The accurate prediction of ship resistance has great significance. Generally, resistance prediction accuracy at the drafts outside of training set is poor compared with that at the drafts inside of training set. In this paper, a prediction method for the resistance at outside drafts is proposed using the hybrid support vector regression - back propagation (SVR-BP) model. Ship resistance is predicted by SVR firstly, and BP is used to make for the errors between SVR prediction results and experimental values. The study shows that the prediction accuracy using the SVR-BP model is better than using the single SVR or BP model.
Resistance prediction has always been one of the hot regions of ship research. Approximate methods such as series data, empirical formula, and parent ship estimation are usually used to predict ship resistance. Some methods are modified for different ship types and working conditions (Jeong et al., 2017; Liu et al., 2014; Kristensen, 2013; Tu et al., 2018). With the development of computer technology, CFD technology has been widely used in ship performance calculation (Guo et al., 2013; Li et al., 2014; Sun et al., 2016; Tu et al., 2018; Lv et al., 2013; Lyu et al., 2018). Compared with the traditional resistance prediction methods, CFD technology has higher accuracy. Both of these prediction methods have shortcomings. The prediction accuracy of approximate methods needs to be improved, and the CFD technology needs more time and requires high computational resources.
Artificial intelligence (AI) algorithms such as machine learning and deep learning have a strong ability to deal with nonlinear and complex problems and are making marks in the regions of image recognition and speech synthesis. These algorithms are also increasingly used in the study of shipping (Atmane Khellal et al., 2018; Niu et al.,2017; Guan et al., 2018) and fluid mechanics (Ling et al.,2016; Wang et al., 2017; Rabault et al., 2019; Raissi et al., 2018). The current application of AI algorithms in ship resistance prediction remains at the stage of using resistance maps for interpolation simply. The research on the application of AI algorithms in ship resistance prediction is not enough, especially for the resistance prediction at the drafts outside of training set for the same ship.
