Robust optimization of flexible wing using Stochastic Kriging surrogate model

The stochastic surrogate model method is applied in the robust optimization design of flexible wing. Comparing with deterministic optimization design, robust design can take into consideration the disturbances of design variables and parameters. Therefore, the performance of design results can be kept stable under uncertainties. A High-fidelity fluid-structure coupling solver (coupled Navier-Stokes equations and static structure equations) is used to analyze the deformation and aerodynamic characteristics of flexible wing. In order to enhance optimization efficiency, Stochastic Kriging (SK) surrogate model, which extends deterministic Kriging (DK) surrogate model into stochastic space, is built. The intrinsic uncertainties of data is acquired by finite number of inputs. The robust optimization of flexible M6 wing illustrates that comparied with robust optimization results of DK, the drag coefficient of optimal result of SK has reduced 2.8 counts and the mean values in variable MachNumberrange have reduced 3.2 counts. The optimal result of SK has higher aerodynamic efficiency in design point and better drag divergence characteristics. It is manifested that stochastic surrogate model is favorable in robust optimizationand the SK surrogate model possesses high prediction accuracy.