规则波中船舶操纵运动预报的灰箱建模研究  

Prediction of ship maneuvering motion in regular wave with gray-box modelling

作　　者：韩阳[1,2] 郝立柱 师超 潘子英[1,2] 鲁江 顾民[1,2] HAN Yang;HAO Lizhu;SHI Chao;PAN Ziying;LU Jiang;GU Min(State Key Laboratory of Hydrodynamics,China Ship Scientific Research Center,Wuxi 214082,China;Taihu Laboratory of Deepsea Technological Science,Wuxi 214082,China)

机构地区：[1]中国船舶科学研究中心水动力学全国重点实验室,江苏无锡214082 [2]深海技术科学太湖实验室,江苏无锡214082

出　　处：《中国舰船研究》2025年第1期47-57,共11页Chinese Journal of Ship Research

基　　金：国家部委基金资助项目。

摘　　要：[目的]针对船舶操纵运动实时准确预报的需求,开展规则波中船舶操纵运动预报的灰箱建模研究。[方法]建立船舶操纵运动方程,以表征操纵运动机理。应用泰勒级数展开方法近似静水水动力,采用经验公式估算规则波中二阶定常波浪漂移力,形成规则波中船舶操纵运动预报的数学模型。采用傅里叶变换方法解决不同频率的操纵与耐波运动数据分离问题,基于操纵运动数据和深度神经网络(DNN)技术,构建静水水动力修正及二阶定常波浪漂移力模型,并将其代入操纵运动机理方程,形成融合机理与数据的规则波中船舶操纵运动预报灰箱模型。然后以ONRT为研究对象,分别应用灰箱模型和数学模型预报规则波中船舶操纵运动。[结果]结果显示,对于所有运动工况,仿真单位时间步长耗时平均约2~3ms,灰箱模型预报结果与试验数据相比其相对精度均值达94.83%,相比数学模型预报精度平均提高了4.50%。[结论]灰箱预报模型可以作为规则波中船舶操纵运动预报的有效方法,能为真实海洋环境中船舶操纵运动的实时预报奠定基础。[Objectives]Aiming at the requirements of the real-time and accurate prediction of ship maneuvering motion,this paper investigates the prediction of ship maneuvering motion in regular waves using graybox modelling to improve the accuracy.[Methods]A maneuvering motion equation is proposed to reveal the known movement mechanism.The hydrodynamic forces in calm water are approximated using the Taylor series expansion,and the second-order steady wave drift forces are estimated using empirical formulas,thereby obtaining the mathematical model for predicting ship maneuvering motion in regular waves.To promote the precision of the hydrodynamic expression,the Fourier transform method is adopted to separate the data of the maneuvering and seakeeping motions.A model for hydrodynamic correction and second-order steady wave drift forces is developed on the basis of the maneuvering motion and deep neural network(DNN)data,then submitted into the mechanistic equation of maneuvering motion.Finally,an gray-box modelling incorporation mechanism and data for predicting ship maneuvering motion in regular waves is established.[Results]Taking the Office of Naval Research Tumblehome(ONRT)as an example,the maneuvering motion is predicted with the adoption of the mathematical model and gray-box model respectively.For all simulated cases,the simulation of the unit time step costs 2–3 ms on average,and the average error between the results of the gray-box model and experiments is 94.83%,with the accuracy promoted by an average of 4.50%compared with the mathematical model.[Conclusions]Gray-box modelling can be recognized as an efficient method for predicting ship maneuvering motion,laying a foundation for the real-time prediction of maneuvering motion in real marine environments.

关 键 词：船舶 操纵性 操纵运动 灰箱建模 多层神经网络 傅里叶变换 ONRT 

分 类 号：U661.1[交通运输工程—船舶及航道工程]