An Adaptive Searching Kriging Surrogate Model for Aerodynamic Optimization

Abstract

The Kriging-based genetic algorithm applied to aerodynamic optimization design encounters a problem of unexpected sample size. In this paper, an adaptive method is proposed that the search space moves with the local optimum. The automatic division and hierarchical approximation of search space are realized by taking the selection of refinement samples into account. The typical function optimization and transonic supercritical airfoil drag reduction design are performed using this method. Results show that the number of samples required is greatly reduced, and the aerodynamic performance of the airfoil is efficiently improved.