Validation of flacs code for risk analysis of hydrocarbon pool fires

Fires have been object of study over the last decades due to their destructive power. Fire’s hazardous nature and its ability to inflict damage to property, the environment and people, has produced a need to understand how it works in every aspect. Currently, the main focus is to estimate the fire characteristics and main effects, in order to accurately design emergency plans and prevention measures. Due to the needs previously stated, fires have been studied and analyzed mainly from an experimental point of view. However, experimental data is arduous and extremely expensive to obtain due to the amount of resources needed. Additionally, small-scale models, which are generally easier to be undertaken, cannot be extrapolated to full-scale models. Considering this, semi-empirical methods were developed, but can only be applied to simple scenarios and they cannot fully model them. To achieve complete models of fires, CFD (Computational Fluid Dynamics) modeling has been recently used as a way to achieve a cheaper and easier method to study the fire development of full-scale fires in a wide range of conditions. Nevertheless, CFD models require a huge validation effort before they could be widely applied. The main objective of this thesis is to analyze the performance and if possible validate the CFD code FLACS-Fire v10.5 (Flame Accelerator Simulator) for pool-fires. FLACS is a Computational Fluid Dynamic (CFD) program, which solves the compressible conservation equations for mass, momentum, enthalpy, and mixture fraction using a finite volume method. To model a fire it is necessary to include, among others, processes that involve submodels for: turbulence, combustion, thermal radiation, and soot generation. It is of utmost importance, while developing fire models, to validate them against experimental data in pursuance of being able to conclude whether the simulation is valid or not, and to determine the inherent error in comparison with reality. This process consists in a replication of the experimental setup in the CFD, in this case FLACS, and compare it with experimental data previously available. In the present work, gasoline and diesel fuel experimental pool fires were modeled with FLACS-Fire v10.5 code. Simulations considered different pool fire experiments with diameters ranging from 1.5 to 4 meters. In addition, simulations were run with the Eddy Dissipation Concept (EDC) as combustion model; with the κ-ε model as turbulence model; and with the Discrete