Ship & Boat ›› 2024, Vol. 35 ›› Issue (03): 81-89.DOI: 10.19423/j.cnki.31-1561/u.2024.03.008

Previous Articles     Next Articles

Neural Network Prediction and Empirical Formula of Ship Bridge Impact Force

YANG Xiaojie1,2, WANG Baolai2,*   

  1. 1. Yantai Salvage Bureau of the Ministry of Transport, Yantai 264012, China;
    2. Yantai Research Institute of Harbin Engineering University, Yantai 264000, China
  • Received:2023-05-30 Revised:2023-08-30 Online:2024-06-25 Published:2024-06-24

船桥撞击力的神经网络预测与经验公式研究

杨晓杰1,2, 王宝来2,*   

  1. 1.交通运输部烟台打捞局 烟台 264012;
    2.烟台哈尔滨工程大学研究院 烟台 264000
  • 通讯作者: 王宝来(1979-),男,博士,副教授。研究方向:结构优化设计。
  • 作者简介:杨晓杰(1997-),男,硕士研究生。研究方向:船舶结构。
  • 基金资助:
    山东省海上航天装备技术创新中心(鲁东大学)开放基金项目(MAETIC202202); 山东省重点研发计划-海上卫星发射及回收(2020CXGC010701); 烟台先进材料与绿色制造山东省实验室开放课题基金(AMGM2022A01); 烟台哈尔滨工程大学研究院校企合作基金(210F0401004)

Abstract: The maximum ship bridge impact force and its relationship with influencing factors are of great significance for the reduction of the harm to the personal security and safety of their belongings caused by ship bridge collision. The finite element model of the ship bridge collision has been established by using ANSYS/LS-DYNA to calculate the ship bridge impact force under different impact velocities and impact angles. And the calculated results are compared with those from the rules of various countries. A model for the prediction of the maximum ship bridge impact force is then established under three training parameters of the maximum impact force value, impact angle and impact velocity by using the Back Propagation (BP) neural network technology together with the simulation data. Finally, an empirical formula of the relationship between the maximum impact force and the impact velocity and impact angle is fitted by analyzing the scatter diagram of the impact velocity and the impact angle. And the results of the empirical formula are compared with the finite element simulation results and the neural network prediction results to validate the accuracy of the empirical formula. It provides a rapid method for evaluating whether the ship bridge collision will have catastrophic consequences.

Key words: ship bridge collision, impact velocity, impact angle, maximum impact force, finite element analysis, BP neural network, empirical formula

摘要: 为了减少船舶与桥梁撞击对人身和财产安全造成的危害,研究船舶与桥梁撞击力的最大值及其与其他影响因素之间的关系具有十分重要的意义。该文首先利用ANSYS/LS-DYNA软件建立了船舶撞桥的有限元模型,计算了不同撞击速度和撞击角度下船桥撞击力的最大值,并将计算结果与各国规范进行比较;然后,利用反向传播(back propagation,BP)神经网络技术,结合仿真计算数据,在最大撞击力值、撞击速度和撞击角度这3个训练参数下,创建了船桥最大撞击力的预测模型;最后,通过分析撞击速度和角度的散点图,拟合了1个最大撞击力与撞击速度和角度关系的经验公式,并将经验公式、有限元仿真计算结果和神经网络预测结果进行比较,验证了经验公式的准确性。该研究为快速评估船桥撞击是否会导致灾难性后果提供了1种有效的方法。

关键词: 船桥撞击, 撞击速度, 撞击角度, 最大撞击力, 有限元分析, BP神经网络, 经验公式

CLC Number: