A Semi-Empirical Model for Conceptual Turbine Vane Cooling Design and Optimization

Abstract

Efficient turbine vane cooling designs are increasingly important to improve the thermal efficiency of gas turbines. Evaluating the performance of a cooling design requires the knowledge of the temperature distribution on the vane surface and the cooling air mass flow rate. The estimation of the vane temperature distribution is considered as a conjugate heat transfer problem, which usually requires a computationally intensive 3D CFD-FEM simulation. However, this approach is not suited for an early design phase, when the cooling design frequently changes. A simplified, yet physical approach is necessary to develop an initial cooling design, which can be used as a baseline for more detailed investigations. This paper presents a predictive model for turbine vane cooling and its integration into an optimization tool chain. The model uses a vane geometry model, the aerodynamic flow field and the coolant conditions from an in-house turbine design tool chain. The cooling geometry is divided into multiple interior sections with their own parameterizations, characterizing the cooling method and its geometric representation. Internal cooling, such as impingement or convective cooling, as well as external cooling, namely film cooling, is considered. Taking the material properties of the vane into account, the model calculates the temperature distribution on the vane surface and the coolant mass flow rate to identify critical hot spots and to evaluate a cooling concept. The capabilities of the model are demonstrated in an optimization process to improve the cooling design of a modern high pressure nozzle guide vane. Compared to a manually created cooling concept, the coolant mass flow rate was reduced by more than 20 \% while simultaneously a more uniform metal temperature was achieved.