Margot PRUVOST, Marc FERRIER, Arnaud MURA
DOI Number: XXX-YYY-ZZZ
Conference number: HiSST 2024-00117
Simulating turbulent combustion in supersonic flows is a challenging yet essential task in designing scramjet combustors. In the high Reynolds number flows passing through the combustion chamber, mixing and chemical time scales have the same order of magnitude. Because of turbulence intermittency, dissipative structures are non-homogeneously distributed, leading to incomplete mixing of chemical species at the molecular level. In order to take into account this uneven distribution of the micro-mixed volumes in turbulent combustion simulations, Partially-Stirred-Reactor (PaSR)-like models have been developed. They assume that each computational cell is composed of a well-mixed region and its surroundings. However, considering a local model to describe the mixing process is a strong assumption since it does not take into account the whole history of micro-mixing. Furthermore, the chemical and mixing time scales are key parameters in these models and their estimation becomes challenging when dealing with complex configurations such as scramjet combustion chambers. In this paper, we introduce a new approach for the simulation of turbulent combustion. It is based on the PaSR concept together with a multi-fluid framework. In the computational domain, two fluids are considered: one relevant to the well-mixed volumes, and the other acting as their surroundings. Two sets of transport equations are considered to describe them. Within this Transported PaSR (TPaSR) framework, species micro-mixing and thermal diffusion are represented by source terms transferring mass and energy between the two fluids. In this study, an expression for mass transfer is proposed and the model robustness is assessed on the NASA Langley Research Center supersonic coflowing burner. The model reproduces the regions featuring the steepest velocity gradients, and ensures mass transfer from one fluid to another at a higher rate in these regions.