﻿ 基于BP网络的球鼻首参数化优化
 舰船科学技术  2021, Vol. 43 Issue (3): 37-40    DOI: 10.3404/j.issn.1672-7649.2021.03.008 PDF

1. 高速水动力航空科学技术重点实验室，湖北 荆门 448035;
2. 中国特种飞行器研究所，湖北 荆门448035

Research on parametric optimization of bulbous bowbased on BP network
GAO Xian-jiao1,2, WANG Li-li1,2, SUN Gang1,2, HAN Xiao-hong1,2, QU Ru-jun1,2
1. Key Aviation Scientific and Technological Laboratory of High Speed Hydrodynamic, Jingmen 448035, China;
2. AVIC Special Vehicle Research Institute, Jingmen 448035, China
Abstract: This paper selects a container ship as research target, and optimization work of bulbous bow when the ship slows down its speed is carried out. Main work of this paper can be summarized as follows. Firstly, three-dimensional ship model is established using the software of Catia, and in order to produce different bulbous bow shapes, four characteristic parameters is selected to describe its basic structure. Secondly, twelve groups of bulbous bow is obtained using Latin hypercube sampling method. Then, the BP network with high nonlinear fitting capability is adopted to establish the relation between characteristic parameters of bulbous bow and resistance coefficient. Finally, a genetic algorithm is used to find the optimal solution of the network, the result shows that resistance coefficient of the optimal individual has been decreased notably, which indicates that this method can be used to optimize ship's bulbous bow.
Key words: bulbous bow     characteristic parameters     latin hypercube sampling     BP network     genetic algorithm
0 引　言

1 三维建模

2 球鼻首参数化

 图 1 球鼻首特征参数 Fig. 1 Characteristic parameters of bulbous bow

 $Error = \dfrac{{{C_{ti}} - {C_{t0}}}}{{{C_{t0}}}} \times 100\%$

3 球鼻首优化分析 3.1 BP网络构建

BP网络又称误差反向传播的神经网络，它具有较强的非线性映射能力和灵活的网络结构，仅凭数据本身就可以建立输入和输出之间的关系模型[5]。输入变量经过输入层、隐藏层向前传至输出层；网络对输出结果进行判断，当输出值和期望值有较大偏差时，误差反向传递，网络依据相关规则调整节点的权值系数和阈值，正向传递和反向传递交叉反复进行，直到获得较为满意的输出，训练到此结束。

 图 2 BP网络拓补结构 Fig. 2 Topological structure of BP network

3.2 球鼻首形状参数极值寻优

 图 3 球鼻首优化流程 Fig. 3 Optimization process of bulbous bow

 图 4 最优个体适应度收敛曲线 Fig. 4 Convergence curve of optimal individual fitness
3.3 优化结果分析

 图 5 船体动压力对比云图 Fig. 5 Comparison of hull dynamic pressure

 图 6 波形对比图 Fig. 6 Comparison of wave profile

4 结　语

 [1] 许欢, 刘伟, 徐梦洁. 船舶减速航行的现状、减排效果及产生的问题[J]. 交通企业管理, 2013, 28(9): 46-47. DOI:10.3963/j.issn.1006-8864.2013.9.020 [2] 贾瑞. 大型油船艏部线型优化[D]. 大连: 大连海事大学, 2014. [3] 邓贤辉, 方昭昭, 赵丙乾. 基于计算流体动力学的最小阻力船型自动优化[J]. 中国舰船研究, 2015, 10(3): 19-25. DOI:10.3969/j.issn.1673-3185.2015.03.004 [4] MATULJA D, DEJHALLA R. Genetic algorithm optimization of a ship's bulbous bow[A]. Annals of DAAAM for 2011 & Proceedings of the 22nd International DAAAM Symposium, Vienna, Austria, EU, 2011, 22(1): 15−17. [5] HECHT-NIELSEN R. Theory of the back-propagation neural network[J]. International Joint Conference on Neural Networks, 1989, 1(1): 593-605.