Abstract:

A self-pressurization model of vacuum multi-layer insulation (VMLI) liquid hydrogen tank coupled with vapor-cooled shield (VCS) is constructed using MATLAB, and the effects of the dimensionless position l_d of VCS, the mass flowrate in VCS {\dot{m}}_{\mathrm{v}\mathrm{c}\mathrm{s}} and the opening moment of VCS t_open on the dormancy of tank are investigated. The dormancy is defined as the duration for tank pressure to rise from the initial pressure to the termination pressure, which is represented by a dimensionless dormancy extension factor {\lambda }_{\mathrm{d}}. Parameters {\lambda }_{\mathrm{c}} and t_d are defiend to represent the mass of hydrogen consumed by VCS and duration time when VCS remains opened respectively; the multi-objective genetic algorithm is introduced to help search for the optimized ({\dot{m}}_{\mathrm{v}\mathrm{c}\mathrm{s}}, t_open, t_d, l_d) in the high-dimensional input parameter space, achieving a Pareto-optimal state for {\lambda }_{\mathrm{c}} and {\lambda }_{\mathrm{d}}, and the resulting set of optimized points ({\lambda }_{\mathrm{c}}, {\lambda }_{\mathrm{d}}) forms the Pareto front. The Pareto front can be approximated as a straight line passing through the origin, with its slope determining the theoretical limit of {\lambda }_{\mathrm{d}} under fixed {\lambda }_{\mathrm{c}}, indicating the maximum dormancy achievable given a specific hydrogen consumption by VCS. When TVCS (triple vapor-cooled shields) is adopted, the three optimized points ({\dot{m}}_{\mathrm{v}\mathrm{c}\mathrm{s}}, t_open, t_d, l_d) in the high-dimensional input parameter space are compared with the initial point by evaluating their corresponding ({\lambda }_{\mathrm{c}},{\lambda }_{\mathrm{d}}). It is found that {\lambda }_{\mathrm{c}} corresponding to the optimized points 1, 2 and 3 are reduced by 5.57%, 18.60% and 31.57% respectively, while {\lambda }_{\mathrm{d}} increase by 64.57%, 42.60% and 20.25% respectively; it indicates that by optimizing the input parameters, the dormancy period is extended when the amount of hydrogen consumed by the VCS is reduced. The research can provide theoretical guidance for improving the utilization efficiency of cold energy in VCS and the efficient storage of liquid hydrogen.

Key words: dynamic simulation, mathematical modeling, optimal design, thermal insulation, liquid hydrogen tank, Pareto-optimal

摘要：

利用MATLAB构建了耦合蒸气冷却屏（VCS）的真空多层绝热液氢储罐的自增压模型，分析了VCS无量纲位置l_d、VCS中质量流量{\dot{m}}_{\mathrm{v}\mathrm{c}\mathrm{s}}和VCS开启时刻t_open对储罐休眠期的影响，休眠期是储罐从初始压力到达设定终止压力的时长，由无量纲的休眠期延长因子{\lambda }_{\mathrm{d}}表征。定义{\lambda }_{\mathrm{c}}和t_d，分别表示VCS消耗的氢气质量和VCS开启的持续时间；引入多目标遗传算法在高维输入参数空间中搜索优化的（{\dot{m}}_{\mathrm{v}\mathrm{c}\mathrm{s}}， t_open， t_d， l_d），使{\lambda }_{\mathrm{c}}和{\lambda }_{\mathrm{d}}达到帕累托最优状态，获得的一系列优化点（{\lambda }_{\mathrm{c}}，{\lambda }_{\mathrm{d}}）构成了帕累托前沿。帕累托前沿可拟合为通过原点的一条直线，其斜率决定了固定{\lambda }_{\mathrm{c}}条件下{\lambda }_{\mathrm{d}}的理论极限，表征在给定VCS消耗氢气量的情况下可以获得的理论最长休眠期。当采用三重蒸气冷却屏（TVCS）时，将高维输入参数空间中三个优化点（{\dot{m}}_{\mathrm{v}\mathrm{c}\mathrm{s}}， t_open， t_d， l_d）对应的（{\lambda }_{\mathrm{c}}， {\lambda }_{\mathrm{d}}）与初始点对应的（{\lambda }_{\mathrm{c}}， {\lambda }_{\mathrm{d}}）进行对比，结果表明：优化点1、2和3对应的{\lambda }_{\mathrm{c}}分别减少了5.57%、18.60%和31.57%，而{\lambda }_{\mathrm{d}}分别提高了64.57%、42.60%和20.25%；说明通过优化输入参数，在VCS消耗氢气量减少的情况下休眠期反而延长了。研究可为VCS中冷能利用效率的提升和液氢高效储存提供理论指导。

关键词: 动态仿真, 数学模拟, 优化设计, 绝热, 液氢储罐, 帕累托最优