Insulation and cost optimization of vapor-cooled shield coupled with para-to-ortho hydrogen conversion based on self-pressurization model of liquid hydrogen tank and NSGA-II

The temperature difference between the inside and the outside of liquid hydrogen storage tank is very large, so MLI (variable-density vacuum multi-layer insulation) coupled with VCS (vapor cooled-shield) is generally introduced to shield external HL (heat leakage). In this research, a self-pressurization model of liquid hydrogen tank is constructed, by considering the effect of MLI integrated with VCS. In addition, VCS tube is filled with catalyst, and it is coupled with the catalytic process of para-to-ortho hydrogen conversion (POC), which can release conversion cold energy. Based on the model, the effect of VCS dimensionless position (DP) in MLI, mass flow rate (ṁ_vcs) in VCS and opening moment (t_open) of VCS on the dormancy of tank is investigated. These three parameters and duration time(t_d) of VCS all affect the dormancy of tank obviously, so we innovatively introduce a multi-objective optimization algorithm, known as the 2nd generation non-dominated sorting genetic algorithm (NSGA-II), to facilitate the search for the optimized (VCS DP, ṁ_vcs, t_open, t_d), that can bring η_c and η_d into pareto optimal state, representing the trade-off between η_c and η_d. η_c is the ratio of the total mass of gas hydrogen consumed by VCS to the initial hydrogen mass within tank, and η_d is the relative increase of the dormancy of tank when VCS is opened compared to that without VCS. The results show that, under condition of VCS-POC, compared with the original design of VCS, η_c of optimized designs 1, 2, 3 respectively decrease by 8.50 %, 18.35 % and 29.53 %, while η_d respectively increase by 58.23 %, 43.10 % and 23.68 %. The optimized distribution of (t_open, VCS DP) is more concentrated than (ṁ_vcs, t_d), and t_open is mainly concentrated in the first 9 days with or without POC. With VCS alone, the optimized VCS DP is mostly distributed in the range of 0.489 to 0.533, while it is mostly distributed in the range of 0.378–0.422, with VCS-POC.