Coupled characteristics and operational optimization of a high-pressure hydrogen cascade fast refueling system

Abstract

To reduce energy consumption while maintaining refueling efficiency and thermal safety, this study develops a simulation-driven prediction and optimization framework for a high-pressure hydrogen cascade fast refueling system. A coupled flow-thermal-solid numerical model is established for an on-board hydrogen storage cylinder, and a convolutional neural network surrogate is used to predict filling time, final temperature, and total system energy consumption. A linearly decreasing weight particle swarm optimization algorithm is then coupled with the surrogate model to search for a low-energy refueling strategy under filling time and temperature constraints. The results show that higher initial pressure shortens filling time and lowers final temperature but increases total energy consumption, whereas higher switching pressures reduce energy consumption at the expense of longer filling time. Lower pre-cooling temperature suppresses temperature rise but increases cooling demand. The optimized strategy reduces total system energy consumption by 3.71%, providing an efficient tool for low-energy operation.