NIST Time|NIST Home|About NIST|Contact NIST

HomeAll Years:AuthorKeywordTitle2005-2010:AuthorKeywordTitle

Predicting the Fire Performance of Buildings: Establishing Appropriate Calculation Methods for Regulatory Applications.


pdf icon Predicting the Fire Performance of Buildings: Establishing Appropriate Calculation Methods for Regulatory Applications. (842 K)
Bukowski, R. W.

Interscience Communications Limited. ASIAFLAM '95. International Conference on Fire Science and Engineering, 1st. Proceedings. March 15-16, 1995, Kowloon, Hong Kong, 9-18 pp, 1995.

Keywords:

fire science; fire protection engineering; fire models; evacuation; fire codes; performance evaluation; regulations; safety factors; acceptance criteria; alternative design; code equivalency; measurement uncertainties; performance based codes

Abstract:

A recently organized effort in the International Council for Building Research, Working Commission 14 (CIB W14), on Engineering Evaluation of Building Fire Safety is examining the various quantitative methods being developed to underpin performance-based codes or for determining equivalency with the implied performance of existing prescriptive codes. These methods share many common features and all recognize the range of fire models and calculational methods that the fire safety engineering profession have begun to embrace as their technical foundation. The broad range of assumptions inherent in the available methods as well as the data required to utilize them raises some interesting questions about their appropriateness in applications to code-regulated situations. Many fire-related computations have inherent uncertainty because of lack of understanding of the physics. Thus, one can ask, where a code defines a minimum level of performance, how far must the fire safety engineer go to minimize uncertainty in a calculation intended to verify compliance? The variability of fire means that there are no unique answers against which to define accuracy; and fire experiments involve measurement uncertainties as well as approximations used to reduce the data which often have similar form to the calculations we wish to verify. These methods all focus on managing fire risk, and their successful application depends on assessing the acceptable level of risk implied by the current codes. Some argue that the lack of a public outcry over fire losses is not a tacit acceptance of those losses by society. Thus, how can acceptable levels of risk be determined when regulatory authorities and legislators are uncomfortable with the notion that there is no zero risk so some fatalities are inevitable? This paper explores these questions from the perspective of the fire scientist, the practicing engineer, and the regulatory offical. The fire scientist needs to be explicit about the impact of assumptions on the applicability of the results. The engineer needs to utilize methods and assumptions which are justified by the application and to assess the sensitivity and uncertainty implications. The regulatory officials are insisting on appropriate and properly documented methods. These is a need for models and calculations incorporated into codes of practice, handbooks, or the codes themselves to be reviewed, verified, documented, and approved for use in specific manners and by qualified persons. There are international efforts to define levels of risk acceptable to society in specific occupancies. Until these points are addressed, the transition to performance-based codes cannot be made with confidence.