| description abstract | A numerical procedure is presented for the scaling of lean aeronautical gas turbine combustors to different thrust classes. The procedure considers multiple operating points and aims for a self-similar flow field with respect to a reference configuration. The developed scaling approach relies on an optimization-based workflow which involves automated geometry and numerical grid generation, unsteady Reynolds-averaged Navier–Stokes (URANS) simulations, and postprocessing of the reacting flow field. Kriging is applied as a metamodel to identify new sets of parameters for combustor geometry generation. A generic lean-burn high-pressure aeronautical combustor has been designed to serve as a first verification test case with reactive flow characteristics comparable to real combustion chambers. The burner geometry is parameterized by 23 free parameters which are altered within the scaling process. The definition of a suitable scaling function is essential for the success of the scaling approach. A scaling function based on pressure loss, axial location of heat release, pilot air split, and the temperature profile at the combustor exit is proposed. The developed procedure is tested and applied for the scaling of an internally-staged lean combustor to a lower thrust class considering multiple operating points simultaneously. In total, 65 different combustor variants have been evaluated by the scaling procedure. Simulations were performed for each of these configurations at takeoff, approach, and idle operating conditions. The final combustor configuration, scaled to a lower thrust class, shows good agreement to the reference configuration in terms of the scaling targets and reasonably resembles the emission indices. Integrating the scaling procedure into the design process of future combustion systems could reduce the required design iterations and thereby contribute to significantly reduced development times and costs. | |
