基于改进NSGA-Ⅱ的载人潜水器多开孔耐压结构优化

为提高载人潜水器耐压结构性能和设计效率，开展优化模型和优化算法的改进研究。进行耐压结构有限元模型的网格无关性分析，利用iSight软件对ABAQUS软件进行集成实现耐压结构的参数化分析，进行设计变量的灵敏度分析和降维处理以降低问题分析难度。利用最优拉丁超立方方法进行样本点的选取，进行样本点的参数化分析。基于典型近似模型对样本点进行拟合，选择拟合精度最高、满足工程需要的近似模型作为耐压结构的替代模型，并建立耐压结构多目标优化模型。结合正态分布交叉算子对第二代非支配排序遗传算法（NSGA-Ⅱ）进行改进，得到NSGA-Ⅱ的改进算法INSGA-Ⅱ，以改善NSGA-Ⅱ局部收敛等问题，在利用测试函数进行INSGA-Ⅱ的有效性验证的基础上，利用INSGA-Ⅱ进行耐压结构多目标优化求解。研究结果表明：参数化分析方法和高精度近似模型可平衡计算精度和设计效率之间的矛盾，在测试对比中，INSGA-Ⅱ所得到的Pareto解集在贴近性、均匀性和收敛性等方面均优于NSGA-Ⅱ，INSGA-Ⅱ得到的耐压结构多目标优化方案较NSGA-Ⅱ更优。

Optimization Design for Multiple Openings Pressure Structure of Manned Submersible Based on an Improved NSGA-Ⅱ Algorithm

To improve the performance and design efficiency of manned submersible pressure structure，the improvement of optimization model and optimization algorithm is carried out. The mesh-independent analysis of the finite element model of the pressure structure is carried out，the parametric analysis of the pressure structure is realized by integrating ABAQUS with iSight software，the sensitivity analysis of the design variables and the dimensionality reduction process are carried out to reduce the difficulty of the problem analysis. the sample points are selected by using the optimal Latin hypercube method，and the parametric analysis of the sample points is carried out. Based on the typical approximation model to fit the sample points，the approximation model with the highest fitting accuracy and meeting the engineering needs is selected as the alternative model for the pressure structure，and a multi-objective optimization model for the pressure structure is established. The improved algorithm INSGA-Ⅱ of NSGA-Ⅱ is obtained by combining the normal distribution crossover operator to improve the second generation non-dominated sorting genetic algorithm (NSGA-Ⅱ) to improve the problems such as local convergence of NSGA-Ⅱ, and the multi-objective optimization solution of the pressure structure is carried out using INSGA-Ⅱ on the basis of validation of the effectiveness of INSGA-Ⅱ by using test functions. The results show that the parametric analysis method and the high-precision approximation model can balance the conflicts between computational accuracy and design efficiency. In the test comparison，the Pareto solution set obtained by INSGA-Ⅱ is better than NSGA-Ⅱ in terms of closeness，uniformity and convergence，and the multi-objective optimization scheme for pressure-resistant structures obtained by INSGA-Ⅱ is better than NSGA-Ⅱ.