Skip to content
From Fire Safety Academy Wiki
Revision as of 18:32, 14 November 2026 by Linda (talk | contribs) (Add SVG figures fse-safety-factor-v1.svg, fse-safety-factor-v2.svg, fse-safety-factor-v3.svg, fse-safety-factor-v4.svg, fse-safety-factor-v5.svg (list-8 visual plan))
Contents
  1. 1Definition and terminology
  2. 2Implied and explicit safety factors
  3. 3Safety factors in structural design
  4. 4Safety factors in egress and hazard calculations
  5. 5Clean agent design concentrations
  6. 6Choosing and combining safety factors
  7. 7Probabilistic approaches and limitations
  8. 8See also
  9. 9Notes
  10. 10References

A safety factor, or factor of safety, is a multiplier applied in design so that the design remains acceptable even if the values used in the calculation are wrong by an uncertain amount.[1][2] The International Fire Engineering Guidelines (IFEG, 2005) define it as an adjustment made to compensate for uncertainty in the methods, calculations and assumptions used to develop engineering designs.[3] The related safety margin is additive rather than multiplicative: it is the difference between the design value and the value that would no longer be safe.[1]

Fire safety engineering uses both forms. A calculated evacuation time may be doubled, the extinguishing concentration of a gaseous agent raised by 20 to 30 per cent, or a predicted radiant heat flux multiplied by two.[1][4][5] Safety factors can be explicit multipliers or implied in conservative assumptions and worst-case scenarios.[1] In performance-based fire design, most safety factors had been set ad hoc by 2008, because few methods provide specific values.[6] In 2002 there were no official rules for combining them, and a factor applied to an intermediate quantity is unlikely to give the same factor on the result.[7]

Definition and terminology

Notarianni (2002) starts from the dictionary definition in terms of stress: the ratio of the maximum stress that a structural part can withstand to the maximum stress estimated for it in use.[1] Safety factors are not limited to stress. In general, the design values are multiplied by the factor of safety, and the design is checked to make sure that it is still safe at the larger value. A safety margin, by contrast, is defined as the difference between the design value and the value that would no longer be safe.[1] For an illustrative demand xd with an upper safety limit xlim, multiplicative and additive checks can be written as:

SF⋅xd≤xlimxlim−xd≥Mmin

In the Fire Protection Handbook of the National Fire Protection Association (NFPA, 2008), Hurley and Bukowski call safety factors the classic method of treating uncertainty in building designs and codes.[2] A sufficient factor is applied so that, if all of the uncertainty resulted in error in the same direction, the result would still provide an acceptable solution. The glossary of the US Nuclear Regulatory Commission report NUREG-1805 (2004) adds that safety factors are used in most engineering designs. In fire protection engineering, they appear especially in evacuation times and in structural fire performance.[8]

In performance-based design, Puchovsky and Hurley (2008) describe safety factors as a way to deal with uncertainty or to provide a margin between outcomes considered acceptable and unacceptable.[6] Ideally, a safety factor should reflect the actual magnitude of the uncertainties. A related term is the analysis margin, the difference between the estimated value of a parameter and the value used in the analysis. A larger margin gives more confidence that the value will not be exceeded, but also a more expensive design. The mean of a set of survey values, such as the fuel load in offices, is a best estimate that is exceeded in half of the cases.[6]

Symbol Description Unit
SF Safety factor (multiplier) –
xd, xlim Design value of a demand and the value at which the design would no longer be safe as quantity
Mmin Minimum required safety margin as quantity
σa, fy Allowable stress and yield strength of steel MPa
tASET, tRSET Available and required safe egress time s or min
Cd, CMEC Design concentration and minimum extinguishing concentration of a gaseous agent % by volume
w Mass of halocarbon agent required kg
V Net volume of the protected space m³
S Specific volume of the agent vapour at the minimum ambient temperature m³/kg
qf,d, qf,k Design and characteristic fire load density MJ/m²
m, δq1, δq2, δn Combustion factor, safety factors for compartment size and occupancy, factor for active fire-fighting measures –

Implied and explicit safety factors

Safety factors are used with both prescriptive and performance-based codes, and they can be implied or explicit.[1] Implied safety factors give an extra margin simply through the choice of a component, or take the form of conservative assumptions and worst-case scenarios. Explicit safety factors are multipliers applied to critical analysis parameters, preferably to the final outcome criteria used to judge whether a design is acceptable. Both types lower the probability that critical values will be reached or exceeded.[1]

Notarianni (2002) gives an example of each kind.[1] Pipe schedules in sprinkler standards carry an implied safety factor, because the pipe exceeds the strength needed. A required sprinkler flow density 1.5 times the minimum shown experimentally to control a fire is an explicit factor in a prescriptive code. In a performance-based design, assuming that a fuel package produces carbon monoxide at the rate of its worst component is an implied factor, and doubling the calculated egress time is an explicit one. Fleming (2008a) describes a further prescriptive case: sprinkler hangers are expected to support five times the weight of the water-filled pipe plus 250 lb (114 kg). The extra load represents a person using the piping for support.[9]

Design inputs often contain implied safety factors. In New Zealand, Verification Method C/VM2 (second edition, 2025) notes that its design fires ignore the incipient phase of fire growth. This provides an inherent safety factor for the pre-travel activity times it prescribes.[10] The IFEG (2005) cite a 1983 compilation by the International Council for Building Research Studies and Documentation (CIB). It warns that a fire load density chosen at its mean would give a negative safety factor for all higher values. At least the 95 per cent fractile should therefore be used.[11]

For offices, the mean in the guidelines' table is 420 MJ/m² and the 95 per cent fractile is 760 MJ/m².[11] As an illustration, choosing the fractile thus implies a factor of about 1.8 on the mean. Conversely, Hurley and Bukowski (2008) note that design fires chosen as a worst likely scenario need no extra safety factor, unless the result is very sensitive to an assumption.[2]

Safety factors in structural design

Structural engineering provides a familiar example. In allowable stress design, the applied stress in any element must not exceed its strength divided by a factor of safety.[12] The factor generally varies from 1.5 to 3.0, depending on the reliability or variability of the strength data and on the consequences of failure. Manley's example is a steel beam with a specified minimum yield strength of 248,000 kPa (248 MPa). With the minimum factor of 1.5 that building codes required for simple gravity loads, the allowable stress becomes 248/1.5, or about 165 MPa.[12]

σa=fySF=248 MPa1.5≈165 MPa

One drawback of this method is that the same factor is used whatever the certainty of the loading. Load and resistance factor design instead splits the factor of safety into components for the types of loading and of elements, so that the overall factor varies with the uncertainty in both.[12] Fire exposure erodes the factor. Kodur and Harmathy (2002) define the critical temperature as the temperature at which a material loses so much strength that it can no longer support the applied load. At that point the safety factor against failure becomes less than 1. North American standards (ASTM E119) assume 538 °C for structural steel.[13]

Because fire is treated as an unusual event, the safety factors on loads are commonly reduced for the fire limit state.[14] In general, dead load is multiplied by 1.0 and live load by a factor from 0.3 to 0.8, depending on occupancy. The British standard BS 5950-8 uses 0.50 for non-permanent imposed loads in offices. The New Zealand standard NZS 4203 adopts 1.0 times dead load plus 0.6 times live load for storage and 0.4 for other occupancies. The IFEG (2005) likewise state that smaller safety factors may be appropriate for structures in fire because of the low probability of fire.[15]

The fire side carries factors too. Lane (2008) describes the time-equivalent method of Eurocode 1, Part 1.2 (2002). In it, the design fire load density is the characteristic value multiplied by a combustion factor, two safety factors for the risk of fire activation and a factor for active fire-fighting measures.[16] The safety factor for compartment size rises from 1.10 at 25 m² to 2.13 at 10,000 m².

qf,d=qf,kmδq1δq2δn

Lane's case study of a 50-storey office building in the United Kingdom used a characteristic fire load of 511 MJ/m², a compartment factor of 1.9 and an occupancy factor of 1.0.[17] With a combustion factor of 1.0 and no credit for active measures, this gives 511 × 1.9 = 971 MJ/m². With a combustion factor of 0.8 and credits of 0.61 for automatic water extinguishing, 0.73 for detection and 0.78 for the fire brigade, the design value falls to 270 MJ/m². Lane adds that in Europe the acceptable level of risk is implicitly captured in the partial safety factors of each country's National Annex.[17]

Waterfall chart of design fire load density from 511 to 971 MJ/m2 with the compartment factor and down to about 270 MJ/m2 with active measures
Design fire load density in a UK office case study: the compartment safety factor raises the characteristic value, and the combustion factor and credits for active measures reduce it.

Safety factors in egress and hazard calculations

For life safety, the available safe egress time (ASET) must exceed the required safe egress time (RSET), and a safety factor is normally added to RSET to provide a margin of error.[18] Proulx (2002) suggests that, in a conservative approach, the time available should be at least twice the time required.[19] The IFEG (2005) say that ASET should exceed RSET by a margin agreed in the fire engineering brief for an absolute evaluation. In a comparative evaluation, the margin of the trial design should equal or exceed that of the deemed-to-satisfy design.[20]

Egress timeline from ignition showing RSET, the factored RSET and ASET, with the safety margin between RSET and ASET
Safety factor and safety margin in an ASET/RSET comparison: the design is acceptable if ASET is at least the factored RSET (multiplicative form) or exceeds RSET by the agreed margin (additive form).

tASET≥SF⋅tRSETtASET−tRSET≥Mmin

Van Rossum (2027) gives both forms and notes that the margin needed depends mainly on the reliability of the input values: the more reliable the input, the smaller the margin.[21] A sensitivity analysis can show whether the margin is sufficient. He reproduces a table of safety factors for this comparison from the Dutch standards for large fire compartments, NEN 6060 and NEN 6079.[21]

Determination method Model parameters New, RSET ≤ 3 min New, RSET > 3 min Existing, RSET ≤ 3 min Existing, RSET > 3 min
Rule-of-thumb method Conservative (80 %) 1.5 2.0 1.0 1.2
Rule-of-thumb method Nominal (50 %) 2.0 2.5 1.0 1.5
Alternative determination Conservative (80 %) 1.2 1.5 1.0 1.0
Alternative determination Nominal (50 %) 1.5 2.0 1.0 1.2

In the table, the factors are larger for new buildings than for existing ones, larger for long evacuations and larger when nominal rather than conservative input values are used.[21] As an illustration, a new building with a calculated RSET of 4 minutes would need an ASET of at least 10 minutes with the rule-of-thumb method and nominal values. With the alternative method and conservative values, it would need at least 6 minutes.

Grouped bar chart of NEN 6060 and NEN 6079 egress safety factors from 1.0 to 2.5 for new and existing buildings
Safety factors on RSET in the Dutch standards for large fire compartments, by building status, RSET length, determination method and input values.

The commentary on the first edition of C/VM2 (2013) shows margins in seconds in a worked hotel example.[22] For a fire in the foyer, RSET was 457 s. The fractional effective dose of carbon monoxide reached 0.3 at 1,065 s, a margin of 608 s, and with 1,000 occupants this was the only tenability criterion that applied. The design therefore complied. As an illustration, that margin corresponds to a ratio of ASET to RSET of about 2.3.

Hazard calculations use explicit factors as well. In a worked example by Budnick et al. (2008), a 2,000 kW pool fire on a 1 m² area exposes a target 5 m from its edge.[5] The point source method of Lawson and Quintiere gives 2.1 kW/m², and the recommended safety factor of two raises this to 4.2 kW/m². The correlation of Shokri and Beyler gives 1.02 kW/m², or 2.0 kW/m² with the same factor. In Beyler's (2002a) large-pool-fire comparisons, applying a factor of two makes the Shokri–Beyler correlation overpredict essentially all the data, and does the same for his point source model below 5 kW/m². He recommends no safety factor where a realistic result is required.[23]

Bar chart of radiant heat flux 5 m from a 2000 kW pool fire: calculated and with a safety factor of two, for two methods
Radiant heat flux at a target 5 m from the edge of a 2,000 kW pool fire, as calculated and with the recommended safety factor of two.

Clean agent design concentrations

For gaseous extinguishing systems, design concentrations are set by small-scale and large-scale testing, independent approval of hardware and the addition of design safety factors.[24] The minimum extinguishing concentration (MEC) for flammable liquids comes from the cup burner test. For Halon 1301, the minimum design concentration was historically the cup burner value plus 20 per cent. For n-heptane, about 3 per cent becomes 3.6 per cent. A separate minimum of 5 per cent in NFPA 12A (1992 edition, as described by DiNenno, 2002), however, meant an effective safety factor of 66 per cent for heptane. Only fuels with cup burner values above 4.2 per cent kept the 20 per cent factor.[24][25]

Cd=SF⋅CMECw=VSCd100−Cd

DiNenno and Taylor (2008) describe the safety factor of NFPA 2001 as a minimum that covers minor errors in design and installation. Examples are errors in room volume, cylinder filling and cup burner values.[4] The 20 per cent factor of earlier editions had been raised to 30 per cent for Class B hazards and for systems operated only by manual means. The ISO standard required 30 per cent for all systems.[4]

An analysis by Schlosser indicated that raising the factor from 20 to 30 per cent reduced the probability of system failure from about 15 to 10 per cent.[4] Variation in cup burner results and some full-scale tests also pointed to higher concentrations.[4] There was no evidence on whether older systems were adequate, and no reported incident in which an existing system designed with the 20 per cent factor had failed.[4]

As an illustration of the effect, the ISO Class B extinguishing concentration of the agent HFC-227ea is 6.9 per cent.[26] A safety factor of 1.2 gives a design concentration of 8.28 per cent, and 1.3 gives 8.97 per cent. The specific volume constants of the handbook give 0.137 m³/kg at an assumed minimum temperature of 20 °C.[27][28] For an assumed protected volume of 100 m³, the agent mass rises from about 66 kg to about 72 kg. The higher factor therefore adds about 9 per cent of agent.

Bar chart of HFC-227ea concentrations: extinguishing 6.9 per cent, design 8.28 per cent with factor 1.2 and 8.97 per cent with factor 1.3
HFC-227ea: ISO Class B extinguishing concentration and design concentrations with safety factors of 1.2 and 1.3, with the agent mass for an illustrative 100 m³ room at 20 °C.

A larger factor does not cure every weakness. NFPA 2001 introduced design factors in its 2000 edition, which add agent for installations where the minimum safety factor may not be sufficient. The only design factors quantified were for systems with multiple flow splits that protect more than four enclosures at once.[27] DiNenno and Taylor list conditions under which a higher safety factor should be considered, such as highly variable room volumes or unenclosable openings. They also stress that increasing the factor is not warranted in many applications. It will not improve system reliability if doors are not closed, and it is not a substitute for other design requirements.[27]

Choosing and combining safety factors

The IFEG (2005) list what the size of a safety factor should depend on. The factors are the redundancy in the design, the reliability of its components, the analysis methods and assumptions, the results of an uncertainty analysis, the acceptance criteria and the consequences of a fire.[29] Factors should be applied only at the end of a calculation sequence, not at every step, because that could lead to over-conservative outcomes. In a comparative evaluation, explicit factors should not be needed, because the same methods and assumptions are used for both designs.[29]

Notarianni (2002) explains why the place of application matters.[1] A factor applied to the final outcome criteria shows most clearly what margin has been achieved, but least clearly how to change the design. Inputs are often not linearly related to outcomes such as upper-layer temperature. Doubling the heat release rate, for example, does not give a safety factor of two on the time available for egress.[7] The size of an explicit factor, such as 1.5 or 2, depends as much on confidence in the equations as on the stakeholders' risk tolerance. Historical performance data can be used to set it; otherwise factors are usually set by expert judgement or mandated in policy.[7]

She also notes that there were no official rules for combining safety factors, and that specific numbers and justifications were lacking in the literature.[7] The effect of each factor should be tracked by repeat calculations. A factor applied to a mean may not cover a varied population: doubling the walking speed of a young, healthy person does not cover a person who walks at less than half that speed. Factors on two parameters do not simply add up, and an assumption can be conservative for one criterion but not for another. A soot yield chosen to be conservative for smoke detector activation, for instance, cannot also be conservative for life safety.[7]

Probabilistic approaches and limitations

Reliability methods offer a way to derive safety factors. Researchers at Lund University applied the first-order second-moment method, in which a safety index β measures the distance to the failure limit state and acts as an overall safety factor for the design.[30] They suggested a target β of 1.4, approximately equal to a probability of failure of 8 per cent given that a fire has started.[30] Meacham (2002) describes the same idea as a limit state function G, where the design must keep the probability that G falls below zero under a target value.[31] In stochastic models of fire spread, the safety factor itself can be a random variable. One model estimates the probability of fire spread through a barrier from the ratio of its equivalent fire resistance to the fire severity, both treated as log-normal.[32]

P(G<0)<Ptarget

Notarianni (2002) found that published numerical safety factors often lacked specific values and justifications.[7] Puchovsky and Hurley (2008) note that buildings to which engineering has traditionally been applied have an excellent fire record but an unknown safety margin. If that margin is too large, its extra cost burdens the economy.[33] In performance-based designs, most factors were developed ad hoc.[6] NUREG-1805 (2004) describes safety factors as a slightly more formal application of engineering judgement, applied for example to fire size, fire growth rate or hazard criteria.[34]

A larger factor does not always mean a safer design. The sources show the opposite in several cases: a factor on the wrong parameter can reduce safety for another criterion, and a higher agent concentration cannot compensate for doors left open.[7][27] The IFEG (2005) also note that less rigorous factors of safety may be appropriate for sensitivity studies, to avoid overly conservative outcomes.[29]

See also

Notes

  1. ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 Notarianni (2002), p. 5-47.
  2. ↑ 2.0 2.1 2.2 Hurley and Bukowski (2008), p. 3-133.
  3. ↑ National Research Council of Canada et al. (2005), Part 0 – New Zealand, p. 0.5-4.
  4. ↑ 4.0 4.1 4.2 4.3 4.4 4.5 DiNenno and Taylor (2008), p. 17-112.
  5. ↑ 5.0 5.1 Budnick et al. (2008), pp. 3-156–3-157.
  6. ↑ 6.0 6.1 6.2 6.3 Puchovsky and Hurley (2008), p. 3-174.
  7. ↑ 7.0 7.1 7.2 7.3 7.4 7.5 7.6 Notarianni (2002), p. 5-48.
  8. ↑ Iqbal and Salley (2004), p. F-49.
  9. ↑ Fleming (2008a), p. 16-56.
  10. ↑ Ministry of Business, Innovation and Employment (2025b), p. 31.
  11. ↑ 11.0 11.1 National Research Council of Canada et al. (2005), p. 3.4-8.
  12. ↑ 12.0 12.1 12.2 Manley (2008), p. 1-46.
  13. ↑ Kodur and Harmathy (2002), p. 1-163.
  14. ↑ Lane (2008), p. 19-93.
  15. ↑ National Research Council of Canada et al. (2005), p. 2.6-7.
  16. ↑ Lane (2008), pp. 19-82–19-83.
  17. ↑ 17.0 17.1 Lane (2008), pp. 19-86–19-87.
  18. ↑ Mowrer and Rosenbaum (2008), p. 3-189.
  19. ↑ Proulx (2002), p. 3-363.
  20. ↑ National Research Council of Canada et al. (2005), p. 1.2-22.
  21. ↑ 21.0 21.1 21.2 Van Rossum (2027), §7.1.
  22. ↑ Ministry of Business, Innovation and Employment (2013), p. 209.
  23. ↑ Beyler (2002a), pp. 3-273–3-275.
  24. ↑ 24.0 24.1 DiNenno and Taylor (2008), p. 17-110.
  25. ↑ DiNenno (2002), p. 4-187.
  26. ↑ DiNenno and Taylor (2008), p. 17-96.
  27. ↑ 27.0 27.1 27.2 27.3 DiNenno and Taylor (2008), pp. 17-113–17-114.
  28. ↑ DiNenno (2002), p. 4-189.
  29. ↑ 29.0 29.1 29.2 National Research Council of Canada et al. (2005), p. 1.2-26.
  30. ↑ 30.0 30.1 Notarianni (2002), pp. 5-48–5-49.
  31. ↑ Meacham (2002), p. 5-160.
  32. ↑ Ramachandran (2002c), p. 3-392.
  33. ↑ Puchovsky and Hurley (2008), p. 3-168.
  34. ↑ Iqbal and Salley (2004), p. B-84.

References

  • Beyler, C. L. (2002a). Fire hazard calculations for large, open hydrocarbon 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. 3-268–3-314). National Fire Protection Association; Society of Fire Protection Engineers.
  • Budnick, E. K., Hunt, S. P., & Wright, M. T. (2008). Closed form enclosure fire calculations. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-145–3-166). National Fire Protection Association.
  • DiNenno, P. J. (2002). Halon replacement clean agent total flooding systems. 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. 4-173–4-200). National Fire Protection Association; Society of Fire Protection Engineers.
  • DiNenno, P. J., & Taylor, G. M. (2008). Halon and halon replacement agents and systems. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 17-93–17-121). National Fire Protection Association.
  • Fleming, R. P. (2008a). Hanging and bracing of water-based systems. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 16-55–16-70). 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.
  • Kodur, V. K. R., & Harmathy, T. Z. (2002). Properties of building materials. 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. 1-155–1-181). National Fire Protection Association; Society of Fire Protection Engineers.
  • Lane, B. (2008). Approaches to calculating structural fire resistance. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 19-77–19-126). National Fire Protection Association.
  • Manley, B. E. (2008). Fundamentals of structurally safe building design. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 1-31–1-49). National Fire Protection Association.
  • Meacham, B. J. (2002). Building fire risk 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. 5-153–5-175). National Fire Protection Association; Society of Fire Protection Engineers.
  • Ministry of Business, Innovation and Employment. (2013). Commentary for Building Code clauses C1–C6 and Verification Method C/VM2.
  • Ministry of Business, Innovation and Employment. (2025b). Verification Method C/VM2: Framework for fire safety design (2nd ed.).
  • Mowrer, F. W., & Rosenbaum, E. R. (2008). Overview of performance-based fire protection design. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-183–3-194). 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.
  • Notarianni, K. A. (2002). Uncertainty. 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. 5-40–5-64). National Fire Protection Association; Society of Fire Protection Engineers.
  • Proulx, G. (2002). Movement of people: The evacuation timing. 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. 3-342–3-366). National Fire Protection Association; Society of Fire Protection Engineers.
  • Puchovsky, M. T., & Hurley, M. J. (2008). Performance-Based codes and standards for fire safety. In A. E. Cote, C. C. Grant, J. R. Hall, Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 3-167–3-181). National Fire Protection Association.
  • 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.
  • Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.