Parametric analysis of input data on the CFD fire simulation

During the last decades the CFD computer modelling of fire has increased significantly. The CFD fire simulator has been purposely developed and optimized for fire protection; nowadays they have a high reliability and ample perspectives of further development. The main drawbacks of CFD fire simulators are mainly due to the complexity of its setting. They are not user friendly and require a good competence on fires and fair competence in computing. To be able to estimate how the simulation outcome is influenced by model assumptions and simplifications, the user has to evaluate the dominating physical processes and involved empirical parameter that has an essential influence on the specific fire scenario. In this paper the Thermo-gravimetric (TGA) and Cone-calorimetric tests will be used to show in which ranges the input parameters such as reference temperature (i.e. the temperature of the released mass rate peak), thermal conductivity, pyrolysis range, specific heat, in a fire simulation with the "Fire Dynamics Simulator", will vary based on the measurement errors, measurement uncertainty, or misinterpretation of the user. Fire simulation validated on experimental tests are used to evaluate the impact of those variant input data on the simulation. A parametric analysis has been carried out and the results are evaluated and discussed.