基于多元回归-神经网络的船型参数优选分析  

Optimization Analysis of Ship Form Parameters with Multiple Regression and Neural Network

作　　者：黄晓玲 高玉玲[1,2] 董国祥[1,2] HUANG Xiaoling;GAO Yuling;DONG Guoxiang(State Key Laboratory of Navigation and Safety Technology,Shanghai Ship and Shipping Research Institute Co.,Ltd.,Shanghai 200135,China;Key Laboratory of Marine Technology Ministry of Communications,Shanghai Ship and Shipping Research Institute Co.,Ltd.,Shanghai 200135,China)

机构地区：[1]上海船舶运输科学研究所有限公司航运技术与安全国家重点实验室,上海200135 [2]上海船舶运输科学研究所有限公司航运技术交通行业重点实验室,上海200135

出　　处：《上海船舶运输科学研究所学报》2023年第2期13-17,68,共6页Journal of Shanghai Ship and Shipping Research Institute

摘　　要：以肥大型船舶模型的试验数据为研究对象,分析单一船型参数与剩余阻力系数和自航因子的相关性。基于多元回归分析和径向基函数(Radial Basis Function,RBF)神经网络,对船型参数的数学建模预测方法进行研究。选取船舶主尺度参数,以及横剖面面积曲线和设计水线面这2类船舶型线特征参数,通过多元二次回归和RBF神经网络分析,得到对船舶性能产生重要影响的主要船型参数及其最佳组合,建立肥大型船剩余阻力系数和自航因子的数值预测模型。数值计算和试验测试结果表明,与只包含主尺度参数的预测模型相比,包含船舶型线特征参数的预测模型具有更好的准确性,可为船舶型线优化设计提供参考。Based on a set of model tests data of large fat ships,the correlation between individual ship characteristic parameter and the ship residual resistance coefficient and the self-propulsion factors is analyzed.With the multiple regression analysis and RBF(Radial Basis Function) neural network,the method for mathematical modeling and prediction of ship characteristic parameters is studied.In addition to the main ship characteristic parameters,the cross-section area curve and the design waterline plane are introduced into the prediction model so that the factors having important impact on the ship performance are rationally combined.The numerical prediction models for the residual resistance coefficient and the self-propulsion factors for large fat ships are established.Through comparison,it is found that the prediction models including the additional hull form characteristic parameters have higher accuracy and are fit for the optimization design of ship hull forms.

关 键 词：肥大型船 回归分析 径向基函数(RBF)神经网络 船型参数 预测模型 

分 类 号：U661.3[交通运输工程—船舶及航道工程]