ISO TC92/Res.244TG N6

CONTENT

1. Background

This document presents proposals for a framework for the long-term future standardisation of fire safety. This framework includes a new set of standards that will support performance-based design (which includes specification of reliability and inspection frequency). Such standards will be essential for a variety of current projects designed either solely by performance-based methods, or more likely, by a blend of performance-based and prescriptive methods, as well as for most pre-engineered construction. There are additional needs for standardisation in areas such as fire safety management and even regulation, but the latter are not within the scope of this document.

ISO TC/92 has agreed that sub-Committees SC1, SC2 and SC3 would be responsible for two types of 'test method' standards:

1. Test Method Standards for categorisation

while SC4 would be responsible for standardisation of the methodologies of fire safety engineering (FSE), i.e., how to 'operate on' the second type of test method result to provide predictions of the 'real world' behaviour. TC/92 was restructured to recognise this "duality" with clear, shared responsibilities established between SC4 and the other Subcommittees.

Fire safety engineering (FSE) is a discipline increasingly being used in support of performance-based national fire safety regulations in many countries throughout the world. The eight parts to ISO/TR 13387 developed by ISO TC92/SC4 outline the fundamental methodologies of FSE. In addition to the purely performance-based regulations, it is becoming very common in many countries for fire safety engineering to supplement prescriptive regulations by being applied to specific design aspects, where reduced costs and improved performance can be achieved.

Many standardised fire test methods give information on the performance of a product or material 'in the test', which may or may not be related to a real fire scenario. These test methods are valuable for product ranking and play an important role in prescriptive regulations, but they are not automatically suitable for supporting performance-based design. Product ranking may be useful for quality assurance purposes or to assess a particular situation, but relative 'success' in a product-ranking test does not ensure relative 'success' in the "real world," where many fire scenarios are possible. Nor does product ranking allow a comparative assessment of performance among alternative (e.g. active versus passive) fire protection strategies. The difficulty is that neither exposure conditions nor performance are adequately quantified to allow extrapolation to be possible from test to different fire situations occurring in the real world.

Because of the need for a new breed of standards to support performance-based design, future fire safety standards are proposed to follow two parallel "tracks" that will reflect the dual needs both of prescriptive regulations and performance-based regulations. These two tracks will contain, respectively:

• STANDARDS FOR CATEGORISATION

Most existing test methods are in this first category. It is agreed that within TC/92, these types of test method standards will continue to be maintained and, where necessary, to be developed. It is recognised that, even if the use of these standards is in prescriptive codes, product data from many of these standards is potentially available for the engineering of fire safety.

• STANDARDS THAT SUPPORT PERFORMANCE-BASED FIRE SAFETY DESIGN

This new standards track will consist of a standards infrastructure for engineering methods and for test methods that integrate all aspects of fire safety design, including disciplines as diverse as the chemistry of combustion or the biological pathways of organic compounds.

• Standards for Engineering Methods

Engineering methods may typically involve:

1. Calculation methods (e.g., algebraic formulas covering limited or specific design aspects)
2. Mathematical models (e.g., computer software for predicting fire behaviour or fire effects)
3. Analysis of reference fire scenarios

The use of reference fire scenarios is an interim technique, until other engineering methods are available to predict real world performance for all scenarios of interest. Reference scenarios are meant to represent a range of "real" fire phenomena for verification of predictive engineering methods or for testing new products, assemblies and fire protection strategies in cases where only real-scale tests will yield reliable design information. Analysis techniques include the linkage or correlation of quantitative results from new or existing test methods to specific performance criteria in the reference fire scenario. Results from reference fire scenarios may also be analysed to determine applicability to real-world environments.

• Standards for Test Methods that Yield Quantitative Data

Although test methods of this type have been agreed within TC/92, the same logic is true for tests within the responsibilities of other technical committees, particularly within TC/21. Standardised testing of active devices such as detectors, extinguishing systems etc. that are constituent parts of any engineered holistic design would need to provide specific information directly to the fire safety designer. New test methods that will meet the requirements of such an engineering analysis are shown schematically in Figure 1. In this figure, the process of providing inputs for engineering methods and the process of the detailed specification of these inputs by engineering methods are both covered by the new standards infrastructure.

There are expectations that with the information available from a quantitative test method, for example the Cone Calorimeter (ISO 5660), coupled with the ability to calculate both 'test' and 'real world' exposure scenarios, it should be possible to assess performance in a full range of practical possibilities (Figure 1). Ultimately, calculations based on test method data input must be able to predict performance for a large number of real scale scenarios.

It is clear that the changes required to achieve the new standards infrastructure will not come immediately. The role of TC92 in this infrastructure is partly that of a catalyst. Without the willing acceptance of the need for these changes by many producers and technical committees, this initiative will fail because required data inputs for engineering methods will not be forthcoming. Time-scales for full implementation are likely to be in the seven to ten year period.

FIGURE 1 SCHEMATIC SHOWING PROPOSED ROLE OF TEST METHODS AND EENGINEERING METHODS IN FIRE SAFETY DESIGN

The new standards infrastructure supporting performance-based design consists of two basic types of fire safety standards: 1) High-level standards that outline the basic engineering requirements for fire safety design and give a comprehensive overview and 2). Working-level standards that implement these requirements for specific situations. There are several reasons why such a standards dichotomy is needed:

1. The information interchange between test method standards and engineering methods shown in Figure 1 indicates that higher level standards are needed to insure that the overall process works smoothly and provides reliable results. For example, standards may be needed to regulate the accuracy of test methods and the format of information transfer to engineering methods. Corresponding standards are needed to prescribe proper engineering methods and the way they are applied, as well as the specification of input information from test methods.
2. It is impossible to anticipate every detailed real-world situation and have a fire safety standard ready, "on the shelf" to deal with fire safety design of all possible products, configurations or processes. Instead, more general standards are needed detailing how fire safety design should be conducted in terms of high level requirements for engineering methods and the test method inputs for such engineering analysis. These more general standards can then be applied to specific situations.
3. ISO TC92 must be able to provide general framework standards that can be adopted by different countries or regional/commercial organisations. These framework-standards would then guide the development of more detailed local standards for different, unanticipated situations. Such higher level standards can also become a part of comprehensive risk or safety management systems for directing organisational behaviour and motivation.
4. Higher-level standards can act as a communication tool for potential customers within and outside of ISO, showing customers (e.g., product committees within ISO) that there is a comprehensive, rational approach to the assessment of fire safety.

The proposed structuring of higher-level fire safety standards, along with some examples, is provided in the following sections.

These standard guides provide the conceptual basis for all aspects of fire safety design and are applicable to all types of structural systems, products and processes.

Standard guides that explain the concept of fire safety design in terms of overall goals and objectives for the use of test methods and engineering methods. Different standards of this type could apply to differing levels of complexity, depending on overall societal impact. Each General Principle standard (together with the appropriate Group A, B and auxiliary standards) would allow fire safety to be proven to regulators or stake holders when there are no applicable, product- specific or performance-specific standards available or acceptable.

The closest example of an existing Level 1/General Principle document would be Part 1 of ISO TR 13387 on Fire Safety Engineering, which gives a framework for the design of fire safety.

5.1.2 Group A Standards

Requirements governing the use of test methods to produce quantitative information for all aspects of fire safety design. Such requirements could cover needs for test documentation, descriptions of sample preparation and conditioning, quantification of repeatability, reproducibility and accuracy and proper calibration intervals and procedures.

The closest example of an existing Level 1/Group A document would be the standard being developed by SC1/WG10 on calibration of heat flux meters (ISO 14943), if this standard were applicable to the calibration of all types of instruments.

Requirements governing the use of engineering methods for all aspects of fire safety design. Such requirements will include, for example, the need for engineering methods to meet verification and validation guidelines, appropriate uses and limitations and the need for adequate documentation. These standards could specify the need for peer review when engineering methods are applied in design projects and the design and proper use of reference fire scenarios as part of a rational engineering analysis. The closest example of an existing Level 1/GroupB document is Part 3 of ISO TR 13387 on verification and validation of models, to the extent that this document is applicable to all types of predictive models. Another example of an existing Level 1/Group B document would be Part 2 of ISO TR 13387 on fire scenarios, to the extent that this document is applicable to the scenario selection process in general and not just to scenarios related to any particular aspect of fire performance.

5.1.4 Auxiliary Standards

1. Terminology Standards (an example is ISO 13943)
2. Maintenance Standards

These are standard guides, consistent with Level 1, that provide the basis for specific aspects of fire safety design but are applicable to all types of structural systems, products and processes.

5.2.1 Group A Standards

Requirements governing the use of test methods to produce quantitative information on specific performance aspects of fire safety for all types of structural systems, products and processes or test method standards that produce such information. Performance aspects may include:

1. Fire initiation, fire growth and the generation of harmful fire products

2. Passive containment of fire and harmful fire products

3. Thresholds for injury to people or damage to materials, products or the environment due to the presence of toxic or harmful fire products

4. Active fire detection, fire protection and fire suppression devices

5. Reliability and Ageing Effects

The closest example of an existing Level 2/GroupA document would be the Calorimetric Bomb (ISO 1716). The Cone Calorimeter (ISO 5660) would be another example if this standard were sufficiently general to apply to all products, not specific types of products.

Requirements governing the use of engineering methods to design for specific performance aspects of fire safety, applicable to all types of structural systems, products and processes or engineering method standards for performing such fire safety designs. Performance aspects may include:

1. Fire Dynamics

2. Response of Structural systems to Fire

3. Safety Behaviour and Movement of People during Fire

4. Damage to Business, the Environment and Heritage due to Fire

The closest example of an existing Level 2/Group B document is ISO 13387, Parts 4-8, to the extent that this document on specific aspects of fire safety design is applicable to all types of structural systems and products. Another example is a document on Verification of Calculation Methods for Structures (SC2/WG7/N27), to the extent that it applies to all types of structural systems and products.

These are working-level standards, consistent with Level 1 and Level 2, used to assess specific performance aspects of fire safety design for specific structural systems, products and processes. As such, these standards can be used for detailed designs and specifications without the need for any additional national/regional/organisational input.

Requirements governing the use of test methods to produce quantitative information on specific performance aspects of fire safety for specific types of structural systems, products and processes or test method standards that produce such information.

Example: A test method that generates information (see 4.3.2) on the heat flux environment during a specified exposure to a sandwich panel.

Example: An engineering method that specifies the type of heat flux data needed from a Level 3/Group A sandwich panel test method and how to use these data to predict real world performance.

A further explanation of the proposed standards hierarchy is provided in Figure 2, where a limited number of examples are illustrated, and in Appendix A. Appendix B contains proposals for development and implementation of the new standards infrastructure to support performance-based design.

FIGURE 2 PROPOSED HIERARCHY OF STANDARDS (WITH EXAMPLES) FOR USE IN THE ENGINEERING OF FIRE SAFETY

Fire safety Engineering-in eight parts, ISO/TR 13387-1 to 8:1999

Standard Guide for Data for Fire Models, ASTM E 1591-94, 1994

Resolution 244 Task Group of ISO TC92

2 April 2002 (Response to 20^th Plenary Meeting of ISO TC92)

APPENDIX A. Further explanations

Conceptual standards should be applicable to all product types, all aspects of fire safety design and all tools for assessing performance-they embody the philosophy of fire engineering application. Their formulation is very general and similar to the simple statements of goals/objectives in performance-based regulation.

The standard requires that any physical, chemical or biological tests on product specimens, to be used in engineering design, will provide those data appropriate to enable 'in-use' product performance to be assessed through exploitation of predictive methods.

It follows that:

• 'product' performance in the test must be provided in quantitative terms for known, controlled and varied exposure conditions

• exposure conditions must be provided in quantitative form and must be representative of in-use fire exposures

• primary processes in the test must be sufficiently well prescribed that they can be modelled theoretically

• performance from the particular conditions of the test must then be translatable by predictive methods to design environments representative of in-use behaviour

Exposure conditions need to be agreed. Some start has been made by the new WG6 of TC92/SC4 the "design fire scenarios" and "design fires" currently being employed in fire safety engineering. A key FSE concept is that it should be possible for all 'products' to be evaluated in a consistent way against the same 'design' fire-albeit at different times during the evolution of the design fire. ISO TC92/SC4 has adopted this concept as the basis for further work.

From the deterministic view point, the Standard requires that predictive tools will model all relevant physical and chemical processes with sufficient accuracy in both standard test methods and in the practical environments intended for use of the test. Data from the test method when used as input to the predictive tool will adequately describe performance in design environments representative of in-use behaviour.

Where "relevant" and "representative" are decided is at Level 2, with a clear concept of where the particular aspect of performance fits into the overall engineered assessment e.g. how can the performance of a sprinkler be related to that of a fire resisting partition using a common fire growth curve.

There is, in addition, a whole class of models not covered by this 'definition'. These are the probabilistic risk models that are used in a very different way. They tend to avoid the technical detail implicit in the deterministic approach but instead utilise data sets from real fire incidents and field surveys. It is clear that standards are required for this approach not only for the risk assessment models but also for the data that they use (e.g. fire statistics). These special requirements for Fire Risk Assessment must be taken into account through the appropriate Level 1, 2 or 3 standards, depending on the generality of the application.

These set the strategy for the particular aspect of performance and relate it to choices of fire scenarios and design fires. What is to be measured-under what conditions-which 'real world' design environment is being targeted?

These standards are generic in the sense that the same requirements will be applied to all product types for assessment of each of the aspects of performance that go to make up the engineering design package.

Examples are:

Standards on how to perform engineering calculations for specific performance aspects of fire safety design; how to decide on a model type, fire scenario, design fire etc. for each such performance aspect (see preceding table).

These are the new working standards that comply with the overall testing strategies established in the broadest sense at Level 1 and in a more focussed sense at Level 2. It is here where the details of test equipment design, measurement devices and tolerances are finalised. Most existing standards are in this category but they lack compliance with the underlying strategy outlined at the higher levels.

Level 3 standards should be developed for modelling methodologies applicable to specific structure, product or process situations. There is a need to standardise either simple, algebraic calculation methods, if available, or the assumptions underlying more advanced, numerical models, when these are applied to various end-use geometric or material configurations.

Algebraic calculation models

Zone models

CFD or Finite Element models

Structural models

APPENDIX B. Proposals for development and Implementation

Implementation of these proposals involve the following steps:

B. Subcommittees ( ISO TC/92 N860) to make explicit the Goals & Objectives of each of its existing standards and to explain into which of the above categories it falls. This work should now be in hand within SC1, SC2 and SC3.

C. Discuss with other ISO technical committees the opportunities presented by and the consequent needs of engineering design methods - the process started with joint meetings on co-ordination etc (Resolution 231 meetings Brussels March 1998, Washington November 1998) and closer TC21 involvement in TC92/SC4 in Paris, April 2000.

D. SC4 have suggested that the following standards be developed:

• Methodology of how to perform the engineering of fire safety
• Modelling methodologies (including rules for selection of design fire scenarios, …)
• Test methods able to provide the necessary data input to calculation models, either for physical, chemical and biological properties (e.g. thermal conductivity versus temperature, specific heat, emissivity, Stress-strain relationships, corrosivity, toxicity…) or more general properties (e.g. RHH, detector activation, behaviour of smoke ventilators, …)
• Test methods for "global" properties, either because it is not yet possible to model them (e.g., integrity or leakage rate for separating elements, behaviour of doors or ducts…) or for obtaining data necessary to validate calculations models.

• Level 1-ISO TC/92 WG11 (TPMG). This work should be conducted with the full involvement of representatives of other ISO/IEC committees with an interest in fire safety. This is probably best achieved by the establishment of a new Task Group-the leader of which needs to be appointed at the 20^th Plenary.
• Level 2-ISO TC/92 Sub Committees 1-4
• Level 3-product committees such as ISO/TC21, ISO/TC61 etc in addition to the TC92 Subcommittees.

The work would need to develop as a cascade starting at level 1-a plan to be agreed at the 20^th Plenary and to be in full operation by the 21^st Plenary in September 2003.

Level 2 could probably start soon after Level 1 development starts with first drafts available for the 21^st Plenary. This is where the overall strategy is made specific and where most work will be needed.

Level 3 would be the most difficult to complete since it involves the 'buy-in' and agreement of the product committees. This work should probably not commence until after the 22^nd Plenary, although preparatory presentations could be made to prepare the ground.

ANNEX - some contributors to this debate

Contributions were provided on a personal basis