Quantifying life safety: Part I: Scenario-based quantification

• Development of a response surface method for probabilistic fire risk analysis.
• Compilation of probabilistic data and stochastic model for input parameters.
• Scenario-based quantification of the failure probability of the life safety design.
• Sensitivity analysis to derive the most significant input parameters and scenarios.

With increasingly complex architecture, new building technologies, etc., compliance with prescriptive code requirements for life safety is often times a sub-optimal solution, overly expensive, and sometimes simply impossible. This leads to the more and more frequent use of performance-based numerical tools in fire protection engineering, especially for the life safety design. Yet the level of safety achieved using those tools remains unknown due to uncertain deterministic input parameters and more or less arbitrarily chosen safety factors. In this paper, an approach is shown to quantify the level of safety for a life safety design using probabilistic risk analysis. The resulting failure probability pf,ipf,i of a life safety design given a specific scenario i yields a metric for the safety level and allows for the objective comparison with other design approaches. Additionally, the approach considers the uncertainty of the input parameters and yields information about the sensitivity of the design to the various input parameters chosen. The methodology will be demonstrated for a multi-purpose community assembly building and the levels of safety are derived based on various scenarios and for different tenability criteria.