A Navier-Stokes Based Strategy for the Aerodynamic Optimisation of a Turbine Cascade Using a Genetic Algorithm

Abstract

The problem of turbine cascade design is considered using automatic optimisation strategies. After the profile parameterisation, two different models are considered to simulate the cascade performances: experimental correlations and a Navier-Stokes code. Interesting considerations on the optimisation algorithm can be drawn using correlations because the evaluation of the fitness is almost instantaneous. By fixing certain of the design constraints two optimised configurations are obtained using simplified correlating function or the viscous flow solver. The optimised configurations are then investigated with an accurate viscous flow computation in order to have a deeper insight into the flowfield.