Session: 15-01-02 Seakeeping and Operability

Submission Number: 183622

Application of Digital-Twin Concept for Prediction of Risk due to Parametric Roll 

ABSTRACT

The occurrence of intact stability accidents associated with parametric resonance remains a critical challenge in maritime operations. Despite advancements in ship design and operational techniques, modern containerships are still susceptible to parametric roll due to increased variations in vertical wetted surfaces.

In this study, adopting the digital twin of ship operation, the occurrence and risk of parametric roll of a containership are predicted. The concept of real-time digital twin for the prediction of ocean wave fields and correspondent ship motion was introduced by Lee et al.(2022). This approach is based on the measurement of wave field by using X-band radar, the prediction of future wave fields near the ship, and the prediction of corresponding ship motions. This sequence was achieved successfully in real time, and all their methods were based on physics-based models. In this study, an AI-based approach is adopted for the prediction of parametric roll. Since the parametric roll problem is nonlinear, the analysis time takes more time than the linear ship motion problem. Very recently, Lee and Kim(2025) showed that a machine-learning technique can be applied to reduce the analysis time without loss of accuracy.

The machine-learning model adopted in the present study is by incorporating a convolutional neural network (CNN) and long short-term memory (LSTM) network to predict ship responses from the surrounding wave field time series. The CNN-LSTM integrated model effectively accounts for spatiotemporal features of three-dimensional wave field inputs. The developed analysis module is attached with the digital twin for ship operation, and the occurrence and degree of parametric roll are validated by comparing with other numerical simulations.

KEY WORDS: parametric roll, digital twin, machine learning, CNN, LSTM

REFERENCES

[1] Jae-Hoon Lee, Yoon-Seo Nam, Yonghwan Kim, Yuming Liu, Jaehak Lee, Heesuk Yang, Real-time Digital Twin for Ship Operation in Waves, Ocean Engineering, 2022.

[2] Jaehak Lee, Yonghwan Kim, Application of Spatiotemporal Wave Field-based Neural Network for Predicting Parametric Roll Motions, Ocean Engineering, 2025.

 

Presenting Author: Jaehak Lee Seoul National University

Presenting Author Biography: Dr.Jaehak Lee is currently a post-doc at Seoul National University. His research interests includes the digital twin for ship operation, wave-field analysis and prediction based on X-band radar, physics- and AI-based seakeeping analysis. He has nice peer-review journal papers and more than 10 conference papers and presentations so far. He is a awardee of JH Hwang Fellowship, young researcher award, by Society of Naval Architects of Korea.

Authors:

Jaehak Lee Seoul National University
Yonghwan Kim Seoul National University