基于多保真深度神经网络的船型优化  

Hull form optimization based on multi-fidelity deep neural network

作　　者：魏亚博 汪杨骏 万德成[1,2] WEI Yabo;WANG Yangjun;WAN Decheng(Computational Marine Hydrodynamics Laboratory,Shanghai Jiao Tong University,Shanghai 200240,China;School of Naval Architecture,Ocean and Civil Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;College of Advanced Interdisciplinary Studies,National University of Defense Technology,Nanjing 210000,China)

机构地区：[1]上海交通大学船海计算水动力学研究中心,上海200240 [2]上海交通大学船舶海洋与建筑工程学院,上海200240 [3]国防科技大学前沿交叉学科学院,江苏南京210000

出　　处：《中国舰船研究》2024年第6期74-81,共8页Chinese Journal of Ship Research

基　　金：国家自然科学基金资助项目(52131102)。

摘　　要：[目的]为了提高优化效率并获得更好的优化结果,将不同精度数据进行有机融合,利用多保真深度神经网络开展船型优化设计。[方法]基于多源数据融合和迁移学习思想,构建了一种多保真深度神经网络。通过将大量低保真数据与少量高保真数据融合学习,构建与高保真数据之间的线性项和非线性项,得到高保真近似模型。基于此方法开展针对DTMB 5415船静水阻力的优化设计。分别采用势流和黏流样本点阻力进行评估,势流计算结果作为低保真数据,黏流计算结果作为高保真数据,构建多保真深度神经网络近似模型。借助遗传算法获得优化解并与只使用单一高保真数据构建的Kriging近似模型的优化结果进行对比。[结果]基于多保真神经网络方法,DTMB 5415阻力减少了6.73%。基于Kriging模型,DTMB 5415阻力减少了5.59%。[结论]多保真深度神经网络近似模型可以兼顾效率和精度,可以用于优化求解,且由其得到的优化船型阻力优化效果更为显著。[Objective] To improve hull optimization design efficiency and obtain better optimization results,different fidelity data is organically integrated and a multi-fidelity deep neural network is applied.[Methods] A multi-fidelity deep neural network is constructed based on the idea of multi-source data fusion and transfer learning.By fusing a large amount of low-fidelity data with a small amount of high-fidelity data,the linear and nonlinear terms between the high-fidelity data are constructed to obtain a high-fidelity surrogate model.Based on this method,the optimization design of the resistance of a DTMB 5415 ship is carried out.The potential flow and viscous flow are used to evaluate the resistance of the sample points respectively.The potential flow calculation results are used as low-fidelity data,while the viscous flow calculation results are used as high-fidelity data.A multi-fidelity deep neural network surrogate model is then constructed.The optimal solution is obtained by genetic algorithm and compared with the optimal solution of the Kriging model constructed by high-fidelity data.[Results] Based on the multi-fidelity deep neural network method,the resistance of DTMB 5415 is reduced by 6.73%.Based on the Kriging model,the resistance of DTMB 5415 is reduced by 5.59%.[Conclusions] The multi-fidelity deep neural network surrogate model can take into account both efficiency and accuracy,which can be used for optimization.The optimized hull form obtained by it has a more significant resistance optimization effect.

关 键 词：船舶设计 人工智能 减阻 船型优化 多保真深度神经网络 数据融合 迁移学习 

分 类 号：U662.2[交通运输工程—船舶及航道工程]