Research on Ship Resistance Prediction Using Machine Learning with Different Samples

Yunfei Yang,Zhicheng Zhang,Jiapeng Zhao,Bin Zhang,Lei Zhang,Qi Hu,Jianglong Sun,Yang,Zhang,Zhao,Zhang,Zhang,Hu,Sun

DOI: https://doi.org/10.3390/jmse12040556

IF: 2.744

2024-03-27

Journal of Marine Science and Engineering

Abstract:Resistance serves as a critical performance metric for ships. Swift and accurate resistance prediction can enhance ship design efficiency. Currently, methods for determining ship resistance encompass model tests, estimation techniques, and computational fluid dynamics (CFDs) simulations. There is a need to improve the prediction speed or accuracy of these methods. Machine learning is gradually emerging as a method applied in the field of ship research. This study aims to investigate ship resistance prediction methods utilizing machine learning across various datasets. This study proposes two methods: employing stacking ensemble learning to enhance resistance prediction accuracy with identical ship samples and utilizing various ship resistance prediction models for accurate resistance prediction through transfer learning. Initially focusing on container ships as the research subject, the stacking ensemble learning model outperforms the basic machine learning model, the Holtrop and Mennen method, and the updated Guldhammer and Harvald method based on comparative prediction results. Subsequently, the container ship resistance prediction model achieves precise resistance prediction for bulk carriers. This study offers dependable guidance for applying machine learning in predicting ship hydrodynamic performance.

oceanography,engineering, marine, ocean

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve

This paper aims to study the use of machine learning methods to predict ship resistance and validate their effectiveness through different datasets. Specifically, the paper proposes two methods:

1. **Stacking Ensemble Learning**: Enhancing the accuracy of resistance prediction through the same ship samples.
2. **Transfer Learning**: Achieving accurate predictions by using resistance prediction models of different types of ships through transfer learning.

The paper first takes container ships as the research object. By comparing experimental results, it shows that the stacking ensemble learning model outperforms basic machine learning models, the Holtrop and Mennen method, and the updated Guldhammer and Harvald method in predicting container ship resistance. Subsequently, the container ship resistance prediction model is successfully applied to the resistance prediction of bulk carriers, providing reliable guidance for applying machine learning to predict ship hydrodynamic performance.

### Research Background and Current Status

Currently, methods for predicting ship resistance mainly include model tests, estimation techniques, and Computational Fluid Dynamics (CFD) simulations. These methods have their own advantages and disadvantages. For example, model tests are reliable but time-consuming and costly, estimation techniques are fast but less accurate, and CFD methods require high computational resources. Therefore, researchers hope to improve prediction speed and accuracy through machine learning methods.

### Main Contributions

1. **Stacking Ensemble Learning Model**: By combining multiple base models (linear regression, K-nearest neighbors, support vector regression, and support forest), a stacking ensemble learning model is constructed to improve prediction accuracy.
2. **Application of Transfer Learning**: Using a pre-trained container ship resistance prediction model, transfer learning is successfully applied to the resistance prediction of bulk carriers, validating the effectiveness of this method.