A Multi-fidelity Aeroelastic Optimization of an Aircraft Wing Using Co-Kriging

In the present study, a multi-fidelity, multi-objective and multi-disciplinary design problem statement for aeroelastic optimization of an aircraft wing has been posed. The problem minimizes the twin objectives of transonic drag and the weight of the structure during cruise flight at Mach 0.8 at 10 km altitude. The wing geometry was parametrized using 11 design variables. A multi-fidelity-based co-kriging metamodel was used to replace the multi-disciplinary analysis routine in the optimization problem. Reynolds-averaged Navier–Stokes (RANS) and Euler solvers were used as high and low fidelity aerodynamic solvers while refined and coarse meshes were used with a default Finite Element (FEM) solver for multi-fidelity structural analysis. Latin hypercube sampling was used to generate 100 design points, out of which 30 were used to perform high fidelity simulations and the rest were used for low fidelity simulations. A high fidelity MDA routine run had a computational time of 4 h while the low fidelity runs were of 30 min. A successful and computationally inexpensive coupled aeroelastic optimization methodology has been demonstrated using MDF and co-kriging. The aerodynamic coefficients Cl and Cd showed improvement of 4.69% and 17.9%, respectively, compared to the baseline values. The structural weight of the optimized geometry was reduced by 355.7 Kg, and there was 14.54% reduction in the maximum von-Mises stress in the optimized structure.