Genetic algorithm-based optimization method for the large-scale hydrogen liquefaction system

Liquid hydrogen (LH 2 ) presents promising prospects for large-scale storage and transportation applications. However, its widespread adoption is limited by the high energy demand of hydrogen liquefaction. Consequently, effective optimization of hydrogen liquefiers is essential for process design. In this study, a hybrid genetic algorithm-simulated annealing (GA-SA) algorithm is proposed that integrates a GA with a SA algorithm to achieve high-efficiency optimization of hydrogen liquefiers. By coupling with the SA, the robustness and search capability of the proposed algorithm are significantly enhanced compared to those of the standard GA. Based on the thermodynamic model of the proposed hydrogen liquefier, parallel and serial GA-SA algorithms are employed to optimize the specific energy consumption, considering the liquid nitrogen consumption (SEC LN 2 ). The serial GA-SA achieves a 74% reduction in computational time. The parallel GA-SA demonstrates superior optimization capability with an optimal SEC LN 2 of 10.54 kWh⋅kg LH 2 −1 and an exergy efficiency (EXE) of 48.2%. Economic and environmental analyses show that the optimized system reduces the total annualized cost by 8.8% to $1.52 million, with annual CO 2 savings of 1,497 tonnes. Furthermore, a novel hydrogen liquefier design incorporating a liquid expander for the final depressurization process is proposed, which improves the liquefaction rate from 94.9% to 99.2% and further reduces SEC LN 2 to 10.09 kWh⋅kg LH 2 −1 . This study also provides a comprehensive comparison of the different algorithms applied to hydrogen liquefier optimization.