Optimum trim prediction for container ships based on machine learning. (1st June 2023)

• Record Type: Journal Article
• Authors: Tu, Haiwen; Xia, Kai; Zhao, Enjin; Mu, Lin; Sun, Jianglong
• Is Part Of: Ocean engineering. Volume 277(2023)
• Publication Date: 2023-06-01
• Subjects: Container ship -- Optimum trim prediction -- Green ship technology; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2022.111322
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: Recent ship-related energy conservation efforts have directed increasing attention on green technology that has the potential to save energy in the shipping industry while protecting the environment. In particular, ship optimum trim energy-saving technology has several advantages, including facile implementation, convenient operation, and high energy-saving effects. This report proposes a method for predicting the optimum trim of container ships based on machine learning, and the developed approach can quickly determine the optimum trim of any container ship to achieve minimum resistance during operation. First, six container ship models from a trim optimization test database were consulted to extract the characteristic parameters of the container ships, thus providing a basic dataset for model training. Four machine learning models were selected to forecast the resistance of the container ship under different trim conditions. The results indicate that the performance of the random forest prediction model was significantly better than the three other tested models (i.e., backpropagation neural network, decision tree, K-nearest neighbor). Therefore, the random forest prediction model was used as the optimal prediction model for determining the optimum trim of container ships. Specifically, the 4700-TEU and 13500-TEU container ships were evaluated; relative to the experimental data, the prediction accuracy reached 85.71% and 88.89%, respectively. Finally, the developed model was applied to the 4250-TEU container ship operation, and the optimum trim angle was predicted under five sets of conditions; the predicted values were consistent with the experimental values. The container ship optimum trim prediction method based on machine learning described in this paper can predict the optimum trim of any container ship (in a certain state), guide its operation, realize energy-saving effects, reduce emissions, and promote the development of green ship technology. Highlights: This paper proposes a method for predicting the optimum trim of any container ship based on machine learning. This paper constructs a machine learning database containing experimental data of six container ships. The procedure for predicting the optimum trim of any container ship is developed to guide practical operation.