Decision Support Framework for Sustainable and Fire Resilient Buildings (SAFR-B)

Increasingly, buildings are becoming complex ‘systems of systems’, with many materials and attributes combining to create a whole that aims to meet a variety of design objectives including but not limited to functionality, aesthetic appeal, sustainability, safety and security. While this has perhaps been the case in complex commercial or multifunctional buildings previously, this complexity is being seen even in the residential sector today. Such buildings are typically designed by professionals seeking to produce stunning, environmentally friendly, healthy, safe, cost efficient and operationally efficient artefacts. They are engineered by experts from diverse disciplines, using innovative materials and technologies, that do not necessarily interact, but focus on their piece of the whole design picture. They are constructed within regulatory boundaries which largely align with the major systems or components of a building (e.g., structure, mechanical systems), albeit sometimes missing important interactions between systems. Recent publications discuss these and other issues, without operationalizing how to solve the question of multi-attribute optimization [1, 2].

While this can result in rather spectacular buildings, with state-of-the-art technologies as part of the building (e.g., building-integrated photovoltaics) and within the building (e.g., automated systems for improved indoor environments or improved user comfort), there can be unintended consequences from design choices that may not manifest until well after construction. In recent years, there have been a series of rather significant fire losses associated in one way or another with choices made to meet societal objectives to be more environmentally sustainable and minimize the potential for climate change. These include numerous high-rise exterior façade fires around the world, notably the Grenfell Tower fire in London [3‐8], the Dietz & Watson cold storage warehouse in Delanco [9], and a spate of fires in buildings under construction using lightweight timber framing [10‐13]. Additional incidents have happened relating non-fire related failures, such as the Champlain Towers collapse in 2021 [14, 15], but the impact of fire as the failure mode is the focus of this work.

The challenge in addressing fire safety issues associated with ‘green’ buildings and attributes has been studied previously by Meacham and McNamee [16], and can be readily understood when one considers the number of attributes and scenarios that exist. The focus of this previous research was on understanding the fire safety implications of choices made to meet sustainability objectives. This is important, but is only one aspect of the optimization needed to properly balance both sustainability and fire safety. It is also necessary to consider the sustainability implications of choices made to meet fire safety objectives. The work presented in this paper represents initial scoping and proof of concept of such research and development, based on research published in more detail in a scientific report [17].

This work presents a holistic, integrated approach to sustainable and fire resilient buildings, and indeed the built environment as a whole. The development of such an approach has been presented conceptually previously [1] as the Sustainable and Fire Resilient Built Environment (SAFR-BE) framework. This paper will extend beyond that conceptual framework to operationalize the methodology. In the following sections, the concept will be introduced together with a discussion of what is meant by sustainability and fire resilience provides as a backdrop to the evaluation of sustainability and fire resiliency model needs. Finally, the multi-criteria risk assessment framework will be developed with implementation of the SAFR-Buildings (SAFR-B) concept into a semi-quantitative decision support methodology and application to a small multi-story residential building.

Fire impacts to the environment come from both natural and human sources. Managing both often requires different strategies and can be pursued without consideration of the interface between the natural and technological worlds. However, as society has looked to minimize its impact on the natural world by implementing sustainability strategies, and more recently has recognized the need to make human settlements resilient as well as sustainable, the need to create sustainable and resilient human development has emerged. When one further considers the interrelationships between carbon emissions, climate change and climate impacts, the need to support both sustainable and resilient building solutions becomes pervasive.

Fire is a particular hazard of concern and focus due to its inherent destructive and dangerous nature to the built environment itself and those residing in it. As a natural hazard, fire is increasing in frequency and intensity due to climate change caused or exacerbated extreme weather leading to, e.g., drought conditions and high temperatures. Further, as a technological hazard, fire can be inadvertently increased by implementation of sustainability measures, e.g. energy sustainability measures creating potential ignition hazards and/or thermal insulation creating additional fuel load. In order to comprehensively address and mitigate the increasing fire risk it is necessary to adopt the concepts of a Sustainable and Fire Resilient Built Environment (SAFR-BE), see Fig. 1 [1]. In this context, the built environment includes buildings (structures, facilities), infrastructure and communities.

Full size image

Fundamentally, application of the framework requires the identification of system boundaries which determine whether the model pertains to SAFR-Buildings (SAFR-B), SAFR-Infrastructure (SAFR-I) or SAFR-Communities (SAFR-C). Depending on the system boundaries, different strategies will be needed to achieve the design objectives of sustainability and fire resilience.

The latest reports from the Intergovernmental Panel on Climate Change (IPCC) state that observed increases in greenhouse gases since 1750 are undoubtedly due to human activity [18‐22]. The observed warming of global surface temperature over the past 200 years is unprecedented when compared to temperature estimates over more than 2000 years. It is urgent to reduce emissions of greenhouse gases, which has resulted in numerous initiatives in recent decades (e.g., [20‐26]).

The construction sector has great potential to positively influence sustainable development by reducing energy and/or material use. While the UN efforts to foster sustainability have identified that there are multiple dimensions of sustainability [27], early efforts to improve the sustainability of the built environment have typically focused on the environmental dimension [28‐31]. In the 1990’s, efforts to improve the environmental impact of buildings led to the emergence of green building certification schemes. In the development of a SAFR-B methodology, the sustainability objectives have been developed by investigating the two dominant green building certification schemes, LEED and BREEAM [32, 33], and the World Building Design Guide (WBDG) [34]. While WBDG is not a certification scheme, the design objectives within the WBDG relate well to the concept of sustainability objectives. Figure 2 provides a comparison between the relative importance of the categories identified within the LEED and BREEAM systems. The WBDG sustainability objectives are not included in this comparison as they do not have scores.

Full size image

It can be seen in the comparison between LEED and BREEAM, that there is considerable alignment in prioritization of the sustainability categories even if the two schemes use slightly different nomenclature and number of points for each category. In both cases, energy is most highly prioritized followed by indoor air quality or health and well-being. In terms of the WBDG objectives, sustainable, accessible, functional all relate well to the categories in LEED and BREEAM, albeit without a points scheme.

Resilience is broadly defined as the ability to prepare and plan for, absorb, recover from, and more successfully adapt to adverse events. With respect to fire resilience of buildings, there is a close relationship to measures that have been used for decades to layer protection into a building in ways that help fires from igniting, controlling the spread and impact of fire if it occurs, and facilitating rapid suppression. These attributes work both to keep occupants safe and to limit damage, which facilitates recovery. Fire safety can be seen as an important component of fire resilience given the aspect of preparing for and assisting recovery from the event of a fire. A well-known structure that illustrates various fire safety strategies that contribute to fire resilience is the National Fire Protection Association’s (NFPA) Fire Safety Concepts Tree (FSCT) [35]. The FSCT reflects the elements that must be considered in meeting the objective of fire safety, and the interrelationships among those elements, which then enables a building, group of buildings, or community to be analyzed or designed by progressively moving through the various concepts in a logical manner [36]. The top levels of the fire safety concept tree decision structure are illustrated in Fig. 3.

Full size image

There are arguably four fundamental objectives of fire safety: protection of life, property, business (mission), and the environment. In addition, the domain of fire safety is broad—from small flames or overheating materials to massive wildland fires. The FSCT is applicable to all objectives and domains. Given the close relationship between fire resilience and fire safety, the FSCT has been used as the basis for developing the fire resiliency side of the SAFR-BE framework.

Risk means different things to different people. It can be challenging to characterize due to the breadth of perceptions, conceptualizations, and definitions that exist [37]. The acceptability of risk has been described as a decision problem [38]. Following that concept, for the purpose of this work, the balancing of risk associated with sustainability and fire resilience in buildings is considered a decision problem. Data, tools and methods for the characterization and analysis of risk exist across an equally broad spectrum (see e.g. [39‐47]). The choice of which underlying method to apply has depended largely on the researcher’s preference and available data. The same is true for quantification and treatment of uncertainty in the risk analysis and characterization (e.g., Paté-Cornell [48]).

Fire risk has many facets, such as risk to people, property, business and environment, from the several fire effects, such as high temperature, thermal radiation, and products of combustion which present many hazards. Various approaches and considerations for fire risk analysis and characterization are widely discussed in the literature and not detailed here (see e.g., [37, 38, 42‐46, 48‐64]). There are many scales that one can consider for fire risk assessment, from product, to building to geographic area (wildland and wildland-urban interface).

There are risks to the environment posed by numerous sources, and the scale is generally larger, for example starting at a local habitat level [65], increasing to an ecosystem level [66], and ultimately impacting the entire world [18, 67, 68]. As a broad generalization, hazard events occur locally, but in many different forms, and their effects can extend beyond the local. This includes ecological damage due to raw material extraction, climate impacts due to materials transportation, fabrication and use, and ecological and climate impacts associated with waste, aspects of which are dealt with in life-cycle assessments [69] or using life-cycle thinking [70, 71].

Because fire and environmental impacts as used in this work are both risk problems, assessing and managing risks associated with unwanted fire impacts and risks associated with unwanted environmental impacts can use the same general processes and approaches [72, 73]. However, the contexts, effects, scale and impacts vary considerably. This makes a one-to-one comparison extremely difficult and perhaps unhelpful (depending on the scale of comparison). Carbon contribution to the atmosphere from intentional human activities (e.g., raw material extraction, processing, manufacturing, combustion of fossil fuels, etc.) dwarf human caused carbon release to the atmosphere by accidental fires in buildings [74]. At the building scale, though, much can be done to decrease the risk of fire and the risk of impact on the environment simultaneously. Measures to reduce both sets of risk are often incorporated into building regulation and building design practice, and can be externally influenced as well (e.g., insurance for fire and ‘green’ building certifications for environment, such as LEED, BREEAM and others).

Unfortunately, as the risks are treated differently, mitigation in one area can result in unintended consequences for the other [16, 75‐81]. It is this confluence of fire risk management and environmental risk management objectives for buildings, in particular where they create the potential for competing objectives, which is a driver of this research effort [2, 39, 40, 72, 73, 82‐84]. More specifically, it is the introduction of technological means for risk mitigation of environmental and fire risks in buildings, how the risks may interact, and how one might understand the preferences and implications of choices between risk mitigation options that is of interest.

Broadly, the risks to the environment can be framed as contribution of building technologies to reducing the impact on the environment. This can manifest as circular building materials [85], energy systems [86], and more (e.g., Anand et al. [87]). Fire risks are framed largely as those attributes of a building that impact the ability of occupants to remain safe in the case of fire. This includes means to prevent fire occurrence or manage the fire and manage the exposed (i.e. primarily the occupants) should fire occur.

Even with such bounding conditions, the risk assessment challenges are significant due to the number of factors involved, difficulty in understanding risk tolerability limits in such a framing, and general lack of data that are specific and/or useful to this framing. This argues for keeping the risk assessment component relatively simple by using a risk indexing decision support approach. In this research, the decision support approach should help to identify the relative importance of both the fire resilience and sustainability attributes to risk reduction, along with the relative weights of building attributes to reducing the risk.

Important to this research is that at present, data remain lacking on fire risks of sustainable building attributes—both likelihood of fire events and consequences thereof—and on perceptions of fire risk associated with sustainable building attributes—that is, the importance of fire risk in comparison with benefits of sustainable construction. As such, a first attempt to address both aspects led to a decision that a qualitative or semi-quantitative approach to risk assessment, coupled with a decision support method to gauge importance, is most appropriate. As a step beyond the qualitative relative risk matrix approach outlined in previous research [16], a combination of the fire safety concepts tree [35], green building certification schemes [32‐34]; and a risk-index approach has been chosen as the basis for the development of a risk assessment and decision support framework. While still qualitative, the fire safety concepts tree provides a set of building fire safety strategies for consideration, and a risk index approach takes a step towards risk quantification.

Risk indexing has been selected because it is simple to understand and easy to work with amongst a group of experts. Koutsomarkos et al. [46] describe fire risk indexing (FRI) methods as heuristic models of fire safety. They define heuristics as procedures that, in the absence of a formal underlying physical theory, provide a practical approach to solving problems, and are typically defined as efficient rules or procedures for converting complex problems into simpler ones; and suggest that such methods reflect a problem solving approach that employs a practical method that is not guaranteed to be optimal or perfect, but is instead considered (by the method’s designers) sufficient for reaching an immediate goal. In keeping with Koutsomarkos et al. [46], a risk index approach is a multi-attribute evaluation used to develop risk assessments where the results are aggregated into a single number. The process of creating a fire risk index commonly includes a procedure of scoring, and mitigating, fire safety attributes—with the result being a rapid and relatively simple fire safety evaluation. The scoring process is typically undertaken by allocation of points to each attribute, considered they are all differently important to the overall risk score. There are numerous examples of fire risk indexing in the literature (e.g., [39, 40, 42, 45, 46]). The same approach is taken for assessing the sustainability level for a particular building, i.e., to allocate points to attributes, considering the different importance for each attribute, that focus on meeting sustainability objectives.

Making decisions on building attributes, which aim to balance sustainability and fire resilience objectives, is effectively an acceptable risk decision problem: how does one balance sustainability and fire resilience without compromising either. This is a complex decision problem with many components, including differing sustainability and fire risks, the relative importance of each to any particular set of choices for building design, and how the relative importance should be weighed or balanced. Use of a decision support tool in making such decisions provides a helpful structure. Recent work explored appropriate decision support tools for this type of problem [17]. It was determined that a multi-criteria decision making (MCDM), or multi-criteria optimization approach, would work best. Research included consideration of a recent paper that summarized numerous multi-criteria optimization methods which have been developed and published through a thorough review of almost 300 peer reviewed publications [88]. The cited research includes a tool to help one select an appropriate MCDM for a particular application, based on aims of the decision problem. After consideration of various MCDMs, the Analytical Hierarchy Process (AHP) was selected. This choice was confirmed by using the MCDA Methods Selection Software of Cinelli et al. [89], which is available on http://​mcdamss.​com.

The Analytical Hierarchy Process (AHP) is an organized framework that helps one think about and address complex interactions and interdependencies between decision factors in a structured way [90]. One of its strengths is the integration of both deductive (focus on parts) and systems (focus on the whole) approaches to solving problems, into a common framework. It does this by breaking down complex, unstructured problems into component variables, arranging the variables into a hierarchical order, assigning numerical values to subjective judgments of the relative importance of each variable, and synthesizing the judgments to determine which variables have the highest priority and should be addressed. It is applicable to single or multiple decision-makers. It provides a means to integrate facts with subjective judgments, incorporate judgments of several people, and resolve conflicts.

The structure of the AHP has at least three levels [90, 91]: the focus (or goal), which is the main objective of the decision problem; criteria levels (criteria and sub-criteria), which describe key components of the decision problem; and alternatives (decision options). Priorities are established so that the criteria or sub-criteria in each level are comparable to each other with respect to the next highest level. The priorities are then weighted, again with respect to the criteria in the next higher level. Finally, the weights are evaluated to determine the overall priority of alternatives to support a decision. Some recent reviews indicate that AHP is by far the most popular approach corresponding to almost 40% of all published methods [88, 92‐94].

The process for setting priorities involves pairwise comparison of elements in a particular level against a given criterion on the next level. The comparisons can be expressed in one of three ways: as the degree of importance of one criterion over another, as the degree of likelihood of occurrence of one criterion over another, or as the degree of the decision-maker’s preference for one criterion over another. The scale for the pairwise comparisons is 1–9, which seems to reflect the degree to which people can discriminate the intensity of relationships between elements [90, 91]. It should be noted that this is a ratio scale, not an interval scale, with the levels expressed as 1: equal importance (or preference, or likelihood), 3: weak importance of one over another, 5: essential or strong importance, 7: very strong or demonstrated importance, 9: absolute importance. For decisions with multiple criteria and/or alternatives, the criteria or alternatives should be clustered into sub-groups that have qualities with no greater preference intervals than can be expressed by the one-to-nine ratio scale.

The mathematical foundation for the AHP is matrix algebra, wherein each of the criteria is included in at least one comparison to another criterion. Details can be found in Saaty [95] with standard application to buildings describing in ASTM standard E1765 [96]. In brief, this means that for \(n\) criteria, the minimum number of pairwise comparisons required is \(n-1\), and the maximum possible number of comparisons is \(n\left( {n - 1} \right)/2\). As a general rule, the more paired comparisons made among the criteria, the better the final decision made should be.

Once hierarchical levels have been identified, components on each level developed, values have been assigned to their prioritization to develop ranking of items, there are three main judgements which are needed in development of an index:

An index is typically structured in a hierarchy built up from a number of different levels; e.g. Level 1: Policy, Level 2: Objectives, Level 3: Strategies, Level 4: Attributes, Level 5: Sub-attributes, Level 6: Survey items. Levels can be eliminated, although the top four levels are generally used. The fewer the levels the simpler the model. This basic structure has been incorporated into the SAFR-B application.

The intention of SAFR-B is that it should provide a framework for using a risk index approach for grading aspects relevant for the building which are important for environmental sustainability and fire resilience. Since different building uses may have different importance levels associated with sustainability and fire resiliency objectives, the initial development of this framework has focused on one building typology. The initial typology that has been selected is an apartment building, not least due to the fact that the majority of fire related fatalities occur in residential occupancies [97].

The SAFR-B framework makes use of the AHP structure for scoring and weighting attributes that contribute to sustainability and to fire resilience, and in concert, to sustainable and fire resilient design. It has been developed through a multistep Delphi process type of engagement of expert input.

The first stage was to develop the hierarchy of the framework in support of a policy of creating a sustainable and fire resilient building by defining the top-level policy, defining target sustainability and fire resiliency objectives, and identifying a finite set of strategies with associated attributes and sub-attributes (survey items). The choice of sub-attribute or survey item depends on the granularity associated with a specific attribute. The experts and non-experts were approached first to provide feedback on the structure as such. Changes to the overall structure (e.g. names and descriptions of the strategies, attributes and sub-attributes) were made based on this external input. The final SAFR-B hierarchy is shown in Fig. 4.

Full size image

There are two branches (objectives) of the hierarchical SAFR-B assessment tree: sustainability and fire resilience. These objectives are supported by seven strategies which have been defined to be associated with sustainability or fire resilience. The wording of these strategies has taken its starting point in sustainability documentation [98] and fire safety documentation [35] in an effort to promote recognizability to both communities.

At the base of the tree, the nine attributes are aspects that are important for the sustainability and fire resilience strategies. These attributes are differently important for each strategy used to reach the ultimate objectives of environmental sustainability and fire resilience. Each attribute is associated with a number of survey items, which are aspects that are considered to be important for the attribute and which can be graded during a survey of a particular building. While it is possible to have more attributes, the AHP works best with a limit of nine. To accommodate this, more complex attributes can be divided in sub-attributes to better capture details.

In order for the framework to be operational, it is necessary to determine the importance of each attribute and the relative grade of each attribute from both the sustainability and fire resilience perspectives. The importance value together with the score for each attribute will be used to derive the sustainability index and the fire resilience index, i.e., how well the proposed building design meets these two objectives. This relative score (and to a certain degree importance) of each attribute of the proposed framework is determined by expert users through a survey of the building.

In a second stage, the experts and non-experts were asked to perform a pairwise evaluation of attributes using the AHP. The experts provided input to the AHP using a Delphi procedure [99, 100], in which the experts conduct their evaluation independently, and each expert has an opportunity to review an initial assessment in a final evaluation step, knowing the combined assessment from all experts. A schematic example of the spreadsheet in which experts provided their initial pairwise comparisons is shown in Table 1.

Numbers were entered into the fields based on relative importance of the attributes in achieving the strategy according the methodology developed by Saaty [90]. Here, a ratio of 1:1 represents that the attributes are equally important, 1:3 that the second attribute is moderately more important, 1:5 that the second attribute is strongly more important etc., up to a maximum relative value of 1:9. A ratio of e.g., 1:5 designates that the second attribute is strongly more important that the first. The relations can, of course, also be assessed the opposite, e.g., 3:1, indicating that the second attribute is moderately more important than the first. Note that all the numbers in the red and green cells were entered by the experts manually while the grey cells were transposed from the expert input automatically.

The SAFR-B index (in this case broken down in the objectives for sustainability and fire resilience) is defined as a sum of all (n) attributes, expressed as the product of a scoring and an importance (weight), as shown below. Using this method, it is also possible to calculate an overall SAFR-B performance index by combining the two individual indices. This index will then give an indication of the Policy performance.

The SAFR-B framework has been applied to a mid-rise apartment building to illustrate how it might be used to inform decision making when weighing sustainability and fire resilience attributes as part of initial building design. Calculations were performed in Excel using derived values. The framework allows users to select building options for each survey item, and then to compare different building choices based on fire resilience and environmental sustainability. The example is based on a typical multi-story multi-family residential building in Sweden. Building 1 is an actual building (see Fig. 6) while Building 2 and 3 are variations of this construction using different attribute choices, i.e., they are both fictious buildings.

Full size image

The building selected as the basis of comparison in this case study (building 1) is a 6-story building in central Malmö, Sweden, built in 1934 with concrete construction. Building 1 is connected to other buildings in the block, although the model calculations have been for this part of the full city block only. The apartments have gypsum walls and ceiling surfaces, with wooden floor surfaces. The façade is covered in plaster while the material used for insulation is unknown and non-fire rated. The building lacks an active fire protection system, with only single detectors installed in the stairwell, attic and cellar. The building management is mostly done by residents, and maintenance routines and information about fire safety and sustainability are weak or nonexistent. The building uses district heating for centralized temperature control during the winter and no central cooling is provided.

Building 2 is a modern 8-storey construction built with precast concrete planar elements. The building is single standing with 13 m to the closest neighbor. The building has wooden inner linings and a façade of wood insulated with fire rated rigid polyurethane foam. There is no attic in the building, and the stairwells are fire rated with door closing system and mechanical ventilation. The building has both active fire extinguishing systems and detectors connected to a central alarm system. The residents are provided with general information about fire safety and sustainability, and the building management test and maintain building systems frequently. The heating and cooling of the building is done by a heat pump integrated with a mechanical ventilation heat recovery system. Building 3 is identical to Building 2, with the addition of solar panels on the roof and a lithium-ion battery energy storage system.

Figure 7 illustrates the differences between the three buildings using the SAFR-B framework. The building method from 1934 is the least favorable option in terms of fire resilience and sustainability. Installing electricity storage and generation proved beneficial for sustainability but detrimental for fire safety, due to decreasing the energy use and increasing the risk of fire ignition.

Full size image

When considering the relative importance of fire resilience and environmental sustainability, the SAFR-B framework indicates that installing electricity generation and storage is the more favorable option, see Fig. 8. The figure indicates a clear difference in the sustainability index, while the fire resilience index is almost the same for the two buildings. The reduction in fire resilience index can be explained by the lower value for "Prevent fire ignition" for building 3 (higher risk).

Full size image

Finally, the objectives can be numerically combined to give an indication of how well each alternative achieves the policy of a sustainable and fire resilient building, see Fig. 9. In this application, the difference between buildings 2 and 3 identified in Fig. 8 is marginal. This is an indication that not only the “policy index” should be assessed but also indices on objective and strategic levels. This is one of the drawbacks with the index method approach that differences could be diminished as sub-indices are combined. The advantage is that a global assessment value is achieved.

Full size image

The intention of the SAFR-B framework is to demonstrate a technique to compare aspects of a building as input to design choices relative to their environmental sustainability and fire resilience. Expert input was used to establish the relevance of the strategies, attributes/sub-attributes and survey items of the model. Once the structure for this first version of the model was established, the Delphi method was used to obtain expert input to the development of scores and weights in the model. When developing the framework, there was a concern that there would be a bias towards resiliency as the authors all have their background within that discipline. Some input from experts indicate that the end result may instead have been biased towards sustainability. More expert input is needed to improve this balance. Related to this comment, the question of the importance of the building structure has been raised. We know from a sustainability perspective that the structure is the part of the building that has the greatest potential to impact on sustainability. The importance of this feature is rather lost when it appears first at the attribute level. Perhaps it should be raised in importance and directly related to a strategy instead in future versions of the model.

As part of this work, an effort has been made to consider the fact that fire safety is part of the regulated built environment while building sustainability has largely been developed through extra-regulatory means. Experts have expressed that there is a need to better relate the pairwise comparisons, used in the AHP method, to what is required from a purely regulatory point of view and what is optional to some degree during building design. As the regulatory setting will vary between countries and contexts it is important to develop guidance concerning how a user can incorporate evaluation of the framework output against the regulatory backdrop provided in their country of application.

Particular effort has been given to defining various strategies and attributes. Nonetheless, much of the expert input related to the difficulty in assessing relative importance of attributes and definition or identification of sub-attributes without precise definitions, e.g. below is a selection of expert queries collated during the evaluation process:

While additional detail will need to be developed over time for all strategies and attributes and some will be very country or project specific, one strength of the method (provided large numbers of experts provide input) is that grading will ultimately be quite robust. It was noted, however, that even experts who are familiar with the approach need to have an understanding of how the model has been designed to ensure that their input is in line with the intentions of the framework. This can only be achieved through improved clarity in component definition.

It should also be mentioned that using a risk index approach implicitly will lead to a simplified description of the object to assess which is indicated in the case study. It will never be possible to include all details that may be present in a specific case. The overall purpose is to provide a general structure for indicating the consequences of design selections. There will, therefore, always be a discussion between feasibility and details and SAFR-BE is not a substitute for a proper and detailed design process.

A first application of the SAFR-B framework has been made as part of this study, and the methodology developed has been applied to a case study of a six-floor apartment building. Fire safety engineering (FSE) and performance-based design (PBD) are at the heart of resiliency in the built environment. The development of engineering methods to weigh sustainability and fire resiliency in building design has the potential to impact on design choices as part of the PBD process. The example given in the case study compares three building alternatives. In all three cases, the buildings have the same function, but make different design choices to achieve that function. The SAFR-B framework provides a methodology to weigh how sustainability and fire safety choices can be optimized and evaluated. In the example chosen, the traditional building method from the 1930’s represented the poorest design solution of the three while the other two alternatives had similar overall SAFR-B performance. The detailed investigation illustrated that building 2 represented the best alternative from a fire safety point of view but building 3 represented the best compromise between fire resiliency and sustainability. The framework is not in its final form but even in this first application, the output provides interesting input to design choices.

Expert input has identified the need for more careful definition of sustainability and fire resilience. As part of this improved definition, it will be necessary to define boundaries for comparisons between survey items associated with the various attributes. A large number of experts were included in the development of the model and the scores, as well as the weighting associated with the AHP model. While the project team is very appreciative of the time and input from all experts, more input is needed for further development of the model moving forward.

While life-cycle thinking lies behind the development of the model, life-cycle analysis is not explicitly used in the framework and the comparisons are somewhat subjective. The design phase develops the basis for all aspects of the building life-cycle, e.g. design choices will affect recyclability of products and material in the building, and the explicit inclusion of life-cycle assessment (LCA) as input to grading between choices will improve the application of the framework. As it becomes more mainstream in the future to include LCA calculations as part of the design process, such information could also be used to develop robust scoring of attributes or weighting between attributes.

The framework aims to improve both sustainability and fire resiliency, not one at the cost of the other. To improve the ability of the model to reflect this, it will be necessary to broaden it to include the impact of fire resiliency choices on a reduction in the number and size of fires. Indeed, when working on development of the framework, one expert input highlighted the difficulty of weighing two desirable outcomes against each other, i.e., sustainability and fire resiliency. Improved fire resiliency can, directly or indirectly, lead to reduced sustainability due to material use and maintenance needs; but, if we do not consider the fact that the building may burn down if fire safety needs are not met, we risk missing an important impact on sustainability. The issue is complex, but this does not mean that future versions of the model should not at least try to provide some design input concerning such potential dilemmas. This inclusion requires models to be developed of the statistical number of buildings associated with a fire as a function of the choices made. This has been outside of the scope of the present study.