Skip to content
Upcoming courses

Fire risk assessment

Tools
From Fire Safety Academy Wiki
Contents
  1. 1Definitions
  2. 2Steps of an assessment
  3. 3Methods
  4. 4Calculating risk from scenarios
  5. 5Use in design and regulation
  6. 6Existing buildings, operations and management
  7. 7Acceptability, uncertainty and limits
  8. 8See also
  9. 9Notes
  10. 10References

A fire risk assessment is a structured process for identifying fire hazards, estimating the likelihood and consequences of the resulting fires, and judging whether the risk is acceptable.[1][2] It can cover a building, a plant or an activity. Watts and Hall summarise the underlying questions as: what could happen, how bad would it be, and how likely is it to happen.[3] A fire risk assessment builds on a fire hazard assessment, which identifies what could go wrong, and then weights each consequence by the likelihood of the scenario that causes it.[4]

Methods range from simple checklists and narratives through risk indexes and risk matrices to fully probabilistic analyses with event trees.[2][5] The results are used to choose between designs, to show that a performance-based design is at least as safe as a compliant reference building, and to set priorities for protection in existing buildings and plants.[6][7] Every assessment ends in a value judgement, because zero risk is not achievable and no universally acceptable level of fire risk exists.[2]

Definitions

The terminology of risk is not consistent. A committee of the Society for Risk Analysis identified 17 different definitions of risk, and the words analysis and assessment are often used interchangeably.[1] In the SFPE Handbook, Watts and Hall define risk as the potential for unwanted, adverse consequences to life, health, property or the environment.[1] When they speak of fire risk, they mean the uncertainty of loss from fire.[8] The risk of an event is usually estimated as the expected value of its probability multiplied by its consequence.[1] The International Fire Engineering Guidelines (IFEG, 2005) use the same idea and define risk as the product of the probability and the consequence of an event.[9]

Watts and Hall distinguish several related activities. Risk assessment is the process of establishing information on acceptable levels of a risk, on actual levels of risk, or on both.[1] Risk evaluation is the part of the assessment in which the significance and acceptability of a risk are judged. Risk analysis is the wider examination that also includes options for managing the risk.[1] A fire hazard analysis, by contrast, looks at what could happen and how bad it would be, without analysing likelihood.[8]

Meacham describes risk as the possibility of an unwanted outcome in an uncertain situation. It depends on three factors: the loss of or harm to something that is valued, the hazard that may cause it, and a judgement about how likely the loss is.[10] In building fire risk analysis, the things that are valued are life safety, property, business continuity, the environment and heritage.[11]

Steps of an assessment

Watts and Hall give a general sequence of five steps.[2] The assessor first identifies the fire hazards, then quantifies the consequence and probability of each. Next, the options for controlling the hazards are identified and their effect on the risk is quantified. Finally, the appropriate protection is selected.[2] The quantification can rest mainly on hard data and established science, or mainly on expert judgement.[2]

Five boxes in sequence: identify hazards, quantify consequence and probability, identify control options, quantify their effect, select protection.
The five generic steps of a fire risk assessment. Based on Watts and Hall (2002), p. 5-4.

Hazard identification usually starts with a survey of the building or process. Meacham lists the information gathered: potential ignition sources, potential fuels, the arrangement of fuel packages, the configuration of the building and compartments, and the fire safety features present.[12] Historical loss data help to avoid focusing on unusual hazards while overlooking more common ones.[12] The hazards are then expressed as fire scenarios, each a sequence of events from ignition to an outcome.[8]

The consequence part asks what harm each scenario would cause. Meacham notes that valuing consequences is often less familiar to engineers than assessing the fire itself. Beyond death and injury, it may include reduced quality of life, smoke and water damage, and loss of market share.[10] The likelihood part asks how often each scenario occurs. Meacham stresses the difference between frequency and probability. Twenty toaster fires per year is a frequency; with 20 million toaster starts per year, the probability of a fire per start is one in a million.[13]

Methods

Watts and Hall classify fire risk methods into four groups: checklists, narratives, indexing and probabilistic methods.[2] Checklists and narratives are not quantitative. They identify hazards, control options and protection, but skip the quantification of probability and of the effect of options.[2] Risk indexing assigns values to selected positive and negative fire safety features based on judgement and experience. It combines them arithmetically into a single value that can be compared with a standard.[5] Introduction to Fire Safety Engineering notes that qualitative assessments often use risk classes such as very low, low, medium, high and very high.[14]

Risk matrices sit between indexes and fully probabilistic methods.[5] Meacham describes a matrix in which the largest realistic consequence of each type of loss is ranked as negligible, low, moderate or high.[15] The frequencies are ranked in bins as well.[16]

In Meacham's example, anticipated events occur more often than once per 100 years (more than 10−2 per year). Unlikely events occur between 10−4 and 10−2 per year, and extremely unlikely events between 10−6 and 10−4 per year.[16] The consequence predictions should bound all outcomes at the 95th percentile or better.[15] Frequencies should refer to exceeding a given loss rather than to single scenarios: 30 scenarios of 10−7 per year together give 3 × 10−6 per year.[15]

Risk matrix: four consequence levels against four frequency bins; each cell marked high, moderate, low or negligible risk.
Example risk ranking matrix: each combination of consequence level and frequency bin is assigned a risk level. Based on Meacham (2002), pp. 5-166–5-167.

Probabilistic methods are the most informative, because their results can be traced through explicit assumptions, data and relationships.[5] The most common tools are event tree analysis, which follows the possible outcomes after an initiating event in time order, and fault trees, which trace the logical causes of a failure.[5] Meacham explains that an event tree starts with an initiating event and asks, for each protection system or strategy, whether it succeeds or fails.[17] A full multi-scenario analysis of this kind is a form of quantitative risk assessment, also called probabilistic risk assessment, which is widely used in the chemical process industry and for nuclear facilities.[18]

Calculating risk from scenarios

In a quantitative assessment, the risk is built up scenario by scenario. For each scenario, the analyst estimates its probability and its consequence, multiplies them, and adds the results over all scenarios.[19][18] The IFEG express this for life safety as the expected number of deaths, given that a fire has started:[19][20]

Two event trees: system A works 0.9 or fails 0.1, risk 0.2 persons; system B works 0.7 or fails 0.3, risk 0.6 persons.
Expected number of people exposed per fire for two systems with 90% and 70% reliability: the risk of system B is three times that of system A. Based on Australian Building Codes Board (2020b), pp. 118–119.

END=∑jPjNj

Symbol Description Unit
END expected number of deaths for a fire, summed over all scenarios persons
j index of the fire scenario –
Pj probability that scenario j develops after ignition –
Nj number of occupants exposed to untenable conditions in scenario j persons

The relation is valid when the scenarios are distinct paths of one event tree, so that each possible outcome is counted once.[21] The consequences Nj come from deterministic analysis of each scenario, and the probabilities from reliability data or statistics.[20] Meacham gives the same sum with the frequency of each scenario instead of its probability after ignition, which gives a risk per unit of time.[18][13] Dividing the expected number of deaths over the building's life by the number of occupants and the design life gives the expected risk to life (ERL) used in risk-cost models.[19]

Worked example. The Australian handbook for the Fire Safety Verification Method (FSVM) compares two fire safety systems.[22] If either system works, no occupants are exposed to untenable conditions; if it fails, two people are exposed. System A, the reference, has a reliability of 90%, and system B, the proposal, 70%.[22] For system A, the expected number exposed per fire is 0.9 × 0 + 0.1 × 2 = 0.2 persons. For system B it is 0.7 × 0 + 0.3 × 2 = 0.6 persons.[23]

A purely deterministic analysis, which looks only at the case in which each system works, would find the two systems equivalent. The risk calculation shows that system B exposes occupants to three times the risk of system A.[23] The example shows why system reliability must enter the comparison whenever it differs between the proposed and the reference design.[22]

Use in design and regulation

The most common use of fire risk analysis is to support a choice, for example between two designs or two formulations of a code.[6] In Australia, the FSVM handbook states that a risk assessment will generally be needed when the reliability of fire safety systems differs appreciably between a proposed design and the reference building.[22] A detailed quantitative risk assessment is the most complete method, but it often needs substantial resources.[23] For designs with minor differences from the reference building, simple event trees for the probability can be combined with deterministic analysis of the consequences.[23] The proposed design must then show that, for each scenario, the risk to life is no greater than in the reference building.[7]

The handbook builds its event tree from the manage-fire branch of the fire safety concepts tree, with added fire brigade intervention.[24] It points out how the parts of a strategy depend on each other. For unassisted evacuation, detection and alarm must occur, occupants must move, and movement means and a safe destination must be provided. The person must also be able to leave without assistance; if any link fails, escape relies on fire brigade rescue.[25] A failed detection system may also delay the call to the fire brigade and fail to start smoke control.[25]

Event tree from fire ignition through occupant suppression, sprinklers and fire brigade to outcomes from minimal to catastrophic.
Simplified event tree derived from the manage-fire branch of the fire safety concepts tree: each later system acts only if the earlier ones fail. Based on Australian Building Codes Board (2020b), pp. 120–121.

The IFEG describe a probabilistic event tree approach that compares the expected risk to life of a design with an acceptance value. That value may be a fixed number, or the expected risk to life of a design that complies with the prescriptive provisions.[26] They warn that probabilistic methods need much statistical data, can be time-consuming, and may be hard to validate.[27] In the Netherlands, Introduction to Fire Safety Engineering notes that there are no risk-based limit values.[14] Acceptable residual risk must therefore be agreed in advance, for example by comparison with a reference situation that complies with the regulations.[28]

Watts and Hall note that fire risk analysis is far more common in the chemical process industry and in nuclear power plants than in other fields.[6] For US nuclear power plants, Siu describes how the core damage frequency due to fire is split into three terms. These are the frequency of the fire scenario, the probability of fire damage to critical equipment, and the probability of core damage given that damage.[29] The three terms correspond to the layers of defence in depth for nuclear plants: fire prevention, fire detection and suppression, and fire mitigation.[29] Qualitative and conservative quantitative screening removes areas that are not expected to contribute to risk, so effort can focus on the scenarios that matter most.[29]

Existing buildings, operations and management

Fire risk assessment also applies to buildings and plants in use. Meacham describes safety reviews of existing facilities, carried out as scheduled visual inspections to identify conditions or procedures that could lead to accidents.[12] "What if" analysis and hazard and operability (HAZOP) studies ask systematically what happens when something goes wrong.[12] Checklists help inexperienced staff to focus on hazards, but users may ignore hazards that are not on the list.[12]

Operations and their control are central to such assessments. Hot work means activities that involve heat, sparks or open flame, such as cutting, welding and grinding.[30] The IFEG list adherence to hot work procedures among the measures that control fire initiation.[31] In US nuclear fire risk assessment, the frequency and nature of maintenance, hot work procedures and housekeeping affect the likelihood of transient fires. The number of work orders in an area is used to rank it.[32] Where a fire watch is posted during hot work, a fire is likely to be detected quickly.[33]

The assessment depends on whether protection works when needed. Meacham distinguishes availability, when a system is in service, from reliability, when an available system actually functions. A detector covered with tape is available but not reliable.[34] Watts and Hall list typical questions: is the battery in the smoke alarm working, is the sprinkler valve open, and is the fire door working or blocked open.[6] Fire safety management, including maintenance and inspection, therefore directly changes the numbers in an assessment.

Acceptability, uncertainty and limits

Deciding whether a risk is acceptable is a value judgement. Watts and Hall note that acceptable risk for a practising engineer is likely to be risk acceptable to the client. For an approving authority, it is risk acceptable to society or to a community.[2] Zero risk is not an achievable goal, and reducing fire risk may increase other risks, as with the effect of halon on the ozone layer.[2] Experts, the public, enforcers and managers may value the same measure, such as residential sprinklers, very differently.[3][8]

Uncertainty is unavoidable. Meacham groups it into scientific uncertainty, uncertainty about human factors, uncertainty in perceptions and values, and uncertainty in decision-making.[35] Fire is a relatively rare event, so data are limited, and changes to buildings and contents make the past an imperfect guide to the future.[13] Even the risk-cost models described by Meacham contain conservative simplifications. They should be used for comparing designs, not for absolute assessments of life risk.[36]

The number of scenarios is a practical limit. The IFEG note that an exhaustive measure of risk to life would require every possible fire scenario. Because the computational effort grows with the number of scenarios, the problem must first be simplified.[19] A fire risk assessment is therefore a model of the risk, built from chosen scenarios, data and judgements. Its value lies in making those choices explicit, so that they can be checked and discussed.[5]

See also

Notes

  1. ↑ 1.0 1.1 1.2 1.3 1.4 1.5 Watts and Hall (2002), p. 5-3.
  2. ↑ 2.00 2.01 2.02 2.03 2.04 2.05 2.06 2.07 2.08 2.09 Watts and Hall (2002), p. 5-4.
  3. ↑ 3.0 3.1 Watts and Hall (2002), p. 5-1.
  4. ↑ Bukowski (2002), p. 5-231.
  5. ↑ 5.0 5.1 5.2 5.3 5.4 5.5 Watts and Hall (2002), p. 5-5.
  6. ↑ 6.0 6.1 6.2 6.3 Watts and Hall (2002), p. 5-6.
  7. ↑ 7.0 7.1 Australian Building Codes Board (2020b), p. 124.
  8. ↑ 8.0 8.1 8.2 8.3 Watts and Hall (2002), p. 5-2.
  9. ↑ National Research Council of Canada et al. (2005), p. 0.5-4 (Part 0, Australia).
  10. ↑ 10.0 10.1 Meacham (2002), p. 5-153.
  11. ↑ Meacham (2002), p. 5-154.
  12. ↑ 12.0 12.1 12.2 12.3 12.4 Meacham (2002), p. 5-155.
  13. ↑ 13.0 13.1 13.2 Meacham (2002), p. 5-159.
  14. ↑ 14.0 14.1 Van Rossum (2027), §1.4.
  15. ↑ 15.0 15.1 15.2 Meacham (2002), p. 5-166.
  16. ↑ 16.0 16.1 Meacham (2002), p. 5-167.
  17. ↑ Meacham (2002), p. 5-156.
  18. ↑ 18.0 18.1 18.2 Meacham (2002), p. 5-161.
  19. ↑ 19.0 19.1 19.2 19.3 National Research Council of Canada et al. (2005), p. 2.3-5.
  20. ↑ 20.0 20.1 National Research Council of Canada et al. (2005), p. 2.3-4.
  21. ↑ National Research Council of Canada et al. (2005), p. 2.3-2.
  22. ↑ 22.0 22.1 22.2 22.3 Australian Building Codes Board (2020b), p. 118.
  23. ↑ 23.0 23.1 23.2 23.3 Australian Building Codes Board (2020b), p. 119.
  24. ↑ Australian Building Codes Board (2020b), p. 120.
  25. ↑ 25.0 25.1 Australian Building Codes Board (2020b), p. 122.
  26. ↑ National Research Council of Canada et al. (2005), p. 1.2-23.
  27. ↑ National Research Council of Canada et al. (2005), p. 1.2-22.
  28. ↑ Van Rossum (2027), §1.2.
  29. ↑ 29.0 29.1 29.2 Siu (2002), p. 5-216.
  30. ↑ Iqbal and Salley (2004), p. F-37.
  31. ↑ National Research Council of Canada et al. (2005), p. 1.2-10.
  32. ↑ Nowlen et al. (2005b), p. 6-21.
  33. ↑ Nowlen et al. (2005b), p. 11-28.
  34. ↑ Meacham (2002), p. 5-162.
  35. ↑ Meacham (2002), p. 5-160.
  36. ↑ Meacham (2002), p. 5-165.

References

  • Australian Building Codes Board. (2020b). Handbook: Fire Safety Verification Method (Print version 1.1).
  • Bukowski, R. W. (2002). Fire hazard assessment for transportation vehicles. 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. 5-227–5-233). 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.
  • 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.
  • 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.
  • Nowlen, S. P., Najafi, B., Wyant, F. J., Forester, J., Kazarians, M., Kolaczkowski, A., Joglar, F., Funk, D., Hannaman, G., & Anoba, R. (2005b). EPRI/NRC-RES fire PRA methodology for nuclear power facilities: Vol. 2. Detailed methodology (NUREG/CR-6850; EPRI 1011989). Electric Power Research Institute; U.S. Nuclear Regulatory Commission, Office of Nuclear Regulatory Research.
  • Siu, N. (2002). Fire risk assessment for nuclear power plants. 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. 5-214–5-226). National Fire Protection Association; Society of Fire Protection Engineers.
  • Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.
  • Watts, J. M., Jr., & Hall, J. R., Jr. (2002). Introduction to 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. 5-1–5-7). National Fire Protection Association; Society of Fire Protection Engineers.
Retrieved from "https://fire-safety.academy/w/index.php?title=Fire_risk_assessment&oldid=2625"