An Integrated Layer-Resolved Heat Transfer and Thermodynamic Model for Boil-Off Gas Prediction in LH2 Storage Tanks

Abstract

Liquid hydrogen (LH₂) is a key energy carrier for decarbonization due to its high gravimetric energy density, which makes it an attractive option for applications ranging from large-scale energy storage to transport and aerospace systems. However, the storage of LH₂ in cryogenic vacuum-insulated vessels introduce challenges associated with heat ingress from surrounding as hydrogen has a very low boiling point at 20 K (1 atm), even small heat leaks can result in continuous mass through boil-off. Quantifying this loss is essential to control the large-scale LH₂ storage economic viability. To address the limitations in conventional approaches that typically rely on constant heat transfer rates or decouple the insulation design from boil-off dynamics, this study presents a layer-resolved heat transfer model integrated with a two-phase boil-off thermodynamic model. The framework considers a horizontal cylindrical tank with hemispherical ends and incorporates an economic model that evaluates the storage-related levelized cost of hydrogen (LCOH), which dynamically uses boil-off predictions to quantify losses in overall revenue due to boil-off. At nominal conditions, LCOH varies from $6.19 to $6.88/kg depending on the hydrogen price. This study also incorporates a comprehensive parametric analysis across seven design and operational variables, including total MLI layers, vacuum pressure, fill level, ambient temperature, cylindrical length, relief pressure, and tank’s wall thickness, followed by a global sensitivity analysis based on normalized swing (NS). Vacuum pressure emerges as the single most influential variable with an NS index of 12.76 on total BOG fraction, increasing it from 10-50% over a 90-days storage run as the vacuum state degrades from a newly commissioned tank to a severe vacuum failure (0.46 Pa). As a result, fill level strongly affects the BOG dynamics through liquid-vapor space heat partitioning, with diminishing-return design point fill level ~75%. The total MLI layer count is the third most influential variable with an diminishing-return design point design point near 60-70 MLI layers, where diminishing returns intersect increasing capital cost. The model confirms that vacuum integrity is both the main thermodynamic as well as economic lever in LH₂ storage design.