Optimization of Low-Pressure Turbine Blade by Means of Fine Inspection of Loss Production Mechanisms

In this work, proper orthogonal decomposition (POD) was applied to large eddy simulations (LES) of two high-loaded low-pressure turbine cascades under unsteady inflow to investigate entropy production in different parts of the blade passage. The turbulent kinetic energy (TKE) production, diffusion, and dissipation terms from the stagnation pressure transport equation were integrated over the computational domain. The POD-based method splits the contribution of different coherent flow dynamics to TKE production and dissipation, including the migration, bowing, tilting, and reorientation of incoming wake filaments, as well as the breakup of streaky structures in the blade boundary layer and the formation of Von Karman vortices in the blade wake. This helps designers identify the dominant POD modes (turbulent flow structures) responsible for loss generation, their dynamics, and the spatial locations where they act, providing insights into the physical phenomena to be controlled. Following this optimization strategy, a new low-pressure turbine profile was designed. After performing LES on the optimized geometry, POD results indicated the possibility of designing a higher-loaded profile with a lower global loss coefficient. Thanks to the stronger acceleration imposed on the bulk flow in the former part of the blade passage, the new loading distribution is shown to be responsible for lower upstream wake migration losses, as well as for a smaller amount of TKE production in the trailing edge wake zone, as a result of the early suction side boundary layer transition.