A New Approach for Modeling the Occupant Response To a Fire in a Building

Abstract

The Risk Assessment Method (RAM) has been adopted as the basic approach for evaluat ing the fire risk and cost associated with fires in apartment buildings, hotels and motels in Australia. It uses a systematic approach to consider a large number of scenarios which are related to fire growth and spread, smoke spread and management, detection, barrier performance, occupant response and evacuation, and fire brigade fire fighting and rescue. For each scenario, a number of submodels are required to be run in order to evaluate the performance of the fire safety system in the building. To limit the overall computational time to a reasonable period of time (hours rather than days on a PC), the computational time for each submodel must be short. The paper proposes a new approach which is termed the Expected Multi-Response Approach for the Human Response and Evacuation submodel. Using this new approach, the computational time is significantly reduced compared with using the conventional Monte Carlo Method.