International Conference

Min-Jae Oh, Myung-Il Roh, Young-Soo Seok, and Sung-Jun Lee, "Optimization of a Ship Hull Form using Deep Learning", Proceedings of ACDDE 2019, Penang, Malaysia, 2019.07.07-10

by SyDLab posted Jul 03, 2019
?

단축키

Prev이전 문서

Next다음 문서

ESC닫기

크게 작게 위로 아래로 댓글로 가기 인쇄
Extra Form
Abstract Design of a ship hull form is very complex and time consuming process. It starts from the selection of a parent ship, and a designer modifies the selected ship to meet the owner’s and Class requirements. When the hull form is modified, the hydrodynamic analysis is conducted. If the hydrodynamic performance is acceptable, the verification is done through the model test. If the hydrodynamic performance is not acceptable, a designer modifies the hull form manually until it satisfies a certain requirement. During this process, a lot of time is consumed, and it requires the designer’s experiences. The modification methods can be different, and the quality of the hull form can be varied from the designer’s proficiency. In this study, the optimization method is proposed to obtain the optimized hull form automatically using the reinforcement learning that is one of the deep learning methods. The smallest total resistance of a hull form is used as the reward in the reinforcement learning, but the other hydrodynamic performance values can be used as the rewards. The KVLCC2 tanker that is a public hull form is used to get the optimal hull form, and the result shows that the proposed method can generate the optimal hull form. It is expected that the proposed method can be used in the hull form design to reduce the time and enhance the performance.
Publication Date 2019-07-08

Min-Jae Oh, Myung-Il Roh, Young-Soo Seok, and Sung-Jun Lee, "Optimization of a Ship Hull Form using Deep Learning", Proceedings of ACDDE(Asian Conference on Design and Digital Engineering) 2019, Penang, Malaysia, 2019.07.07-10