Investigation of Adaptive Mesh Refinement on an Industrial Gas Turbine Combustor

Abstract

An adaptive mesh refinement (AMR) method is demonstrated for Large Eddy Simulations (LES) of an industrial gas turbine combustor, the SGT-100 provided by Siemens Energy Industrial Turbomachinery Ltd. In this paper, the simcenterstar-ccm+® solver is used to dynamically refine a series of Large Eddy Simulations with a Flamelet Generated Manifold (FGM) combustion model as applied to the SGT-100. Mesh refinement criteria are defined using second gradients of mixture fraction and reaction progress. Two meshes are assessed with and without AMR. The results are then compared to a refined static mesh and experimental data. The accuracy and computational cost of the static and adaptively refined meshes are discussed. It is shown that AMR can provide close to 2× speed up compared to a refined static mesh with similar predictions of mean and RMS quantities of the flow field, flame temperature and major species.