Numerical modeling accidental release of cryogenic liquid : application to liquid hydrogen (LH2) risk assessment

In the face of energy shortages and increasing pollution, many countries are developing policies to replace fossil fuels with clean alternative energy sources. Storage and transportation are crucial in the hydrogen sector. Compression and liquefaction are common, with LH2 being preferred for medium to long-distance transport due to its high density. However, leaks can lead to fires and explosions. Hundreds of accidents, some of them severe, have been recorded. To study LH2 behavior during accidental leaks and enhance safety, this thesis employs the open-source CFD code OpenFOAM to model the release process and perform detailed analyses.The thesis reviews nearly all studies on LH2 release—experimental and modeling—and traces its research history. A bibliometric search in the Web of Science Core Collection using keywords like "liquid hydrogen" + "leakage" and "liquid hydrogen" + "release" found 571 articles, with 346 selected after manual screening. In addition, some meaningful analyses were conducted. This macro-micro analysis offers comprehensive insights.A model using OpenFOAM was developed and validated against a hydrogen jet experiment. Based on the PIMPLE algorithm, it effectively simulates hydrogen-air mixing. The complete model was built on this model.The review of previous LH2 studies showed that nearly all numerical models were validated using NASA experiments, and our model will be similarly validated. We first review NASA's experiments and extract the typical physical phenomena of LH2 release. By simplifying the process, we identify the core thermodynamic and fluid dynamic processes, then introduce the necessary numerical models and use the PIMPLE algorithm to build a predictive model. Comparison with NASA's data demonstrates the good agreement, with deviations within an acceptable range, confirming the model's validity.Subsequent analysis examined the effects of wind speed, ground type, pressure, temperature, and release duration on LH2 release by comparing temporal changes in hydrogen concentration, maximum vertical and downwind spread, and cloud volume. Results show that higher wind speed increases downwind diffusion and hinders cloud rise, raising downstream risks. Ground roughness affects the wind profile, with greater roughness yielding a less steep profile and more pronounced wind speed differences. Higher atmospheric pressure inhibits hydrogen diffusion and reduces cloud volume. Temperature has little effect. Longer release duration enlarges the affected area. Wind speed and release duration have the most significant impact, so special attention is needed in windy conditions.An air wall method is proposed to enhance LH2 safety. Compared with bund wall, the air wall demonstrates a strong protective effect, and its mechanism of action is explained. Its protection under different airflow velocities was investigated, showing that higher velocities improve its effect. The timing of activation is crucial—it can still protect if activated after a leak, but earlier activation yields better results. Faster airflow speeds up the air wall's establishment, and a larger air wall offers better protection. Additionally, stronger airflow is required under higher wind conditions.