基于BP神经网络的海上发射船耐波性优化研究  

Research on seakeeping optimization of marine launcher based on BP neural network

作　　者：王宝来[1] 杨晓杰 刘大辉 WANG Bao-lai;YANG Xiao-jie;LIU Da-hui(Yantai Research Institute of Harbin Engineering University,Yantai 264000,China;CIMC Offshore Engineering Research Institute,Yantai 264670,China)

机构地区：[1]烟台哈尔滨工程大学研究院,山东烟台264000 [2]中集海洋工程研究院有限公司,山东烟台264670

出　　处：《舰船科学技术》2023年第24期47-51,共5页Ship Science and Technology

摘　　要：为了保证海上火箭发射的安全,应用BP神经网络对火箭发射船加以优化。按照海上发射在设备布局方面要求进行总布置设计,建立火箭-发射船刚性连接端的弯矩模型,以方形系数、船长和船宽为优化变量,通过BP神经网络模型对母型船进行优化设计,优化得到的船型弯矩比母型船减少了19.41%;对部分优异样本点进行再建模和数值仿真,仿真结果表明,BP神经网络模型的误差不到1%,验证了模型的准确性。该研究为海上发射船设计优化提供了一种研究思路,为海上火箭发射是否存在风险提供了一种预知方法。In order to ensure the safety of sea launch,BP neural network is used to optimize the launch ship.Carry out general layout design according to the requirements of equipment layout for offshore launch.The bending moment model of the rigid connection end of the rocket launcher is established.Taking the square coefficient,the length and the width of the ship as the optimization variables,the parent ship is optimized through the BP neural network model.The bending moment of the optimized ship is 19.41%less than that of the parent ship.The accuracy of the model is verified by re modeling and numerical simulation of some sample points.The results show that the error of the BP neural network model is less than 1%.This study provides a research idea for the design optimization of offshore launch ship,and a prediction method for risk of offshore rocket launch.

关 键 词：BP神经网络模型 全因子试验 火箭弯矩 发射船优化 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]