RIASSUNTO
ABSTRACT
Seakeeping is a very important indicator for ship performance, the shape of ship is the key factor that affects seakeeping, many researchers have been done a lot of work about it. In this paper, a bow lines optimization method using deep learning and (1+1)-evolution strategy, which are two popular algorithms, is proposed and applied to the ship with a vertical bow. A reasonable parameterization method and Latin hypercube sampling method are adapted as auxiliary tools for optimization, the numerical calculation method of the potential theory is used to simulate the seakeeping performance of ships at different given sea states. The results show that the bow lines optimization method for better seakeeping performance by the approach of deep learning and (1+1)-evolution strategy is feasible.
INTRODUCTION
In order to improve ship performance, the optimization of ship form has always been a hot research topic for ship designers, many new technologies (Sun et al., 2016) have been studied to improve navigation resistance performance and new ship standards (Tu et al., 2019) have been introduce for ship energy saving. In the respect of seakeeping study, researchers also done a lot of work to find the factors that affect the seakeeping performance and optimize the hull form for a better seakeeping performance using technology of single-objective optimization, multi-objective optimization and so on. S. M. Wang et al. (2018) revealed that the outrigger layout of a trimaran has a significant influence on the seakeeping performance. A. Scamardells, V. Piscopo (2014) proposed a new index, Overall Motion Sickness Incidence (OMSI) and assumed it as the parameter to minimized in the singleobjective optimization procedure. Hassan Bagheri et al. (2014) obtained the seakeeping performance using computational method, and optimize hull form by genetic algorithm with the ship displacement in a constraint range. They took the well-known S60 hull and Wigley as the initial hulls and carried the optimization work. Mark A. Gammon (2011) took a fishing boat as a example to carry out a Multi-Objective, Genetic Algorithm (MOGA) optimization. The ship resistance, seakeeping and stability were the three indices for optimization, and finally got the optimal hull offset and optimal values for the principal parameters of the fishing ship. Miao. A. et al. (2018) conducted the multi-objective optimization design of ship hull based on seakeeping performance in different waves. Bonfiglio. Luca et al. (2018) combined the strip theory and a boundary element method based on potential theory, took the combined method as the tool to conducted a multi-fidelity optimization research.