| Abstract | It is necessary to predict the stability of the ship at the moment when it is damaged on the sea. It can also help minimize casualty by predicting the ship’s stability and the time to reach equilibrium. In an emergency, the flooding analysis should be performed within a short time. However, since the analysis is a time domain simulation with complexity, it takes too long time to be used in such circumstance. To reduce the computational cost of estimating ship’s stability, surrogate modeling method is used in this study. Surrogate modeling is a method that uses a simplified model of an actual model to drive results with reduced computational cost while maintaining accuracy as much as possible. The surrogate model can predict the results quickly without performing analysis in the time domain. To construct the surrogate model, various types of a deep artificial neural network (ANN) are tried. Each result of the analysis is used to train the neural network. In the result, accurate estimations are obtained by the surrogate model within a dramatically reduced time. Keyword: Stability, Flooding simulation, Surrogate model, Deep artificial neural network. |
|---|---|
| Publication Date | 2017-09-26 |
Jisang Ha, Myung-Il Roh, Sung-Jun Lee, Ki-Su Kim, Seung-Min Lee, "Toward Rapid Flooding Analysis of a Ship Using Surrogate Model by Deep Learning", Proceedings of the 31st Asian-Pacific TEAM(Technical Exchange and Advisory Meeting on Marine Structures) 2017, Osaka, Japan, pp. 397-400, 2017.09.25-28