船舶运动模型辨识在海事事故复现中的应用  

Application of Ship Motion Model Identification to Maritime Accident Reproduction

作　　者：蒋柳杨 孙洪波[1] 张春雨 沈莉婷 JIANG Liuyang;SUN Hongbo;ZHANG Chunyu;SHEN Liting(Navigation College of Jimei University,Xiamen 361021,China)

机构地区：[1]集美大学航海学院,福建厦门361021

出　　处：《集美大学学报(自然科学版)》2025年第2期141-149,共9页Journal of Jimei University:Natural Science

基　　金：福建省自然科学基金项目“基于耦合模型的船舶锚泊安全辅助系统关键技术研究”(2023J01325)。

摘　　要：为了再现海事事故发生过程,模拟事故前船舶操纵及运动轨迹,以“BALAWAN”轮碰撞事故为研究对象,以事故船航行数据记录仪记录的航行数据为训练样本,提出采用最小二乘支持向量机(least square support vector machine,LS-SVM)方法,对船舶分离型运动数学模型的水动力导数进行辨识。首先对实船航行数据进行修复、滤波和筛选,然后构建模型训练样本进行模型参数辨识,最后开展事故船轨迹复现仿真试验。通过与真实航行轨迹对比,验证了LS-SVM辨识方法可以作为海事事故复现的一种途径。In order to reproduce the process of maritime accidents and simulate the ship s maneuvering and motion trajectory before the accident,this article studies the collision case of the vessel Balawan,using the VDR navigation data of the accident ship as training samples,the least square support vector machine(LS-SVM)method employed for determining the hydrodynamic derivatives of MMG(ship manoeuvring mathematic model group)model.Firstly,the real ship navigation data was repaired,filtered and screened,then model training samples were constructed and model parameter identification was conducted.Finally,the simulation tests of the trajectory reproduction were carried out.By comparing with the real voyage trajectory,it was verified that the LS-SVM identification method can be used as a way to reproduce maritime accidents.

关 键 词：海事事故 船舶运动数学模型 最小二乘支持向量机回归 轨迹复现 参数辨识 

分 类 号：U661[交通运输工程—船舶及航道工程]