Quantitative Risk Assessment in Fire Safety

Ganapathy Ramachandran, David Charters

• Publisher: Routledge
• Year: 2011
• eBook ISBN: 9781135818388
• Edition: 1
• Topic: Architecture
• Subtopic: Architecture Methods & Materials

Fire safety regulations in many countries require Fire Risk Assessment to be carried out for buildings such as workplaces and houses in multiple occupation. This duty is imposed on a "Responsible Person" and also on any other persons having control of buildings in compliance with the requirements specified in the regulations.

Although regulations only require a qualitative assessment of fire risk, a quantitative assessment is an essential first step for performing cost-benefit analysis of alternative fire strategies to comply with the regulations and selecting the most cost-effective strategy. To facilitate this assessment, various qualitative, semi-quantitative and quantitative techniques of fire risk assessment, already developed, are critically reviewed in this book and some improvements are suggested.

This book is intended to be an expanded version of Part 7: Probabilistic risk assessment, 2003, a Published Document (PD) to British Standard BS 7974: 2001 on the Application of Fire Safety Engineering Principles to the Design of Buildings. Ganapathy Ramachandran and David Charters were co-authors of PD 7974 Part 7.

Quantitative Risk Assessment in Fire Safety is essential reading for consultants, academics, fire safety engineers, fire officers, building control officers and students in fire safety engineering. It also provides useful tools for fire protection economists and risk management professionals, including those involved in fire insurance underwriting.

1 Introduction

We all take risks all the time, whether it is crossing the road, driving to work or watching television. The risk may vary from being knocked down, to being in a car accident or suffering ill health due to lack of exercise. The same can be said of fire safety in buildings. As long as we occupy buildings where there is a chance that ignition sources and combustible materials may be present together, there will be a risk of death and injury due to fire in addition to property damage. We need not be fatalistic, however, this simply identifies the need to manage the risk. It also indicates that, although we should work towards reducing risk, the ultimate goal of zero risk is not currently a realistic expectation.

As Benjamin Franklin once said, ‘But in this world nothing is certain, but death and taxes’. It follows that whilst we live there is a risk of death and the only way of not dying is not to live in the first place. This may be self-evident, but it is very important when we start to consider specific risks that we consider them in the context of other risks.

There is a practical benefit to looking at risks in context. Society may decide that it would like to dedicate more resources to addressing one risk than another. For example, for healthcare, it may typically cost about £20,000 to save a life, whereas for fire safety in buildings, it may cost more than say £1million to save a life (Charters 1996). Therefore, society (and/or its representatives) could decide to put more resources into healthcare than into fire safety in buildings. Equally, society may be more concerned about the suffering of people killed and injured by multi-fatality fires than it is about the provision of every possible healthcare intervention to all patients, irrespective of need or prognosis.

For fire safety in buildings, the annual fire statistics indicate that there is a finite level of fire risk in buildings (Office of the Deputy Prime Minister 2005). This may also indicate that if a building complies with the appropriate fire safety standards, then its level of fire risk is broadly tolerable (or possibly acceptable). It could also be said that applying the fire standards to a non-standard building could result in intolerable levels of fire risk. However, no criteria for fire risk in buildings have been set in the UK (British Standards Institute PD 7974 Part 7 2003).

For the fire safety engineering of a non-standard building, this means that the level of risk should be designed to be the same or lower than that for an equivalent standard building (Office of the Deputy Prime Minister 2005; British Standards Institute BS 7974 Code of Practice 2001). Since there are no quantitative risk criteria, this means comparison of a non-standard building with a compliant building. Similarly, since regulations are not generally framed in terms of risk, this results in an assessment of physical hazards and the balancing of a qualitative arguments about risk. So, in conclusion, we could say that, with respect to fire safety in buildings, we are all taking a risk but rarely, if ever, is it calculated.

The risk of an undesirable event can be defined as the combination of:

• the severity of its outcome (its consequences); and

• how often it might happen (its frequency) (British Standards Institute PD 7974 Part 7 2003).

If we simply consider the potential consequences of fires, then we may not be adequately addressing fires with lower consequences but whose risk is much higher due to their frequency. It is easy to imagine very severe fire events in buildings that have no fire precautions. We can identify ignition sources and combustible material, estimate how quickly a fire would grow and how quickly untenable conditions would develop. We can then consider the potential number of occupants, their likely behaviour and how quickly they might escape from the building. It is less easy to imagine very severe events in buildings with many fire precautions, but they can, and do, occur. Severe fire events like these seem to catch us by surprise and are characterised by a series of failures of the many fire precautions. The lower frequency of these events leads to the sense of surprise. The fact that a series of failures had to occur before the severe event occurred indicates that we had ‘defence in depth’. Defence in depth is where many systems are present, but only a few need to work to achieve a safe outcome. Finally, this indicates that all fire precautions have a level of unreliability and occasionally many of the fire precautions may fail at the same time, leading to the severe event. Fortunately, these very severe events are relatively rare, but this rarity means that it is very difficult to assess their frequency directly. For example, if we have a thousand buildings of a certain type and we want to address fires that may occur once every million building years, then on average this fire could be expected only to happen once every thousand years. This is a long period over which to collect and analyse data and make a decision about the adequacy of the fire precautions. What makes it even harder is that this ‘one in a million building years’ event could happen tomorrow and the day after. This makes the direct estimation of high consequence/low frequency events very uncertain. Therefore, we use events that happen more often to estimate the frequency of rarer events. The frequency of ignition is the usual starting point for fire safety in buildings. To estimate the frequency of the severe fire events we need to understand how ignition may, or may not, lead to them. This involves the study of the reliability of fire safety systems. For example, probabilities can be attributed to the following series of events:

• Does the fire grow?

• If yes, is the fire detected early?

• If yes, is the fire extinguished using first aid fire fighting?

• Is the fire suppressed and/or vented by automatic systems?

• Does the fire spread beyond the compartment of fire origin?

In this way we can assess the impact of fire retardant or non-combustible materials, fire detection, extinguishers, sprinklers, vents, compartmentation etc on reducing the frequency of severe fires if, say, all these fire precautions fail.

We can combine the frequencies and consequences of these severe events to estimate their levels of risk. The levels of risk can be expressed as a frequency distribution (or ‘Fn’ curve) of risks, with lower risk events near the origin and higher risk events to the upper right-hand side. See Section 1.4 on acceptance criteria for risk and the role of Fn curves. Some events may have trivial or no consequences and so can be discounted from the analysis. Other events are high risk and should be addressed.

Because there are no quantitative criteria for the fire risk in buildings in the UK, it is not possible to say in absolute terms when the risk is low enough. However, if we know where a compliant standard building lies on the distribution, it is possible to do it by comparison. So we can say that we are safe enough, when the risk is lower, than a compliant standard building.

1.1 Fire engineering

Fire engineering can be defined as (BS 7974 2001):

the application of scientific and engineering principles to the protection of people, property and the environment from fire.

Fire engineering can address one or more objectives. These objectives can include:

• life safety;

• property loss prevention/business continuity; and

• environmental protection.

Fire engineering is generally more demanding, technically and in terms of resources, than the application of simple prescriptive fire safety guidance used to support building regulations (ADB 2007). Therefore, fire engineering is generally used where simple fire safety guidance may not adequately address the fire scenarios or issues of concern. Simple fire safety guidance may not adequately address the fire scenarios or issues of concern when the building is large, complex or unique, or when application of the simple prescription conflicts with the function of the building (usually rendering the fire precaution highly unreliable) or when application of the simple rules is not the most cost-effective approach. Other factors that might indicate where fire engineering is typically used include:

• when there is an atrium;

• when there are multiple purpose groups in one building;

• where a highly innovative design is used to facilitate the function of the building; or

• when there are unique or challenging fire hazards.

Examples where fire engineering is typically used include the larger assembly buildings, hotels, hospitals, industrial and commercial premises, transport interchanges and tunnels, landmark, heritage and headquarters buildings, ships and offshore installations.

1.2 Deterministic approaches

Deterministic approaches use quantitative analysis of physical processes, such as smoke movement and evacuation, to aid decision-making during the design process (BS 7974 2001). To understand the deterministic approaches, a more detailed consideration of the design process may be of benefit.

Design can be characterised as an essentially creative, largely intuitive, often divergent process where many design parameters are manipulated in three and sometimes four dimensions, to meet multiple design objectives. Design objectives may include function, efficiency, cost, buildability, safety, durability, reliability, aesthetics etc. Design parameters may include the location of the site, the size and interconnection of different spaces, means of access and egress, the number of levels, the type of construction etc. For fire safety design, the main design objective is life safety (occupants, fire fighters and others in the vicinity), but other objectives such as property protection/business continuity and the environment may also be included. In many respects, the use of simple prescriptive guidance suppresses the design aspect of fire safety and can encourage the perception that there is only one solution and ‘this is it’. However, the nature of many modern buildings means that simple fire safety rules are increasingly difficult and costly to apply. In these circumstances alternative fire safety design solutions are required to ensure that an appropriate level of safety is achieved. This is where analysis is used to assess the level of safety and indicate whether the fire safety design objective(s) have been met. Analysis can be characterised as an essentially logical, structured, rigorous process which takes certain input parameters, undertakes a series of operations/calculations and produces one or more output parameters. For fire safety analysis, the input parameters could include the material contents and their burning characteristics, size of a space, the number, width and distribution of its exits, the number of occupants, a walking speed and a flow rate through the exit. Deterministic analysis can be defined as the use of point values for the above variables and a purely physical model. Through the use of deterministic egress calculation methods a single value of the output parameter of time for occupants to move through an exit can be calculated (see Section 3.6 on consequence analysis). This single value is regarded as an exact value, ignoring the uncertainties governing the input and output parameters.

Decision-making can be characterised as an essentially convergent process of identifying an issue that needs to be addressed, gathering information on it (including that from analysis) and reaching a conclusion based on the evidence. For fire safety, the issue could be the number and width of exits required for egress from a large and complex space. Based on knowledge of the number and width of exits for smaller/simpler spaces, the basis of the input data, analytical model and results of the analysis, a decision on the appr...

Table of contents

• Cover
• Halftitle
• Title
• Copyright
• Contents
• List of Figures
• List of Tables
• About the authors
• Preface
• Acknowledgements
• 1. Introduction
• 2. Qualitative and semi-quantitative risk assessment techniques
• 3. Quantitative risk assessment techniques
• 4. Acceptance criteria
• 5. Initiation
• 6. Design fire size
• 7. Fire spread beyond room of origin
• 8. Performance and reliability of detection, alarm and suppression
• 9. Performance and reliability of detection, alarm and suppression
• 10. Performance and effectiveness of fire service intervention
• 11. Whole project analysis
• 12. Interactions
• 13. Combining data from various sources – Bayesian techniques
• Index