基于遗传算法的碟型水下滑翔机结构优化  

Structural Optimization of the Round Dish-Shaped Underwater Gliders Based on the Genetic Algorithm

作　　者：甄春博[1] 刘兆瑞[2] 王天霖[1] 英扬 于鹏垚 

机构地区：[1]大连海事大学交通运输装备与海洋工程学院,辽宁大连116026 [2]大连理工大学运载工程与力学学部船舶工程学院,辽宁大连116024

出　　处：《海洋技术学报》2017年第2期10-15,共6页Journal of Ocean Technology

基　　金：国家重点研发计划重点专项资助项目(2016YFC0301500);国家自然科学基金资助项目(51379025,51609031);海洋工程国家重点实验室开放基金资助项目(1513);中央高校基本科研业务费专项资金资助项目(3132016346);辽宁省博士启动基金资助项目(201601070)

摘　　要：以结构质量和应力作为目标函数,结构变形为约束条件,设计碟型水下滑翔机结构优化流程,采用单参数和多参数敏感度分析方法完成关键结构参数的筛选。采用拉丁超立方试验设计方法完成了对设计空间的采样布点工作,利用样本点数据创建了滑翔机结构优化的Kriging代理模型,并对Kriging代理模型进行了联合训练,使模型的拟合达到非常高的可用精度。采用NSGA-2第二代非支配排序多目标遗传算法对滑翔机进行了结构优化求解,得到了优化的Pareto前沿面最优解。优化结果显示,结构质量和应力较优化前分别降低5.57%和14.91%,文中所提方法在滑翔机结构优化设计中具有可行性。Considering the structural mass and stress as the objective function, and structural deformation as constraint condition, the structural optimization design of the dish-shaped underwater glider is studied, with the analysis method of single parameter and multi-parameter sensitivity used to select the key structural parameters. The sample points in the design space are selected by using the experimental design method of Latin Hypercube Sampling. Then, the Kriging agent model of the glider structural optimization is established by using the sample point data. After the joint training of the Kriging agent model, the fitting of the model has obtained very high precision. Finally, the second generation of non-dominated sorting genetic algorithm （NSGA-2） is adopted to solve glider optimization problem, and the Pareto front surface optimal solution is obtained. The optimization results show that the structural mass and stress is reduced by 5.57% and 14.91%, respectively. The proposed method is feasible in the structural optimization design of underwater gliders.

关 键 词：水下滑翔机 结构优化 敏感度分析 Kriging代理模型 非支配排序多目标遗传算法 

分 类 号：U661.4[交通运输工程—船舶及航道工程]