A Study on Reliability of Kriging Based Approximation Model and Aerodynamic Optimization for Turbofan Engine High Pressure Turbine Nozzle
Sanga Lee, Saeil Lee, Young-Seok Kang, Dong-Ho Rhee, Dong-Ho Lee, Kyu Hong Kim · The KSFM Journal of Fluid Machinery · 2013
In the present study, three-dimensional aerodynamic optimization of high pressure turbine nozzle for turbofan engine was performed. For this, Kriging surrogate model was built and refined iteratively by supplying additional experimental points until the surrogate model and CFX result has effective difference on objective function. When the surrogate model satisfied this reliability condition and developed enough, optimum point was investigated. Commercial program PIAnO was used for optimization process and evolutionary algorithm was used for searching optimum point. As a result, difference between estimated value from Kriging surrogate model and CFD result converges within 0.01% and the optimized nozzle shape has 0.83% improved aerodynamic efficiency.