An efficient dimensionality reduction approach for modelling cryogenic hydrogen release

Abstract

This study develops a precise and cost-effective method for reducing the complexity of cryogenic hydrogen unintended release. For this purpose, the dominant modes of a well-characterised cryogenic hydrogen jet (Sandia National Laboratories), with stagnation pressure and temperature fixed at 5 bar and 50 K, are extracted using the proper orthogonal decomposition (POD) method. Building on validating the large eddy simulation (LES) against experimental measurements, the velocity field within the LES results is analysed near the release point. The mass fraction half-width of cryogenic hydrogen jets grows linearly with a slope of 0.1115 against the downstream distances, resembling room-temperature jets. The first 10 extracted modes from 500 LES snapshots demonstrate dominance over the remaining modes. In addition, this study evaluates the flammable gas dispersion using LES and 10-mode POD, showing POD effectively identifies high-risk areas within hydrogen's flammability range. Minor deviations (2.1–3.3%) confirm POD's reliability for hazard assessment compared with LES. Furthermore, an in-depth analysis is performed on the radial distributions of hydrogen mass fraction and centreline hydrogen mole fraction.