International Conference

Jin-Hyeok Kim, Myung-Il Roh, In-Chang Yeo, Ki-Su Kim, Min-Jae Oh, Sejin Oh, "Estimation Model of Hydrodynamic Performance Using Hull Form Variation and Deep Learning", Proceedings of PRADS 2022, Dubrovnik, Croatia, pp.??, 2022.10.09-13

by SyDLab posted Dec 17, 2021
?

단축키

Prev이전 문서

Next다음 문서

ESC닫기

크게 작게 위로 아래로 댓글로 가기 인쇄
Extra Form
Abstract In the design process of hull form, several candidates of hull forms are generated, and CFD (Computational Fluid Dynamics) analysis is typically used to evaluate the hydrodynamic performance of the candidates. If the performance of the evaluated hull form is not good, it is improved through the iterative process of redesigning or fairing the hull form. However, there is a problem that CFD analysis takes a long time to calculate. As the design period of the ship is limited, the iteration is not sufficient to find the optimal hull form. To solve this problem, in this study, we proposed a method to evaluate the performance within a short time by skipping CFD analysis using a deep learning model. To train a deep learning model for evaluating the performance of hull forms, it takes a long time to generate data and train the model, but once the model is trained well, the performance of the hull form can be estimated quickly using the trained model. The hull forms used for training the model are generated by deforming the reference hull form using FFD (Free Form Deformation). The performances derived from the CFD analysis are used as a ground truth. For the better precision of estimation, various structures of the deep learning model were compared, and we selected an appropriate model to predict performances of the hull forms. By using the proposed model, many candidates can be evaluated when designing the hull form. In addition, the efficiency of the design process of the hull form can be increased by selecting only a few good alternatives and performing CFD. In this study, from data generation for the deep learning model, a prediction model’s structure and learning process were proposed and applied to evaluate the performance of various hull forms.
Publication Date 2022-10-09

Jin-Hyeok Kim, Myung-Il Roh, In-Chang Yeo, Ki-Su Kim, Min-Jae Oh, Sejin Oh, "Estimation Model of Hydrodynamic Performance Using Hull Form Variation and Deep Learning", Proceedings of PRADS 2022, Dubrovnik, Croatia, 2022.10.09-13