A Semi-Markov Fire Growth Model

Abstract

This paper aims to describe an alternative fire growth model that allows prediction of fire development, including time to flashover. The unpredictable nature of real fire development is incorporated into the model through use of probability distributions which can be defined using appropriate fire test data, when available. By assuming that the fire goes through five different stages starting from ignition and progressing to flashover, the total time to flashover may also be estimated. The model considers potential variability in the times at which the fire will undergo transitions between the various stages of development, using a state transition method called the semi-Markov process model. Different fire data may be incorporated into the model by defining appropriate statistical distributions for the transition descriptors, making the model flexible enough for use in a variety of applications important to both product design engineers and fire safety regulators. © 2012 Springer Science+Business Media New York.