High-Performance Yacht Shape Optimization by CFD Analysis Coupled with POD

This thesis, conducted in collaboration with ESTECO SPA and OPTIMAD SRL, presents a study on the single-objective shape optimization of a high-performance yacht. The MANTA MKII ASD, designed by Adragna Yacht Studio Design, is used as the baseline reference model. The original configuration has been modified to optimize its aerodynamic and hydrodynamic performance, with a specific focus on minimizing drag. The optimization process relies on computational fluid dynamics (CFD). However, due to the high computational cost associated with optimization techniques, which require a significant number of design evaluations, reduced-order models (ROMs) were developed to significantly decrease computational time and cost while maintaining acceptable accuracy levels. The study begins by introducing the theoretical foundations underlying the employed methodologies and tools. It then focuses on selecting the most suitable models and simulation setups for the investigation. High-fidelity CFD analyses were conducted to construct a comprehensive database for training the ROMs. Geometrical modifications of the baseline model were performed using surface morphing techniques, leveraging free-form deformation algorithms and radial basis functions to modify key hull parameters. The design of experiments employed a Full Factorial algorithm to ensure adequate exploration of the design space. The ROMs were based on a Proper Orthogonal Decomposition (POD) approach combined with interpolation using radial basis functions. The single-objective optimization problem was solved using an algorithm which effectively combines global and local exploration to identify optimal solutions. The study achieved significant results, including hull designs with drag reductions of up to 16%. Moreover, computational times for fluid dynamic analyses were reduced from approximately 4 hours to just 25 seconds, with errors limited to 6% when compared to full-order model results.