Author(s): Dai Chenxiao; Liu Kun; Yu Tongqiang; Wang Rong
Linked Author(s):
Keywords: Sea ice; Polar ship and offshore structure; Ship-ice interaction; Structural stress; Machine learning; Convolutional neural network
Abstract: As global climate change accelerates, the melting Arctic ice opens new maritime shipping routes, particularly in the polar regions, presenting significant opportunities for resource development and trade efficiency. However, this also introduces critical challenges related to the safe navigation of vessels in icy waters. One of the key risks involves ship-ice interactions, which can cause substantial damage to ship structures, particularly in ice-heavy environments. The structural stress and damage of ship hull frame structures under ship-ice interaction are examined, focusing on rapid stress evaluation and predictive models. Using a combination of numerical simulations and machine learning techniques, we develop a fast, precise method for predicting the structural responses of ship hull frames under various ice loading scenarios. The research emphasizes the integration of artificial intelligence methods, including convolutional neural networks (CNN), to enhance the speed and accuracy of stress predictions. Through detailed simulations of ice conditions such as ice thickness, collision angles, and ship speeds, the study provides a comprehensive framework for evaluating the safety and resilience of vessels in polar ice zones. The outcomes of this work are intended to assist in the design and optimization of ships for polar navigation, ensuring safer operations and contributing to the development of polar shipping routes.
Year: 2026