A Study of the CST Parameterization Characteristics

Abstract

This paper presents a study of the recently introduced Class-Shape function Transformation (CST). The study assesses some of the main characteristics that parameterization exhibits, such as the sensitivity to parameter perturbations, the uniqueness of a geometric representation, its filtering capability and parameter inter-dependence. In a gradient-based shape optimization process, the solutions of intermediate design cycles are naturally subject to numerical inaccuracy. Under such circumstances, those characteristics of the CST may affect the path and, eventually, compromise the performance of the optimization process depending on the gradient calculation method and minimum-search algorithm. A comparison of two different gradient calculation methods is presented in the context of inverse aerodynamic design along with some validation test cases.