To address the issue of position drift caused by water flow impact during bottom monitoring in high-flow areas by underwater monitoring robots, a method combining parametric modeling, fluid simulation, and multi-objective optimization algorithms was proposed for improving the external structure. Initially, a simplified parametric model of the robot was constructed, followed by a sensitivity analysis of the fluid domain conducted on the computational fluid dynamics platform. Optimal Latin hypercube sampling and a second-order response surface model were applied to fit the design variables and assess their sensitivity. Finally, a multi-objective optimization model was established based on the non-dominated sorting genetic algorithm Ⅱ (NSGA-Ⅱ), targeting minimal drag and minimal lift. After 20 generations of evolutionary iteration, the Pareto optimal solution set was obtained,and NSGA-Ⅱ algorithm was compared with other algorithms to verify its superiority in multi-objective optimization. The optimization results show a 6.3% reduction in drag, a 7.1% decrease in lift, and an 18.8% reduction in volume compared to the initial structure. Additionally, the corner rounding treatment eliminates stress concentration, and the design demonstrates superior stability under different flow velocities and turbulence intensities. The effectiveness of the optimization method is validated, providing theoretically and technically extensible support for the structural design of mission-oriented underwater robots.
DU Hai-yang, ZANG Hong-bin, LI Jia-long, WANG Hao-ran, ZHANG Hong.
Multi-objective Structural Optimization Design of Underwater Monitoring Robots Based on NSGA-Ⅱ Algorithm[J].
Science Technology and Engineering, 2026, 26(19): 7978-7988 DOI:10.12404/j.issn.1671-1815.2507398
闫兴亚等[2]针对自主水下航行器(autonomous underwater vehicle,AUV)耐压壳体的多目标优化设计需求,通过折减因子法构建综合应力裕度与临界载荷裕度模型,以双裕度和质量为优化目标,采用非支配排序遗传算法Ⅱ(non-dominated sorting genetic algorithm Ⅱ,NSGA-Ⅱ)对耐压及起吊工况进行寻优,并结合博弈组合赋权-逼近理想解排序(technique for order preference by similarity to ideal solution,TOPSIS)法筛选最优尺寸方案。Chen等[3]为实现带附体流线型水下航行器的体积最大化与阻力最小化,提出基于自适应采样算法的元模型设计优化框架,通过计算流体动力学(computational fluid dynamics,CFD)仿真与Myring(克里金模型)模型结合带拥挤距离的多目标粒子群优化算法求解多目标优化问题。苗怡然等[4]将参数化设计与稳健性理念引入水下航行器外形设计,通过最优拉丁超立方法选样、响应面模型拟合构建艇体阻力近似模型,采用非支配排序遗传算法Ⅲ(NSGA-Ⅲ)完成确定性多目标优化,并考虑制造误差开展稳健性优化,有效提升了设计方案的稳健性与工程实用性;不过该研究存在流体域无关性验证不足、优化算法参数未作敏感性分析、优化方案工程可行性论证不充分等问题。王妹婷等[5]通过Fluent对比4种回转体线型的直航阻力,选定Myring线型并结合型值点约束,利用Visual C++确定首部锐度因子n=2、尾部离去角θ=25°,优化水下机器人回转体外形以降低阻力、提高能源利用率;但该方法仅聚焦直航阻力单一目标,未考虑升力对机器人坐底位置稳定性的影响,高流速下易因升力波动导致偏移。周宏根等[6]通过三维扫描获取河鲀外形点云数据,基于非均匀有理 B 样条 (nonuniform rational bspline,NURBS)曲面建模技术将河鲀背部曲面划分为主区和侧区,构建了精确的几何与数学模型;然而该建模过程依赖高精度点云数据,且流体仿真与优化算法脱节,计算成本较高。唐军等[7]为提升水下机器人阻力性能与设计效率,基于Myring线型构建含10个设计变量的参数化模型,通过Fluent计算直航阻力,采用最优拉丁超立方设计采样并建立2阶响应面模型,进而利用NSGA-Ⅱ算法开展多目标优化;但研究仅考虑10个设计变量对直航阻力的影响,未验证流体域对仿真精度的作用,且忽略多目标协同,NSGA-Ⅱ算法生成的Pareto解集实用性较弱。唐元贵等[8]针对“海斗号”全海深自主遥控水下机器人(autonomous remotely-controlled vehicle,ARV)的深渊探测需求,在分析使命导向型典型性能指标的基础上,提出基于模糊层次分析法构建总体设计参数化模型;通过模型参数配置与优化,实现以最短时间成本完成深渊探测的目标;不过针对“海斗号”的参数化设计聚焦深渊探测的“时间成本优化”,未涉及高流速下的低阻低升设计,场景适配性有限。
在水下监测机器人外观仿真研究中,阻力与升力等特性往往依赖于多维设计变量的变化。为了在有限计算资源下获得具有代表性的样本点集,采用最优拉丁超立方设计(optimal Latin hypercube design, OLHD)进行采样。该方法通过在输入变量的全局设计空间中均匀布点,并结合最大最小距离或熵准则优化样本分布,确保样本在高维空间中具有良好的覆盖性与差异性。相比于单纯的随机采样,OLHD能够显著减少样本聚集现象,提升阻力、升力数据在多维输入空间中的代表性和可解析性[10]。体积V、阻力F1和升力F2试验设计样本点及响应值的各项数据如表2所示。
YanXingya, TangYuangui, LiangHui, et al. Multi-objective optimization design of AUV ring-stiffened cylindrical shell based on comprehensive margin[J]. Machine Design and Research, 2024, 40(5): 30-35.
[4]
ChenS P, LiuF. Multi-objective shape optimization of underwater vehicles based on an adaptive sampling algorithm[J]. Ocean Engineering, 2024, 146: 103950.
MiaoYiran, GaoLiangtian, LiuFeng, et al. Robust optimization design of underwater vehicle shape based on parameterization[J]. Journal of Harbin Engineering University, 2018, 39(4): 622-628.
ZhouHonggen, CuiJie, TianGuizhong, et al. Surface fitting and mathematical modeling of puffer fish back morphology[J]. Journal of Jilin University (Engineering and Technology Edition), 2020, 50(3): 1131-1137.
TangJun, LiChuanqing, DangZhaokai, et al. Structural optimization design of underwater robot based on NSGA-Ⅱ algorithm[J]. Journal of Machine Design, 2025, 42(4): 56-62.
TangYuangui, WangJian, LuYang, et al. Parametric design method and experimental research on haidou full-depth ocean autonomous and remotely-operated vehicle[J]. Robot, 2019, 41(6): 697-705.
YangRuigang, GaoLin, DongQing, et al. Structural optimization design of stacker based on kriging model and NSGA-Ⅱ algorithm[J]. Machinery Design & Manufacture, 2022(5): 298-304.
WeiHongmei. Optimization of spider web-shaped microchannel structure based on multi-objective genetic algorithm[J]. Science Technology and Engineering, 2025, 25(9): 3698-3703.
GongLei, QinYongjun, TianYuan, et al. Optimization and performance study of desert sand concrete mix proportion based on response surface and NSGA-Ⅱ algorithm[J]. Science Technology and Engineering, 2025, 25(13): 5571-5578.
ZhaoShuen, YangMingsen, PengGuangxu. Lightweight optimization design of body-in-white based on NSGA-II mixed sensitivity analysis[J]. Journal of Mechanical Strength, 2019, 41(4): 887-894.
[25]
ChenC S, HeM W. A many-objective evolutionary algorithm with pareto front estimation and angle-based selection[C]// 2022 IEEE 13th International Conference on Software Engineering and Service Science (ICSESS). Beijing: IEEE, 2022: 276-279.
HuangChaozhi, LiSiying, LiuXiaobo, et al. Multi-objective optimization method for permanent magnet-assisted synchronous reluctance motor based on KELM-NSGA-Ⅱ[J]. Science Technology and Engineering, 2025, 25(3): 1065-1074.
ZhongQingwei, TangHaoming, YuYingxue, et al. Multi-objective optimization of aircraft taxiing path based on improved NSGA-Ⅱ algorithm[J]. Science Technology and Engineering, 2025, 25(20): 8737-8744.
TaoZhicong, WuJuncen, MengXianda, et al. Multi-objective optimization design and analysis of pressurehull based on NSGA-Ⅱ algorithm[J]. Ship & Boat, 2025, 36(4): 66-74.
[34]
ClarkA J, WangJ X, TanX B, et al. Balancing performance and efficiency in a robotic fish with evolutionary multiobjective optimization[C]// 2014 IEEE International Conference on Evolvable Systems. Orlando: IEEE, 2014: 227-234.