基于观察窗蠕变特性的耐压球壳优化

Optimization design for the spherical pressure shell of manned submersible with consideration of creep of window

针对载人潜器耐压球壳优化问题,本文设计了基于观察窗蠕变特性的耐压球壳参数化分析流程。采用最优拉丁超立方方法进行了样本点的选取与分析,构建了观察窗有机玻璃本构模型,建立了基于径向基神经网络耐压球壳近似模型。以耐压球壳质量为目标函数,采用多岛遗传算法进行了单目标优化求解。以观察窗蠕变位移、耐压球壳质量为目标函数,采用第二代非支配排序遗传算法进行了多目标优化求解。结果表明:单一质量优化效果不明显,而观察窗蠕变变形明显增大,多目标优化方案质量最大增幅2.07%,观察窗变形最大降低-18.09%,说明在耐压球壳设计的过程中考虑观察窗蠕变效应,可提高载人潜器的安全性。

Aiming at the optimization of the spherical pressure shell of manned submersible,this study designs a parametric analysis process of spherical pressure shell on the basis of the creep characteristics of the observation window.The constitutive model of PMMA for the creep characteristics of the observation window is established.The optimal Latin hypercube method is used to select and analyze sample points,and the pressure spherical shell approximation model based on the radial basis function neural network is established.Taking the spherical pressure shell mass as the objective function,the multi-island genetic algorithm is used to solve the single-objective optimization problem.Taking the creep displacement of the observation window and the mass of the spherical pressure shell as objective functions,the second-generation non-dominated sorting genetic algorithm is used to solve the multi-objective optimization problem.Results reveal that the effect of single mass optimization is not obvious,but the creep deformation of the observation window increases significantly.The maximum increase of the multi-objective optimization scheme is 2.07%,and the maximum decrease of observation window deformation is-18.09%.It shows that the safety of manned submersible can be improved by considering the creep effect of observation window on the design process of pressure spherical shell.