Optimization of a Cylindrical Film Cooling Hole using Surrogate Modeling

Abstract

Optimization of a cylindrical film-cooling hole has been performed by surrogate modeling approach using three-dimensional Reynolds-averaged Navier-Stokes analysis. SST model has been employed as a turbulence closure model for the analysis of turbulent convective heat transfer. Spatially-averaged film-cooling effectiveness has been maximized for optimization. For two design variables, the ratio of length to the diameter of the hole and ejection angle, 12 experimental points are selected by Latin hypercube sampling. Performances of three basic surrogate models and three weighted average surrogate models have been evaluated. Among the surrogate models tested, the Kriging model predicts the optimum point with the highest objective function value calculated by Reynolds-averaged Navier-Stokes analysis, which gives a 3.6% improvement of the spatially-averaged film-cooling effectiveness in comparison to the reference geometry. The objective function is more sensitive to the ejection angle than the ratio of length to diameter of the hole.