摘要
船型优化可以减少船舶航行过程中的阻力,是提高船舶快速性的主要途径.在现代船体型线优化设计过程中,通常反复使用计算流体力学(CFD)软件进行仿真计算,船舶模拟模型船型样本多、计算量大,使优化的时间成本大幅提高,引入变精度模型将有效解决此问题.以油船球艏优化为例构建了球艏降阻优化BP神经网络(BPNN)模型,可作为低精度模型在优化迭代过程中对大量设计点进行快速阻力预报,通过变量的相关分析,预报出总阻力的变化趋势,寻求逼近最优解的设计点,并为下一步在基于变精度模型的球艏降阻优化研究中神经网络的应用提供经验与支持.
Ship form optimization can reduce the resistance of the ships,which is the main way to improve the rapidity of ships.In the process of modern ships form optimization design,the software of computational fluid dynamics(CFD)is often used repeatedly for simulation calculation.Due to the large number of samples and large amount of calculation,the time cost of optimization is greatly increased.If the variable precision model is introduced,the problem of time cost will be solved effectively.Taking an oil tanker as sample,a BP neural network(BPNN)model for bulbous bow resistance optimization is established,which can be used as a low fidelity model to predict the resistance to a large number of design points during optimization iteration.Through the correlation analysis of variables,the variation trend of the total resistance is predicted,and the design point approaching the optimal solution is determined,which can provide experience and support for the application of neural network in bulbous bow resistance optimization based on the variable fidelity model.
作者
张维英
周俊秋
于博文
于洋
ZHANG Weiying;ZHOU Junqiu;YU Bowen;YU Yang(School of Navigation and Naval Architecture, Dalian Ocean University, Dalian 116023, China)
出处
《大连理工大学学报》
EI
CAS
CSCD
北大核心
2021年第2期160-171,共12页
Journal of Dalian University of Technology
基金
国家自然科学基金青年基金资助项目(51509124).
