at the time of a design or implementation is the code of record.[8]

The documents that quote the guide therefore cite different editions. NUREG-1805 cites the 1992 edition on explosions and lists the 2001 edition as current in its table of NFPA documents.[9][10] The 2008 handbook reproduces data from the 2004 edition and a photograph credited to a "User's Manual for NFPA 921, 2005 edition".[11][12]

Contents of the 2004 edition

Custer and Sheppard (2008) list the chapters of the 2004 edition.[2] The list opens with basic methodology, basic fire science, fire patterns, building systems, electricity and fire, building fuel gas systems and fire-related human behaviour. It continues with legal considerations, safety, sources of information, planning and documenting the investigation, and physical evidence. Then come origin determination, fire cause determination, analysing the incident for cause and responsibility, and failure analysis and analytical tools. The list closes with explosions, incendiary fires, fire and explosion deaths and injuries, appliances, motor vehicle fires, wildfire investigations and the management of major investigations.[2]

The guide is also a source of definitions, several of which Madrzykowski and Stroup (2008) quote. Pyrolysis is the chemical decomposition of a material into one or more other substances due to heat alone. Smouldering is a slow, exothermic surface reaction, usually characterised by glowing and smoke production.[3] The ignition energy is the heat per unit of exposed surface area that a material must absorb to pyrolyse, ignite and burn: the absorbed heat flux times the exposure time until ignition.[12] The ignition temperature is the minimum temperature a solid must attain to ignite under specific test conditions.[4]

The 2004 edition had a chapter on explosions, and NUREG-1805 quotes two descriptions of an explosion attributed to NFPA 921.[2][5] One is a sudden conversion of potential energy, chemical or mechanical, into kinetic energy in the form of rapidly expanding gases. The other is a physical reaction with four elements: high-pressure gas, confinement of the pressure, rapid production or release of that pressure, and change to the confining structure.[5]

NUREG-1805 reproduces from the guide a table that explains electrical systems by analogy with water systems. It pairs a pump with a generator, pressure with voltage, flow with current, friction with resistance and pipe size with conductor size.[13] The report adds that current through a conductor generates heat, in an amount that depends on the resistance and the current.[13] Custer and Sheppard expect investigators to know electrical power distribution systems and the failure modes of common electrical equipment and appliances.[14]

The guide also compiles fire test data. The handbook chapter on compartment fire growth reproduces, from Table 5.4.2.1 of the 2004 edition, representative peak heat release rates of items burning unconfined.[11] The values span almost three orders of magnitude, from 4 kW for a small wastebasket to 3,120 kW for a polyurethane sofa.[11]

Selected representative peak heat release rates, unconfined burning, from NFPA 921 (2004), Table 5.4.2.1, as reproduced by Custer (2008)[11]
Item Mass as printed (lb) Mass converted (kg) Peak heat release rate (kW)
Small wastebasket 1.5–3 0.7–1.4 4–18
Cotton mattress 26–29 11.8–13.2 40–970
Television set 69–72 31.3–32.7 120–290
Cotton easy chair 39–70 17.7–31.8 290–370
Dry Christmas tree 14–16 6.4–7.3 500–650
Polyurethane mattress 7–31 3.2–14.1 810–2,630
Polyurethane easy chair 27–61 12.2–27.7 1,350–1,990
Polyurethane sofa 113 51.3 3,120

Investigation method

Custer and Sheppard (2008) describe the complete reconstruction and failure analysis process as developing a time history of the incident and locating the fire origin. It also establishes the factors that led to ignition and that increased or lessened the loss.[2] The basic steps are examination of the scene, collection of background data, testing, incident reconstruction and specific failure analysis, and many of these steps are outlined in NFPA 921.[15] Often only parts of the process are carried out, by different organisations at different times, so that each party relies on information gathered by others.[15]

Time is the framework of the analysis, on a macro scale of days to years for the prefire history and a micro scale of seconds to hours.[2] Events are arranged in timelines anchored on benchmark events, which are well documented in time and space. Examples are the issue of a building permit, the arrival of the fire department, an explosion and a roof collapse.[16] In a "hard time" timeline every event has a known time; in a "soft time" timeline only the order is known, and benchmark events bound the other times.[15]

The scene examination gathers physical and photographic evidence and documents the damage in maps and diagrams.[15] Each photograph should be documented on a diagram with its direction of view, and an original archival copy of every image should be kept that can be testified to in court. Maps should show the layout, damage, fuels, victims and fire protection features, and these features should be recorded both as built and at the time of the fire.[15] Interviews with firefighters are needed because some damage may come from fire fighting and overhaul, or from faster burning where the building was ventilated.[15]

The investigator should collect evidence, develop hypotheses about the origin and cause, and test them to see which the evidence supports.[17] Burn patterns such as a "V" pattern, low burns and multiple "points of origin" can be misleading. They should be used only when unambiguous or supported by other physical evidence.[17] The patterns show the total burning history, including extinguishment and overhaul, so patterns from fires extinguished early are more meaningful than those after near-total destruction. In some cases of total or near-total destruction, neither an on-scene investigation nor a later analysis can establish origin or cause.[17]

Energised electrical circuits should be examined for arcing, because the arc locations help to show how the fire progressed, a technique called arc fault mapping.[17] Common tests are identification of unknown materials, flammability tests, performance tests of fire protection systems and demonstration tests.[18] Laboratories identify possible ignitable liquids in debris by gas chromatography and mass spectroscopy, and such analysis should follow the appropriate ASTM standards.[18] Demonstration tests can mislead when ventilation, fuel arrangement or ignition point differ from the real event.[19]

Ignition, fire growth and failure analysis

Fire reconstruction is an organised, step-by-step portrayal of the most likely course of the fire, from the ignition sequence to the limits of flame and smoke at extinguishment.[20] It should explain the speed and direction of fire and smoke spread, the effects of barriers and suppression, and the performance of life safety features.[20] Failure modes and effects analysis and fault tree analysis help to identify alternative sequences of events, which are then evaluated against the evidence. Computer fire models can also compare possible scenarios with the actual events, as in the analyses of the DuPont Plaza and First Interstate Bank fires.[20]

Analysis of the ignition sequence concerns three items: the size, nature and source of the ignition energy, the fire properties of the material ignited, and the circumstances that brought energy and fuel together. These circumstances are human, such as careless welding, or mechanical, such as a valve rupture.[20] A candidate source is compared with the ignition energy the fuel needs in its actual form. For the same material and heat source, ignition is more likely if the material is finely divided than if it is a solid mass.[20] A competent ignition source has sufficient energy and temperature to raise a particular fuel to its ignition temperature during the postulated exposure time. The concept applies to electrical ignition: an electric arc may be far hotter than the ignition temperature of a fuel, yet too brief to supply enough energy.[20]

If established burning followed, the analysis must explain why the fire grew to flashover or to its final extent. The factors include the heat release rates and arrangement of the fuels, ceiling height, ventilation and interior finish.[20] A fire reported to have grown "unusually fast" is in many instances called suspicious or incendiary, when its growth may in fact be due to the burning properties of the materials and the compartment.[20] In 2008 the heat release rate of complex materials could not be calculated, and the only accurate data came from experiments.[20] These questions belong to fire dynamics, and investigators are expected to understand heat release rate, fire plumes, ceiling layers and flashover.[14]

For manual suppression, the fire department response time can be divided into detection, alarm transmission, dispatch, travel, and the time from arrival to agent application.[21] A delay in any segment extends the limits of flame and smoke travel, for example through missing detectors, traffic, frozen hydrants or blocked access.[21] For example, a sprinkler system may have failed to control a fire because a very high ceiling delayed its response and low water pressure could not deliver the needed flow.[21]

In a fire safety context, failure can be defined as a fire that results in injury, death, or property or monetary loss, or that prevents buildings or mechanisms from functioning as designed.[21] Failure analysis identifies the primary and contributing causes, whose sources range from basic design and equipment defects to postconstruction modifications and abuse. The causes are compared with the codes, standards and practice at the time of the fire. Custer and Sheppard regard it as even more important to note where compliance failed to provide the expected fire safety than to note noncompliance.[21]

Use in practice and training

In the United States the local fire department generally has the primary responsibility to document a fire and to make an initial determination of origin and cause.[7] When a fire may have been incendiary, the state or provincial fire marshal's office is responsible for investigating it in much of the United States and in all Canadian provinces.[22] Federal agencies such as the National Transportation Safety Board and the Bureau of Alcohol, Tobacco, Firearms and Explosives investigate fires under their jurisdiction. Insurers and consultants reconstruct fires to evaluate claims, to identify parties for possible subrogation action and to support litigation.[22]

When major financial consequences depend on the conclusions, it is not uncommon for two or more parties to retain investigators who present different theories of the fire. Custer and Sheppard name the need to resolve such disputes as a major reason for the increasing interest in applying scientific methods, fire models and test results to investigations.[22] Most scenes are on private property, and investigators must have a legal right to be there, otherwise their findings may not be admissible in court. In the United States the three primary mechanisms are exigent circumstances, the permission of the owner and a search warrant.[23]

NFPA 921 was widely used for state and local training programmes throughout the United States. It served as the basis for the performance criteria in NFPA 1033, Standard for Professional Qualifications for Fire Investigator.[2] It was also recognised for training by the International Association of Arson Investigators and the National Association of Fire Investigators.[2] NFPA 1033 is one of 15 professional qualifications standards for fire service positions that are the responsibility of the NFPA Professional Qualifications Correlating Committee.[1] Custer and Sheppard describe it as prescribing minimum basic competencies, beyond which a basic knowledge of physics and chemistry is needed, and college-level study for complex fires.[14]

The level of knowledge varies with the task. A fire ground officer at a residential fire extinguished early needs less than an investigator from the fire investigation unit of the department, who deals with advanced fires in all types of occupancy.[24] All investigators should be trained in scene documentation and in origin and cause determination, and should have some knowledge of the legal aspects of access, evidence and witnesses.[14]

Cause classification and arson

Custer and Sheppard (2008) state that a determination of arson can usually come only after a thorough fire investigation. In it, other possible causes are ruled out and physical evidence of an incendiary fire is established.[24] Investigators should not come to believe that their role is to find arson, and "fire investigation" is a more appropriate term than "arson investigation".[24] Burn patterns alone are not proof that flammable liquids were present: corroborative evidence such as chemical traces is needed, and smoke colour is not necessarily evidence.[17]

According to Ahrens et al. (2008), NFPA 921 does not consider "suspicious" to be a valid cause, and many fire investigators concur. There was broad consensus that "suspicious" is never a valid final cause after an investigation, although it was considered appropriate for fires unlikely ever to be fully investigated.[25] Older versions of the US National Fire Incident Reporting System offered "incendiary" and "suspicious", which analysts grouped together as arson. Version 5.0 dropped both choices and added the cause "intentional", which implies an illegal act less strongly.[25]

Investigations also feed fire data, but with limits. Hall and Ahrens (2002) note that anecdotal information, such as NFPA fire investigations or the in-depth investigations of the US Consumer Product Safety Commission, can indicate the range and diversity of fires that can happen. Such anecdotes cannot be used to estimate incident rates or probabilities.[26]

Limits and cautions

Custer and Sheppard (2008) note that research related to fire investigation was ongoing and that past techniques were being refined. As the state of the art develops, previous determinations may need to be re-evaluated.[24] The burning rate of wood, measured as char depth over time, is an example. It was traditionally taken as about 0.6 mm/min (1.4 in. in 60 minutes). Research showed that it varies with radiant heat flux and ventilation and can be as high as 4.3 mm/min.[17]

The traditional value was used to estimate the time of burning and to detect flammable liquids.[17] In that use, the burning time implied by a char depth is the depth divided by the charring rate, so an uncertain rate gives an uncertain time. The handbook therefore states that indicators such as depth of char should be used with care.[17] To illustrate the sensitivity to the assumed rate, this relation can be written as:

t=xcβ

Symbol Description Unit
t burning time implied by the char depth min
xc char depth measured at the scene mm
β assumed charring rate of the wood mm/min

As an illustration, an assumed char depth of 20 mm gives about 33 minutes at the traditional rate of 0.6 mm/min. At the rate of 4.3 mm/min reported by research, the same depth gives only about 4.7 minutes.[17] The depth is an invented example, but the sevenfold spread follows directly from the two rates in the handbook.

Fire models were increasingly used in forensic work, but Custer et al. (2008) caution that they must be applied with care to gain credibility there.[27] Their guidance is to know the limitations of the models, to use more than one model, and to check the model against experiments. Users should also apply ranges of values for sensitive variables and consider ambient conditions.[28] Some real fires may be too complicated for models to analyse as a whole, although parts, such as early growth in the compartment of origin, may still be studied.[27] The organisational, legal and qualification details in this article follow the handbook's 2008 account of US practice, with a separate mention of Canadian provincial fire marshals.[22][14]

See also

Notes

  1. ↑ 1.0 1.1 1.2 1.3 Farr and Sawyer (2008), p. 1-83.
  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 Custer and Sheppard (2008), p. 3-33.
  3. ↑ 3.0 3.1 Madrzykowski and Stroup (2008), p. 2-32.
  4. ↑ 4.0 4.1 Madrzykowski and Stroup (2008), p. 2-35.
  5. ↑ 5.0 5.1 5.2 Iqbal and Salley (2004), p. 15-1.
  6. ↑ 6.0 6.1 6.2 Farr and Sawyer (2008), pp. 1-90–1-91.
  7. ↑ 7.0 7.1 7.2 7.3 7.4 Custer and Sheppard (2008), p. 3-31.
  8. ↑ Iqbal and Salley (2004), p. E-1.
  9. ↑ Iqbal and Salley (2004), p. 15-17.
  10. ↑ Iqbal and Salley (2004), p. E-9.
  11. ↑ 11.0 11.1 11.2 11.3 Custer (2008), p. 2-50.
  12. ↑ 12.0 12.1 Madrzykowski and Stroup (2008), p. 2-34.
  13. ↑ 13.0 13.1 Iqbal and Salley (2004), pp. B-11–B-12.
  14. ↑ 14.0 14.1 14.2 14.3 14.4 Custer and Sheppard (2008), p. 3-40.
  15. ↑ 15.0 15.1 15.2 15.3 15.4 15.5 Custer and Sheppard (2008), p. 3-34.
  16. ↑ Custer and Sheppard (2008), pp. 3-33–3-34.
  17. ↑ 17.0 17.1 17.2 17.3 17.4 17.5 17.6 17.7 17.8 Custer and Sheppard (2008), p. 3-35.
  18. ↑ 18.0 18.1 Custer and Sheppard (2008), p. 3-36.
  19. ↑ Custer and Sheppard (2008), pp. 3-36–3-37.
  20. ↑ 20.0 20.1 20.2 20.3 20.4 20.5 20.6 20.7 20.8 Custer and Sheppard (2008), p. 3-37.
  21. ↑ 21.0 21.1 21.2 21.3 21.4 Custer and Sheppard (2008), p. 3-38.
  22. ↑ 22.0 22.1 22.2 22.3 Custer and Sheppard (2008), p. 3-32.
  23. ↑ Custer and Sheppard (2008), pp. 3-32–3-33.
  24. ↑ 24.0 24.1 24.2 24.3 Custer and Sheppard (2008), p. 3-39.
  25. ↑ 25.0 25.1 Ahrens et al. (2008), p. 3-68.
  26. ↑ Hall and Ahrens (2002), p. 5-70.
  27. ↑ 27.0 27.1 Custer et al. (2008), p. 3-118.
  28. ↑ Custer et al. (2008), pp. 3-117–3-118.

References

  • Ahrens, M., Frazier, P., & Kelch, G. (2008). Use of fire incident data and statistics. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-61–3-92). National Fire Protection Association.
  • Custer, R. L. P. (2008). Dynamics of compartment fire growth. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 2-49–2-58). National Fire Protection Association.
  • Custer, R. L. P., & Sheppard, D. T. (2008). Fire loss investigation. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-31–3-41). National Fire Protection Association.
  • Custer, R. L. P., Sheppard, D. T., & Wood, C. B. (2008). Applying models to fire protection engineering problems and fire investigations. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-111–3-120). National Fire Protection Association.
  • Farr, R. R., & Sawyer, S. F. (2008). Fire prevention and code enforcement. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 1-81–1-93). National Fire Protection Association.
  • Hall, J. R., Jr., & Ahrens, M. J. (2002). Data for engineering analysis. 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. 5-65–5-78). National Fire Protection Association; Society of Fire Protection Engineers.
  • 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.
  • 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.