Aim: The purpose of this study is to verify the usability of the k-ω SST turbulence model for the description of the combustion process during a vertical propane jet fire. Simulating a jet fire using computational fluid mechanics involves an appropriate selection of a mathematical model to describe the turbulent flow. It is important as the variables from this model also describe the rate of the combustion reaction. As a result, they have an impact on the size and shape of the flame. The selection of an appropriate model should be preceded by preliminary simulations. Project and methods: For this purpose, a vertical jet fire in no wind conditions was selected for simulation. Consequently, it was possible to develop a two-dimensional axisymmetric geometry. A good numerical mesh can be applied to such axisymmetric geometry. Selected process conditions allowed to create an axisymmetric numerical grid. Its values, proving the quality, are shown in a chart demonstrating the distribution of the parameter quality depending on the number of elements from which the numerical grid was built. In the work, a two-stage model of the combustion reaction was selected in order to verify whether the area in which the mole fraction of carbon monoxide will have significant values is so large that the selected kinetic reaction model will have an impact on the flame length. Results: Three simulations of jet fire taking place in the direction opposite to the force of gravity were performed. The simulations performed allowed for setting the basic L f parameter, which determines the flame length. Additionally, the length of the mixing path slift-off, needed to initiate the combustion reaction, was determined. The simulations performed allowed for comparing significant parameters characterizing the flame with the parameters calcu- lated using correlations included in the literature on the subject. Due to this comparison, it was possible to define an interesting scope of research work, because the length of the gas mixing path determined from the CFD simulation differed significantly from the values calculated from the correlation. Conclusions: Interestingly, such large differences between CFD results and correlations were not observed for the L f parameter. The correlations based on the Froude number give slightly higher values of the flame length than the results of the CFD simulation. On the other hand, the correlation based on the Reynolds number gives slightly lower values of the L f parameter than the values obtained from the CFD calculations. This may indicate that the effects related to the inertia forces (Re number) better describe the simulation process conditions than the correlations based on the influence of inertia forces and gravity forces (Fr number). Keywords: jet fire, mathematical modelling, computational fluid dynamics Type of article: short scientific report
Read full abstract