Robust optimization of turbomachinery cascades using inverse methods

Abstract

It is widely known that real-world airfoils and turbomachinery cascades are subjected to small perturbations in their geometries due to the manufacturing errors and operational and environmental conditions. As a result, the performance requirements of such systems can be significantly affected, leading to poor designs or even severe failures during operation. In this context, the concept of robust optimal design in computational fluid dynamics is very interesting and can be conveniently carried out by using modern multiobjective numerical optimization techniques, in which the interest is to generate optimal and robust aerodynamic shapes that are less sensitive to small perturbations during their useful life. However, the large number of evaluations of the cost functions makes the use of direct methods in the robust optimization of turbomachinery cascades in aerodynamics very costly, sometimes unfeasible. Those difficulties motivate the study reported in this paper, in which a robust optimization strategy by using inverse methods is proposed consisting in the use of multiobjective evolutionary algorithms to be combined with robustness functions specially adapted to turbomachinery cascades.