A novel surrogate-based aerodynamic optimization method using field approximate model

摘要

An efficient surrogate-based aerodynamic shape optimization method is developed to improve the optimization efficiency. In this method, the field approximate model is presented firstly to predict the flow field parameters of interest for specific aerodynamic optimization problems with respect to the design variables and sequentially updated. The differential evolution is used to locate the optimum of field approximate model coupled with the analytical post-processing to calculate the objective and constraints for aerodynamic optimization. This optimal point is calculated by time-consuming computational fluid dynamics simulation and the result is added to the sampling set to update the sampling points and field approximate model. The proposed method is compared with conventional sequential approximate optimization and shows great advantages in accuracy and efficiency. Two shape optimization test cases are provided to verify the efficacy and efficiency of the proposed method.