RIASSUNTO
ABSTRACT
For a hull form optimization design with a relatively large number of design variables, numerous new sample hulls need to be evaluated which cause plenty of computation resource. Therefore, in order to avoid highdimensional sampling, a design-space dimensionality reduction method based on Proper Orthogonal Decomposition is implemented while retaining the geometric variation in the original design space as much as possible. The Wigley hull is considered as the initial ship in this paper. After the hull form deformation in the original design space is set, the design-space dimensionality reduction is carried out and a new, lowerdimension design space can be obtained. Finally, a comparative study of total drag optimization with or without design-space dimensionality reduction is given by doing several optimizations using OPTShip-SJTU. The optimal hulls are further validated by high-fidelity CFD solver naoe-FOAM-SJTU, showing that doing the design-space dimensionality reduction can reduce computation costs while giving relatively good optimization results compared to those in the original high-dimension design space.
INTRODUCTION
In the past few years, huge development of computer technology and calculation theories allow us to do the Simulation-Based-Design (SBD) other than use empirical or semi-empirical formulas to optimize a ship hull form in order to improve its comprehensive hydrodynamic performances. It is a new design method which integrates hull form modification method, numerical simulation tool and optimization technology. For instance, Campana, Peri, Tahara, and Stern (2006) used the Non-Uniform Rational B-Spline (NURBS) surface modeling method to modify a bulbous bow, and the modified hulls were evaluated by RANS-based solver, and the bulbous was optimized; Yang and Huang (2016) used surrogate models to perform three optimization cases and were validated by cross validation, where each sample point was evaluated from the surrogate model constructed by the rest of the sample points; considering uncertainty, Liu, Wang, and Wan (2018) gave a probability optimization of KCS hull's resistance at full speed range. As mentioned above, researches have focused on the hull form modification, hydrodynamic evaluation, surrogate model construction, and optimization algorithms in the field of hull form optimization.