Large Eddy Simulation of Turbulent Combustion using Adaptive Mesh Refinement in a typical Micro GasTurbine Combustor

Antoine Verhaeghe1, ⋆, Alessio Pappa1, Ward De Paepe1, Pierre Bénard2, Laurent Bricteux1
: antoine.verhaeghe@umons.ac.be
1 Université de Mons
2 INSA Rouen
Mots clés : Large Eddy Simulation (LES), Turbulent Combustion, micro Gas Turbine (mGT), Adaptive Mesh Refinement (AMR), Thickened Flame Model (TFLES)
Résumé :

The past few years developments in advanced cycle modifications aimed at making micro Gas Turbines (mGTs) more fuel and operational flexible. However, accurate data, on real industrial combustor configuration, assessing the performances and emissions of the combustion under unconventional diluted conditions or fuels involved by these novel cycles are still missing. In this framework, Large Eddy Simulations, which allow to accurately assess the unsteady effects coupled to turbulent-chemistry interaction of reacting flows, offer an opportunity to better assess the combustion behavior under these specific conditions. However, the computational cost still remains much higher than RANS simulations. Moreover, an adequate mesh has to be generated to perform valid LES. However, for such complex geometries, the mesh generation might be complex, especially when the region of interest is not intuitively known. In this work, a dynamic adaptive mesh refinement strategy is proposed and applied on the LES of a typical industrial mGT combustor, the Turbec T100. This strategy consists in automatically refining the mesh along the simulation only in regions where finer cells are required to capture essential effects, i.e. the flame front, at a reasonable cost. Thus, the flame sensor, allowing to identify the flame front location, is used as adaptation criterion. The triggering of the mesh adaptation over time is based on the relative metric error, representing the deviation from the desired mesh metric. The results show that the adaptation strategy allows to generate automatically a dynamic mesh that is able to capture correctly the flame. Moreover, although the mesh adaptation process has a non-negligible calculation cost, a much heavier static mesh would be required to achieve the same resolution accuracy. However, this strategy, simply based on the flame sensor, does not allow the automatic and correct meshing of the dilution zone at the combustion chamber outlet, which is necessary to predict emissions and flue gas properties precisely. Therefore, to ensure a fully automatic meshing of the combustion chamber to avoid human intervention, an additional adaptation criterion must be considered.

doi : https://doi.org/10.25855/SFT2022-070

PDF : download