High-Fidelity Hydrostructural Design Optimization of Lifting Surfaces

Recent advances in high-performance computing and the efficiency of numerical solvers have made it possible to use sequential high-fidelity hydrodynamic and structural simulations to carry out design and optimization of marine lifting surfaces such as hydrofoils and propulsors. However, the design optimization of flexible hydrofoils and propellers requires coupled hydrodynamic and structural analysis to achieve a truly optimal, physically realizable, and structurally sound design. To address this need, the thesis presents an efficient high-fidelity hydrostructural design optimization with large numbers of design variables, multiple design points, as well as design constraints to avoid cavitation, avoid excessive stresses, and satisfy manufacturing tolerances. The hydrostructural solver couples a 3-D nearly incompressible Reynolds-averaged Navier–Stokes solver with a 3-D structural finite-element solver. The hydrostructural solver is validated by comparing the hydrodynamic load coefficients and tip bending deformations of a cantilevered aluminum hydrofoil with a NACA 0009 cross section and a trapezoidal planform. A coupled adjoint approach for efficient computation of the performance and constraint function derivatives with respect to 210 shape design variables is used. Using this state-of-the-art hydrostructural design optimization tool, a multipoint optimization yields improved performance over the entire range of expected operating conditions with significantly increased cavitation inception speed. The hydrostructural optimal result is compared to an equivalent hydrodynamiconly optimization, and results show that only the hydrostructural optimized design satisfies the stress constraint up to the highest expected loading condition, highlighting the need for coupled hydrostructural optimization. The proposed approach enables multipoint optimization of the hydrostructural performance for hydrofoils and marine propulsors, and it constitutes a powerful new tool for improving existing designs, and exploring new con-