Fire Safety Science Digital Archive

The rate of fire growth in a real scale fire test is estimated using an inverse modelling approach of computational fire dynamics (FDS v5). Measurements from the fire test are assimilated into the fire model and based on this the parameters of a decoupled fire growth model are estimated. A forecast of the fire development is then made with the estimated parameters. It is shown that a simplified fire growth model can give a robust representation of the underlying physics and that the necessary parameters can be estimated at an acceptable computational cost. The forecasts are shown to accurately predict the fire development. The results are based on a simplified single parameter fire growth model of a well characterized scenario, but the methodology allows for an extension to a more complicated model that would require less previous characterization of the fire scenario.