The Computer Model "BGRAF": A Cognitive Approach to Emergency Egress Simulation

During the past decade, fire safety researchers have come to the understanding that human factors in fires play an important role in controlling the spread of fire; and in decreasing the number of fire casualties in buildings. With the current developments in computer technology, computer modeling of human behavior in fires emerged as an effective method of research. Such computer modeling techniques offered the advantage of being able to experiment with hypothetical fires in buildings without Note endangering human life. Consequently, a study to develop a computer model that will simulate the emergency egress behavior of people in fires was undertaken. Changes in the information processing capacity of the individual as a result of time pressure and stress was considered as part of the emergency egress decision process. Theories from environmental psychology identified a range of cognitive factors, such as visual access in buildings, architectural differentiation, signage and plan configuration that affect way finding and route selection in buildings. These factors needed to be incorporated into emergency egress models. The model was based on the integrated building data base of the CAD system developed at the University of Michigan, Architecture and Planning Lab., which provided a comprehensive building definition, and allowed both graphic and tabular output. Two actual fire incidences were simulated as part of the validation study. These studies have stressed the importance of the cognitive aspects of the physical environment as a factor in emergency egress. A goal structure that represented the total decision process during fires was incorporated into the model. This structure allowed the inputting and testing of a variety of goal structures by using actions as model blocks. The objectives of the model developed in this study can best be summarized as to study and eventually to predict the route selection and exiting behavior in fires, with the purpose of using such information in making building design and code development decisions, and in suggesting action sequences that will best support the safety of the occupants of a building under different emergency conditions.