Fire Australia 2017 Quantification of Fire Safety Fire Safety Engineering Stream Title Authors Topics Case Study: Risk based approach for the design of a transport infrastructure Edmund Ang, Imperial College London AECOM Australia E: chin.ang13@imperial.ac.uk edmund.ang@aecom.com Ray Christie, Transport for NSW E: ray.christie@transport.nsw.gov.au - Risk based fire safety engineering - Methodology and case study Abstract In Australia and internationally, the use of the risk based approach for the fire safety design in buildings and infrastructure are gradually recognised as a viable option in addition to the currently adopted absolute, i.e. safe or unsafe design approach. The risk based design has a significant advantage as it recognises the inherent probabilistic nature of a fire risk, and allows the level of mitigation measures needed plus the acceptance criteria of the risk to be objectively quantified in line with the probability. In addition, the risk based approach recognises although there are countless options to potentially eliminate the risks, not all of these options are effective from a cost benefit perspective. In this case study, we present the 7 steps risk based approach we adopted for the design of a 250 m long tunnel to evaluate the necessity to provide a new fire hydrant system in the tunnel by considering the risk of a fire to the safety of the passengers, to the fire fighters, to the surrounding areas and to the train operations. The 7 steps process of the risk based approach includes establishing the baseline requirement, understanding the acceptance criteria for the risk, to conducting the probabilistic study of the fire risks and the consequences of the risks. We explained in this case study the process to integrate the outcome of the risk and consequences assessment with the potential mitigation measures being considered. This is a key advantage and a key step in the risk based approach as this method allows us to establish if the mitigation measure is cost effective, i.e. if the cost outweighs the consequences of the risk. Whilst the technical approach to a risk based design may change depending on the type of projects and environment, a suitably implemented 7 step risk based approach can ensure a more objective outcome to enable decision makers to reach closer to a quantifiably fair and impartial decision.
1.0 Introduction to Risk Based Design In Australia and internationally, for the fire safety design of both buildings and infrastructure, the fire safety design of these developments are often carried out using a combination of prescriptive guidance, e.g. the Building Code of Australia [1], the National Fire Protection Association [2] and performance based design. For the prescriptive guidance, the design based on this approach is relatively straightforward whereby practitioners need only to follow the prescribed recommendations in these standards. On the other hand, performance based fire engineering design is used when the conventional prescriptive guidance cannot be complied with, or a more efficient and safer design can be achieved by the use of the performance based design. In the performance based design, for example the IFEG [3] notes a number of approaches that can be taken for the quantitative analysis including the deterministic approach, i.e. safe or unsafe or a comparative approach, i.e. benchmarking against code guidance. Whilst the IFEG recognises a risk based approach exists, and this method has been applied in establishing the performance requirements of code guidance [4], the risk based design approach has not been applied extensively by practitioners in the design for fire safety engineering over the last 20 years compared to the deterministic approach. Figure 1: Performance Based Solutions with an Absolute and Probabilistic Approaches In this paper, we discuss the challenges to the widespread adoption of the risk based design, and examine a case study on the application of this approach in a transport infrastructure. 2.0 Challenges to the Adoption of the Risk Based Design The risk based design has a significant advantage as it recognises the inherent probabilistic nature of a fire risk, and allows the level of mitigation measures needed plus the acceptance criteria of the risk to be objectively quantified in line with the probability. In addition, the risk based approach recognises although there are countless options to potentially eliminate the risks, not all options are effective from a cost benefit perspective. The risk based approach allows this to be quantified objectively. This is compared to the deterministic approach whereby the viability of a design is based strictly on a pass fail criteria, ignoring the probability of an event from occurring, e.g. 1 in
1,000 years or 1 in 10,000 years and the cost benefit on the option to reduce the risk, e.g. is $10,000,000 a justifiable cost in eliminating 5 major injuries per 1,000 statistical years. Whilst the risk based approach has an advantage over a deterministic approach, we recognise this approach s inherent challenges and we see three main challenges with this: Firstly for a probabilistic analysis, the key input is the quality of the historical data for the events considered. This is because the outcome of a probabilistic study is determined by the reliability and accuracy of the data collected. This is compounded by the complexities including if the historical data is only applicable to one region and the methodology for the data collection, e.g. the precise definition of a major and minor injury. To overcome this challenge, an industry body, e.g. Engineers Australia can play a key part in setting a mutually agreeable standard and methodology for the collection of data. Secondly, unlike probabilistic analysis with a pass fail criteria, the risk based approach explicitly recognises an event could occur but with a varying likelihood. This means from an approvals perspective, there needs to be legally established acceptance criteria that clearly define what level of risks and consequences are acceptable both from design and legal perspectives. For this, we believe the government or a state agency is in the strongest position to establish a minimum benchmark for the acceptance of a risk based design. Thirdly for the cost benefit analysis which is integral to a probabilistic analysis, a key topic to be considered is assigning a monetary value to a human life, i.e. the value of a statistical life. A guideline published by Transport for NSW [5] provided a number of references to the value of a statistical life. Table 1: Value of Statistical Life [5] Value of Statistical Life in NSW (VSL) $6,698,897 Value of Statistical Life Year (VSLY) *Based on a real discount rate of 4% over 40 years $325,434 Whilst this method is commonly adopted in other disciplines, e.g. National Health Service (NHS) UK [6] in evaluating whether a type of treatment should be funded, this is still perceived as a politically challenging topic in the fire safety industry. To overcome this challenge, we believe a key element is through engaging with the public to help the public and the stakeholders to understand this is a necessary step in ensuring an equitable outcome is achieved in the decision making process. 3.0 7 Steps Process to Risk Based Design From our experience, a possible misconception of the risk based design approach is the method of analysis is focused solely on the probability of the event, i.e. if the probability of an event is deemed low enough there would not be a need to put in place a mitigation measure. Although a probability analysis is a critical element, it only forms part of the process in the risk based design as illustrated in Figure 2. A key component of the risk based design is that it combines the probability of an event with the consequences and the cost to implement it. This way, the cost effectiveness of the mitigation measure can be benchmarked against the
event, and this enables the decision makers to decide if the limited funding should be channelled to address this risk, or another risk. A second key component of the risk based design is the design process relies on feedback from the key stakeholders. This closed loop feedback ensures the design process is refined in each iteration. In addition as we will demonstrate later, a risk based design often includes ambiguous or uncertainty. In this instance, active consultations are crucial in achieving a mutually agreeable decision making process. Figure 2: Risk Based Design Approach Adopted for the Case Study 4.0 Risk Based Design for Infrastructure In the current state, the adoption of the risk based design is more frequently seen in the transport infrastructure due to a number of reasons. Firstly the risk based approach is enshrined in the Australian Rail Safety National Law and covered under the approach of So Far As Is Reasonably Practicable, or commonly known as SFAIRP [7]. The recognition of the risk based approach on a legislative level provided the legal backing necessary for this approach. Secondly because of the scale of the projects and due to the finite funding or physical constraints, often the only viable means of achieving a practical design outcome is through
the use of a risk based approach to quantify the effectiveness of a measure needed against an event, e.g. a train fire disabled in mid-journey. If only an absolute approach is adopted in the design analysis, the outcome would show an impractical hundreds of millions of dollars would be needed to reduce or mitigate this extreme event. Thirdly as suggested in Figure 2, a key element in the risk based design is the acceptance criteria of the risk. Although it is possible to show it is not cost effective to mitigate a risk and its consequences, designers and policymakers accept catastrophic consequences due to an event will never be tolerable. To determine this, Transport for NSW for example in Figure 3 has a risk ranking framework developed to categorise the tolerability of a given risk based on the consequences. Type A risk is not tolerable and an alternative will be needed. Type B to D risks will only be tolerable if it can be demonstrated, i.e. via a risk based study and cost benefit analysis that all reasonably practicable mitigations measures have been implemented. Figure 3: Safety Risk Ranking and Guidelines
In this paper, we present a case study to show how the 7 steps risk based approach is adopted in the railway industry as we believe the same approach can be adopted in the wider fire safety engineering design, including in the buildings and industrial sectors. As the intent is not to focus on the technical details, but to illustrate the 7 steps process for the risk based approach from conception to conclusion, we have only provided a summary of the detailed technical design consideration. 5.0 Case Study: Rail Enclosure Structure 5.1 Case Study Introduction The RES or Rail Enclosure Structure is a short tunnel section located in the rail corridor in North Sydney. The RES is a 250 m long 24 m wide short tunnel as shown in Figure 4. As part of the proposed works in the RES, it became necessary to determine if a new fire hydrant system is required in the RES. Figure 4: Rail Enclosure Structure (250 m long tunnel section) Following a desktop review of the current design and standards, we together with the stakeholders determined a risk based design approach is needed to determine if a new fire hydrant system is required in the RES. 5.2 Step 1: Identify All Credible Scenarios The main intention of providing a fire hydrant system in the RES is to allow the fire brigade to perform fire fighting operations for a fire in the RES. Therefore, the first step of the analysis is to identify credible scenarios that will result in a fire event in the RES. We identified the potential scenarios resulting in a fire in RES using the event tree as shown in Figure 5. For this event tree, we have also undertaken a detailed qualitative analysis considering each event but this is not included in this paper.
Figure 5: Identification of the Credible Fire Scenarios 5.3 Step 2: Constructing an Event Tree from Event Start to Consequences Once the credible fire scenario, i.e. the passenger train fire has been established, we have developed an event tree to example the possible consequences chronologically initiating from a fire in the RES and the impact due to the absence of the fire hydrant in the RES. The event tree is based on a passenger train fire caused by equipment failure and arson. Figure 6: Chronological Event Tree for a Passenger Train Fire
5.4 Step 3: Evaluation of the Probability of the Event As shown in Figure 6, we have considered the chronological events associated with the passenger train fires by considering the type of fires (accidental or arson), the severity of the fire, the location where a train is disabled (within or outside of the RES), and the wind condition during a detrainment scenario (presence of wind would spread smoke in the RES). We have established the probability of these events using the relevant statistical data and to account for the uncertainties, we have also included a margin of safety. For example, based on the current fire safety provisions which are expected to improve over time and from the statistic, less than 0.06% of a train fire continues to grow beyond the inception stage. However in our calculations, we increased this to 1% to account for the uncertainty and as a margin of safety. By allowing for a margin of safety in each chronological event from the fire type to the wind condition, this resulted in a compounded safety margin in the design. Using the event tree and the probabilities of the events, we derived the likelihood ranking table in accordance with the Safety Risk Ranking Guidelines in Figure 3. Table 2: Likelihood Ranking for a Passenger Train Fire Event Fire Type Arson Equipment failure Wind Condition Probability (a) Probability of major fire detraining in the RES (b) Not adverse 0.8 0.008 Adverse 0.2 0.008 Not adverse 0.8 0.0009 Adverse 0.2 0.0009 Probability of (a) and (b) 0.0064 (1 in 156 years) 0.0016 (1 in 625 years) 0.00072 (1 in 1,389 years) 0.00018 (1 in 5,500 years) Likelihood Ranking Per Figure 3 L6 L6 L6 L6 5.5 Step 4: Evaluation of the Consequences of the Worst Credible Events Based on the event tree analysis conducted in the previous step, we identified the following worst credible scenarios that would lead to a fire incident within the RES. Once the likelihood has been established, the next step is to consider the consequence (threat to life safety) associated with the scenarios. We have considered the following fire scenarios: 1. Scenario 1 Arson fire with no adverse wind 2. Scenario 2 Arson fire with adverse wind 3. Scenario 3 Equipment failure with no adverse wind 4. Scenario 4 Equipment failure with adverse wind 5. Sensitivity scenario Arson fire with adverse wind and train stopping near portal. We have modelled the fire scenarios and the consequences using advanced computer modelling techniques to establish the tenability conditions in the RES when a passenger train on fire stops within the RES.
The consequence analysis considered a train carriage on fire stopping in the middle of the RES. This resulted in the scenario where passengers have to travel the furthest to reach the portal (open air place of relative safety). However as a sensitivity analysis, an additional scenario assuming wind blowing southward and the train fire carriage at northward near a portal resulting in smoke spread along the entire length of the RES is also considered. The evaluation of the consequences is intended to inform the practitioners and stakeholders the potential severity of the event. Fire Location Plan view of the RES Tunnel Portal Figure 7: A Computational Fluid Dynamics Model of a Train Fire in the RES 5.6 Step 5: Determining the Risk Level with the Probability and the Consequences Combined From the computer modelling in the previous step and from evaluating the tenability criteria for a passenger train fire within the RES, we concluded the consequences for the worst credible scenarios (out of the 5 fire scenarios) are low. See Table 3. Figure 8: Arson and Equipment Failure Fire Risk Ranking At this stage, it is important to note although the consequences of the event are considered low (D), this is not considered satisfactory until it can be demonstrated via a cost benefit analysis the provision of a fire hydrant system has a net negative value. This is a critical element, particularly under the Rail Safety National Law and the SFAIRP (So Far As Is
Reasonably Practicable) to demonstrate all reasonably practicable mitigation measures have been implemented regardless of the level of risks or consequences. 5.7 Step 6: Mitigation Measures plus a Cost Benefit Analysis As part of the cost benefit analysis, it is crucial to establish the consequences cost for providing and omission of the fire hydrant system in the RES. From consultation with stakeholders and from a literature review, we established an indicative consequences cost as shown in Table 3. Active consultations are important throughout the risk based design process, but specifically in instances where there is uncertainty, e.g. in the consequences costs in the consideration. A thorough consultation with stakeholders ensures an informed consensus and a fair outcome is achieved. Note that the disproportionate factor is to account for uncertainties in the calculations of the consequences costs and is a mutually agreed factor with the stakeholders. Table 3: Likelihood Ranking for a Passenger Train Fire Event Impacted Areas No Fire Hydrants With Fire Hydrants Delay in Restoring Service in the RES 6 days 3 days Loss in Ticket Sales $94k per day $94k per day Material and Engineering Costs $100k per day $100k per day Other costs (legal, management and reputational) Factor of 3 Factor of 2 Total Indicative Consequences Costs $3.49m $1.16m Indicative Disproportionate Factor (2 to 5) $6.98m to $17.45m $2.32m to $5.8m Using the consequences costs, and considering the costs for providing a fire hydrant system, we derived the cost benefit analysis for the provision of the fire hydrant with the probability of the fire event factored in. This is summarised in Table 4. Table 4: Cost and Benefit Analysis Variables Total cost for providing the fire hydrant Probability of an arson fire escalating into a major fire within the RES (no adverse wind) Consequences cost for omitting the fire hydrant system Consequences cost for providing the fire hydrant system Design life of the fire hydrant system Consequence cost with the probability of the event factored in Consequences cost over 30 years (design life of a typical fire hydrant) With Fire Hydrant ($2.32m to $5.8m) x 0.0064 = $0.014m to $0.037m per year $0.014m to $0.037m per year over 30 years = $0.42m to $1.11m Description $1.32m 1 in 156 years or 0.0064 $6.98m to $17.45m $2.32m to $5.8m 30 years No Fire Hydrant ($6.98 to $17.45m) x 0.0064 = $0.044m to $0.11m per year $0.044m to $0.11m per year over 30 years = $1.32m to $3.3m
As per the analysis above, the consequences cost for omitting the fire hydrant over 30 years is indicatively between $1.32m to $3.3m. The cost of providing and maintaining the fire hydrant is at $1.32m. Separately as shown in the consequences cost where a fire hydrant is installed, the consequences cost + the installation of the fire hydrants mean the combined costs are between $1.74m to $2.43m. Comparing $1.74m to $2.43m (with fire hydrant) to the consequences cost without a fire hydrant system $1.32m to $3.3m, the difference is between $0.42m to $0.87m where the negative means the cost to install the fire hydrant is not beneficial. This outcome is normal in a risk based design whereby depending on the sensitivity of the consideration, i.e. the disproportionate factor and the cost to install the fire hydrant, the cost benefit analysis could show the provision of a fire hydrant provides a net positive outcome. This scenario demonstrates the importance of consultation as in our experience, a risk based approach will involve a number of uncertainties and the best outcome can often be achieved by active consultations with the key stakeholders. 5.8 Step 7: Consultation with the Stakeholders As we discussed earlier, an important step in the risk based design is to consult with the key stakeholders and to ensure their feedbacks have been accounted for in the design. From our discussions with the various stakeholders including various state agencies, a consensus was reached that although the cost benefit analysis shows the provision of a new fire hydrant system in the RES can be a net negative, the cost involved is negligible when considering the proposed works in this rail corridor. Based on the above, the stakeholders agreed a fire hydrant system should be provided. 5.9 Discussions From the case study, we demonstrated the decision making process for the provision of a new fire hydrant system can, and have been quantitatively assessed and considered. In this instance, although the outcome where a fire hydrant system should be provided may have been foreseeable, this is often not the case when considering other complex engineering designs. For example the design of an open gangway in a train carriage compared to compartmented train carriages. This requires consideration not only from a fire incident perspective, but also from a daily passenger security perspective. Without an approach that can quantify these considerations, it is unlikely a fair and equitable outcome can be reached, and in the unfortunate scenario where an accident occurs, indefensible questions will be raised as to why one decision has been made in favour of another. With a quantitative risk based design approach, practitioners will be able to demonstrate the thought and design process that has been considered to reach an outcome.
6.0 Conclusions In this paper, we discussed the benefit and challenges to the adoption of the risk based design approach in solving engineering challenges. In addition, we provided a 7 steps design process for the risk based design approach that combines both a probability plus a cost benefit analysis. Whilst the technical approach to a risk based design may change depending on the type of projects and environment, fundamentally the process to consider for the risk based design will be in line with the 7 steps we have outlined, from considering the credible scenarios to carrying out a cost benefit analysis plus active consultation with stakeholders. The intent of the case study in this paper is to demonstrate the application of the 7 steps process. Often in a risk based design, there is a misconception this design approach is only on considering the probability of an event, disregarding to the consequences and the cost benefit of the measures to mitigate the event. With a suitably implemented risk based design approach, it is possible to enable decision makers in reaching an equitable outcome which may not always be attainable when using an absolute analysis approach. In engineering and other disciplines with finite resources, a suitably implemented risk based approach can ensure a more objective outcome. This is in our opinion a beneficial approach in assisting decision makers to reach closer to a quantifiably fair and impartial decision. 7.0 References [1] Australian Building Codes Board. The National Construction Code Building Code of Australia. ABCB, April 2016 [2] National Fire Protection Association. NFPA 101: Life Safety Code. NFPA, January 2015. [3] Australian Building Codes Board. International Fire Engineering Guidelines. ABCB, March 2015. [4] B. Meacham. Performance-Based Building Regulatory Systems Principles and Experiences. IRCC, February 2010. [5] Transport for NSW. Principles and Guidelines for Economic Appraisal of Transport Investment and Initiatives. TfNSW, March 2013. [6] National Institute for Health and Care Excellence. Developing NICE guidelines: the manual. NICE, October 2014. [7] New South Wales Government. Rail Safety National Law (NSW) No 82a. [Online] [http://www.legislation.nsw.gov.au/#/view/act/2012/82a], December 2016 [Accessed on March 2017].