Safety

If Hydrogen is expected to be highly valuable some improvements should be conducted mainly regarding the storage safety. To prevent from high pressure hydrogen composite tanks bursting the comprehension of the thermo-mechanics phenomena in the case of fire should be improved. To understand the kinetic of strength loss the heat flux produced by fire of various intensities should be assessed. This is the objective of this real scale experimental campaign which will allow studying in future works the strength loss of composite high-pressure vessels in similar fire conditions to the ones determined in this study. Fire calibration tests were performed on metallic cylinder vessels. These tests with metallic cylinders are critical in the characterization of the thermal load of various fire sources (pool fire propane gas fire hydrogen gas fire) so as to evaluate differences related to different thermal load. Radiant panels were also used as thermal source for reference of pure radiation heat transfer. The retained thermal load might be representative of accidental situations in worst case scenarios and relevant for a standardized testing protocol. The tests performed show that hydrogen gas fires and heptane pool fire allow reaching the target in terms of absorbed energy regarding the results of risk analysis performed previously. Other considerations can be taken into account that will led to retain an hydrogen gas fire for further works. Firstly hydrogen gas fire is the more realistic scenario: Hydrogen is the combustible that we every time find near an hydrogen storage. Secondly as one of the objectives of the project is to make recommendations for standardization issues it's important to note that gas fires are not too complex to calibrate control and reproduce. Finally due to previous considerations Hydrogen gas fire will be retained for thermal load of composite cylinders in future works.

Fast response and high durability hydrogen sensor is required in the safety management of hydrogen station and fuel cell vehicle. We had developed the catalytic combustion type hydrogen sensor in the shape of the miniature beads. It is using the optimized Pd-Pt/Al2O3 catalyst and the Pt micro-heater coil. Both warm-up time and response time of this sensor achieved less than 1 second by downsizing the element to 200μm diameter. Furthermore we improved the resistance of sensor poisoning to silicone vapor and confirmed long term stability within +/-10% of output error up to 8 years. Therefore we assume that our sensor technology contribute to hydrogen safety.

Reliable predictive tools for hydrogen safety engineering are needed to meet increased and more widespread use of hydrogen in the society. Industrial models and methods used to establish thermal radiation hazard safety distances from hydrogen jet fires are often based on models previously developed for hydrocarbon jet fires. Their capability of predicting radiative heat fluxes from hydrogen jet fires has often only been validated against small-scale or medium-scale jet flame experiments. However large-scale hydrogen jet fire experiments have shown that thermal radiation levels can be significantly higher than one might expect from extrapolation of experience on smaller hydrogen flames. Here two large-scale horizontal hydrogen jet fires (from a 20.9 mm and a 52.5 mm diameter release respectively) have been modelled and simulated with the advanced industrial CFD code KAMELEON FIREEX KFX® based on the Eddy Dissipation Concept by Magnussen for turbulent combustion modelling. The modelling of the high-pressure hydrogen gas releases is based on a pseudo-source concept using real-gas thermodynamic data for hydrogen. The discrete transport method of Lockwood and Shah is used to calculate the radiative heat transfer and radiative properties of water vapour are modelled according to Leckner. The predicted thermal radiation is compared to data from large-scale hydrogen jet fire experiments and discussed. This work was conducted as part of a KFX-H2 R&D project supported by the Research Council of Norway.

This paper examines the helium dispersion behaviour in a 16.6 m3 enclosure with a small opening in the floor and distributed leaks along the edges. Helium a simulant for hydrogen was injected near the center of the floor with an injection rate ranging from 2 to 50 standard liters per minute (Richardson number of 0.3–134) through an upward-facing nozzle. In a short-term transient the helium distribution predicted with the models of Baines & Turner (1969) and Worster & Huppert (1983) matched the measured distributions reasonably well. In a long-term transient the vertical helium profile always reached a steady state which consisted of a homogenous layer at the top overlaying a stratified layer at the bottom. The helium transients in the uniform layer predicted with the models of Lowesmith (2009) and Prasad & Yang (2010) assuming a vent was located in the ceiling were in good agreement with the measured transients.

The possibility of flame acceleration (FA) and deflagration-to-detonation transition (DDT) when homogeneous hydrogen-carbon monoxide-air (H2-CO-air) mixtures are used rises the need for an efficient simulation approach for safety assessment. In this study a modelling approach for H2-CO-air flames incorporating deflagration and detonation within one framework is presented. It extends the previous work on H2-air mixtures. The deflagration is simulated by means of the turbulent flame speed closure model incorporating a quenching term. Since high flow velocities e.g. the characteristic speed of sound of the combustion products are reached during FA the flow passing obstacles generates turbulence at high enough levels to partially quench the flame. Partial flame quenching has the potential to stall the onset of detonation. An altered formulation for quenching is introduced to the modelling approach to better account for the combustion characteristics for accelerating lean H2-CO-air flames. The presented numerical approach is validated with experimental flame velocity data of the small-scale GraVent test rig [1] with homogeneous fuel contents of 22.5 and 25.0 vol-% and fuel compositions of 75/25 and 50/50 vol-% H2/CO respectively. The impact of the quenching term is further discussed on simulations of the FZK-7.2m test rig [2] whose obstacle spacing is smaller than the spacing in the GraVent test rig.

The modelling of Rayleigh–Taylor instability during premixed combustion scenarios is presented. Experimental data obtained from experiments undertaken by FM Global using their large-scale vented deflagration chamber was used to develop the modelling approach. Rayleigh–Taylor instability is introduced as an additional time-dependent combustion enhancing mechanism. It is demonstrated that prior to the addition of this mechanism the LES deflagration model under-predicted the experimental pressure transients. It is confirmed that the instability plays a significant role throughout the coherent deflagration process. The addition of the mechanism led to the model more closely replicating the pressure peak associated with the external deflagration.

We prepared several gas barrier resins based on amorphous PVA derivative that has the T1C (13C spin-lattice relaxation time) of a long time component in amorphous phase. We confirmed it was important to control state in amorphous phase of gas barrier resin in order to achieve both moldability and good gas barrier property. Polymer alloy was designed to improve flexibility. Polymer alloy made of amorphous PVA and elastomer resin showed good hydrogen resistance. Even after its polymer alloy were repeatedly exposed to 70MPa hydrogen gas the influence on higher-order structure in amorphous phase was in negligible level.

In this work we evaluate the consequences of the combustion of a stoichiometric mixture of hydrogen-air on a mechanical device which can be considered as a long tube. In order to choose the most dangerous combustion regime for the mechanical device we devote a particular attention to the investigation of the 1D deflagration-to-detonation transition. Then once established the most dangerous combustion regime we compute the reacting  flow and the stress and strain in the mechanical device. Analyses are performed using both semi-analytical solutions and Europlexus a computer program for the simulation of  fluid-structure systems under transient dynamic loading.

In order to simulate an accidental hydrogen release from the low pressure pipe system of a hydrogen vehicle a systematic study on the nature of transient hydrogen jets into air and their combustion behaviour was performed at the FZK hydrogen test site HYKA. Horizontal unsteady hydrogen jets with an amount of hydrogen up to 60 STP dm3 and initial pressures of 5 and 16 bar have been investigated. The hydrogen jets were ignited with different ignition times and positions. The experiments provide new experimental data on pressure loads and heat releases resulting from the deflagration of hydrogen-air clouds formed by unsteady turbulent hydrogen jets released into a free environment. It is shown that the maximum pressure loads occur for ignition in a narrow position and time window. The possible hazard potential arising from an ignited free transient hydrogen jet is described.

The aim of the study is to identify and quantify the additional risks related to hydrogen explosions during the operation of a hydrogen-driven car. In a first attempt the accidents or failures of a simple one-tank hydrogen storage system have been studied as a main source of risk. Three types of initiators are taken into account: crash accidents fire accidents without crash (no other cars are involved) and hydrogen leakages in normal situation with following ignition. The consequences of hydrogen ignition and/or explosion depend strongly on environmental conditions (geometry wind etc.) therefore the different configurations of operational and environmental conditions are specified.<br/>Then Event Tree/Fault Tree methods are applied for the risk assessment.<br/>The results of quantification permit to draw conclusions about the overall added risk of hydrogen technology as well as about the main contributors to the risk. Results of this work will eventually contribute to the on-going pre-normative research in the field of hydrogen safety.