Robust Design Optimization of Film Cooling Hole Array for High Pressure Turbine Nozzle

Abstract

Film hole array optimization has been started and considered recently, despite its various difficulties. With its early stage attention in the research field, there are many issues that should be addressed. One of the significant issues in film hole array optimization is the existence of uncertainty in a high-pressure turbine. Film holes in the 1st stage turbine nozzle work under high uncertainty conditions, the main source of which arises from the manufacturing tolerance and varying flow conditions of the turbine inlet and cooling system. Without consideration of these factors, the optimization results can be ineffective or cause critical failure of the mission if there is any difference between the operating and design condition. In this study, three separate robust design optimization studies for a film cooling hole array are performed under consideration of manufacturing and operational uncertainties. To determine the design variables, the film hole array is parameterized by using newly suggested shape functions with five design variables. The Efficient Design Optimization method coupled with the Kriging model and Monte Carlo simulation, as well as the Genetic Algorithm are used as robust design optimization methods. The manufacturing tolerance and blowing ratio variance of a film hole and the turbine inlet temperature profile are considered as an uncertainty and probabilistic density function, and the variation range of these uncertainties are quantified referring to the open literature by several random variables Thus, film hole arrays showing high cooling performance and high robustness to the uncertainties are successfully obtained

sequentially, the results are compared with each other to derive the following conclusions. Manufacturing tolerance is the most influential followed by variation of the blowing ratio, while the variation of the turbine inlet temperature profile hardly affects the film cooling performance. The region whose temperature fluctuates the most on the nozzle surface appears differently according to the uncertainties, but the random variables related to the holes near the leading edge of the nozzle have a larger impact on the cooling performance than the others regardless of the type of uncertainty.