Safety

Hydrogen is widely recognized as an attractive energy carrier due to its low-level air pollution and its high mass-related energy density. However its wide flammability range and high burning velocity present a potentially significant hazard. A significant fraction of hydrogen is stored and transported as a cryogenic liquid. Therefore loss of hydrogen containments may lead to the formation of a pool on the ground. In general very large spills will give a pool whereas moderate sized spills may evaporate immediately. Accurate hazard assessments of storage systems require a proper prediction of the liquid hydrogen pool evaporation and spreading. A new pool model handling the spread and the evaporation of liquid spills on different surfaces has recently been developed in the 3D Computational Fluid Dynamics (CFD) tool FLACS [1-4]. As the influence of geometry on the liquid spread is taken into account in the new pool model realistic industrial scenarios can be investigated. The model has been validated for LNG spills on water with the Burro and Coyote experiments [56]. The model has previously been tested for LH2 release in the framework of the EU-sponsored Network of Excellence HySafe where experiments carried out by BAM were modelled. In the large scale BAM experiments [7] 280 kg of liquid hydrogen was spilled in 6 tests adjacent to buildings. In these tests the pool spreading the evaporation and the cloud formation were investigated. Simulations of these tests are found to compare reasonably well with the experimental results. In the present work the model is extended and the liquid hydrogen spill experiments carried out by NASA are simulated with the new pool model. The large scale NASA experiments [89] consisted of 7 releases of liquefied hydrogen at White Sand New Mexico. The release test 6 is used. During these experiments cloud concentrations were measured at several distances downwind of the spill point. With the new pool model feature the FLACS tool is shown to be an efficient and accurate tool for the investigation of complex and realistic accidental release scenarios of cryogenic liquids.

In many areas European research has been largely fragmented. To support the required integration and to focus and coordinate related research efforts the European Commission created a new instrument the Networks of Excellences (NoEs). The goal of the NoE HySafe has been to provide the basis to facilitate the safe introduction of hydrogen as an energy carrier by removing the safety related obstacles. The prioritisation of the HySafe internal project activities was based on a phenomena identification and ranking exercise (PIRT) and expert interviews. The identified research headlines were “Releases in (partially) confined areas” “Mitigation” and “Quantitative Risk Assessment”. Along these headlines existing or planned research work was re-orientated and slightly modified to build up three large internal research projects “InsHyde” “HyTunnel” and “HyQRA”. In InsHyde realistic indoor hydrogen leaks and associated hazards have been investigated to provide recommendations for the safe use of indoor hydrogen systems including mitigation and detection means. The appropriateness of available regulations codes and standards (RCS) has been assessed. Experimental and numerical work was conducted to benchmark simulation tools and to evaluate the related recommendations. HyTunnel contributed to the understanding of the nature of the hazards posed by hydrogen vehicles inside tunnels and its relative severity compared to other fuels. In HyQRA quantitative risk assessment strategies were applied to relevant scenarios in a hydrogen refuelling station and the performance was compared to derive also recommendations. The integration provided by the network is manifested by a series of workshops and benchmarks related to experimental and numerical work. Besides the network generated the following products: the International Conference on Hydrogen Safety the first academic education related to hydrogen safety and the Safety Handbook. Finally the network initiated the founding of the International Association for Hydrogen Safety which will open up the future networking to all interested parties on an international level. The indicated results of this five years integration activity will be described in short.

A crucial problem of new hydrogen technologies is the lightweight and also safe storage of acceptable amounts of hydrogen for portable or mobile applications. A new and innovative technology based on capillary arrays has been developed. These systems ensure safe infusion storage and controlled release of hydrogen gas although storage pressures up to 1200 bar are applied. This technology enables the storage of a significantly greater amount of hydrogen than other approaches. In storage tests with first capillary arrays a gravimetric storage capacity of about 33% and a volumetric capacity of 28% was determined at a comparative low pressure of only 400 bar. This is much more than the actual published storage capacities which are to find for other storage systems. This result already surpassed the US Department of Energy's 2010 target and it is expected to meet the DOE's 2015 target in the near future.<br/>Different safety aspects have been evaluated. On the one hand experiments with single capillaries or arrays of them have been carried out. The capillaries are made of quartz and other glasses. Especially quartz has a three times higher strength than steel. At the same time the density is about three times lower which means that much less material is necessary to reach the same pressure resistance. The pressure resistance of single capillaries has been determined in dependence of capillary materials and dimensions wall thickness etc. in order to find out optimal parameters for the “final” capillaries. In these tests also the sudden release of hydrogen was tested in order to observe possible spontaneous ignitions. On the other hand a theoretical evaluation of explosion hazards was done. Different situations were analyzed e.g. release of hydrogen by diffusion or sudden rupture.

A three-dimensional mathematical model of gaseous hydrogen deflagration in the enclosed space is developed. The process is described by the system of gas dynamics differential equations. Thermodynamic parameters of the mixture and its components are defined as functions of the local temperature and mixture composition. The concentration changes of the fuel and combustion products are determined using conservation laws taking into account rates of component disappearance and formation and turbulent diffusion. It is assumed that the chemical reaction takes place only in the volume where the fuel concentration is within the limits of inflammability. The mathematical model is validated during an intercomparison test to predict deflagration of a large-scale hydrogen-air mixture in open atmosphere. An algorithm of numerical solution based on the Godunov method is developed. A computer system of engineering analysis of gas-dynamic processes of hydrogen-air mixture formation and combustion in enclosed space with natural ventilation is created. It allows predicting the history of the changes of overpressure temperature concentrations of hydrogen and combustion products and other thermogasdynamic parameters of the mixture in space. This prognosis can be used to estimate dangerous zones of destruction and recommend some safety measures.

For hydrogen fuelled vehicles to attain significant market penetration it is essential that any potential risks be controlled within acceptable levels. To achieve this goal on-board vehicle hydrogen storage systems should undergo risk analyses during early concept development and design phases. By so doing the process of eliminating safety-critical failure modes will help guide storage system development and be more efficient to implement than if undertaken after the design-freeze stage. The focus of this paper is the development of quantitative risk analyses of storage systems which use onboard reversible materials such as conventional AB5 metal hydrides the complex hydride NaAlH4 or other material candidates currently being researched. Collision of a vehicle having such a hydrogen storage system was selected as a dominant accident initiator and a probabilistic event tree model has been developed for this initiator. The event tree model contains a set of comprehensive mutually exclusive accident sequences. The event tree represents chronological ordering of key events that are postulated to occur sequentially in time during the accident progression. Each event may represent occurrence of a phenomenon (e.g. hydride chemical reaction and dust cloud explosion) or a hardware failure (e.g. hydride storage vessel rupture). Event tree branch probabilities can be quantified using fault tree models or basic events with probability distributions. A fault tree model for hydride dust cloud explosion is provided as an example. Failure probabilities assigned to the basic events in the fault tree can be estimated from test results published data or expert opinion elicitation. To account for variabilities in the probabilities assigned to fault tree basic events and hence to propagate uncertainties in event tree sequences Monte Carlo sampling and Latin Hypercube sampling were employed and the statistics of the results from both techniques were compared.

Because of air quality problems the problem of CO2 related greenhouse gas emissions and shortage of fossil fuels many vehicles with gaseous fuels (CNG biogas hydrogen etc.) are under research and development. Such vehicles have to prove that as well as their exhaust emissions their overall system emissions (including running loss) remain below certain safety limits before they can be used in practice. This paper presents a cost-effective way of monitoring such system emissions from hydrogen or other gaseous fuel powered vehicles within an air-conditioned chassis dynamometer test cell as commonly used for low ambient emission tests on gasoline vehicles. The only additional equipment needed is a low-concentration sensor for the gas of interest (e.g hydrogen). The method is based on concentration measurements and a dynamic mass balance model. This method is based on the fact that atoms cannot vanish. Applied to a room containing a gas mixture this means that the change of mass of a gaseous matter (called gas G subsequently) inside the chamber is the difference of all mass of G flowing into the chamber and all mass of G flowing out of the chamber. This assumes that no chemical reactions of the gas in mind with other matter take place. By measuring the flow rates and concentrations of ventilation-in flow and ventilation-out flow as well as room concentration the emissions of G of a source i.e. the vehicle to be tested can be calculated. These concentrations need to be measured as functions of time to be able to give values of emissions per time unit. It is shown by a real experiment that very low emissions can be recorded. Additionally error bounds and sensitivities on different parameters such as air exchange ratio are quantified.

A buoyant round vertical hydrogen jet is investigated using Large Eddy Simulations at low Mach number (M = 0.3). The influence of the transient concentration fields on the extent of the gas envelope with concentrations within the flammability limits is analyzed and their structure are characterized. The transient flammable region has a complex structure that extends up to 30% beyond the time-averaged flammable volume with high concentration pockets that persist sufficiently long for potential ignition. Safety envelopes devised on the basis of simplified time-averaged simulations would need to include a correction factor that accounts for transient incursions of high flammability concentrations.

Computational fluid dynamic simulations have been performed in order to study the consequences of a hydrogen release from a pressure swing adsorption installation operating at 30 barg. The simulations were performed using FLACS-Hydrogen software from GexCon. The impact of obstruction partial confinement leak orientation and wind on the explosive cloud formation (size and explosive mass) and on explosion consequences is investigated. Overpressures resulting from ignition are calculated as a function of the time to ignition.

Most studies on blast waves generated by gas explosions have focused on gas explosions occurring in open spaces. However accidental gas explosions often occur in confined spaces and the blast wave generates from a bursting vessel as a result of an increase in pressure caused by the gas explosion. In this study blast waves from bursting plastic vessels in which gas explosions occurred are investigated. The flammable mixtures used in the experiments were hydrogen-air mixtures at several equivalence ratios and a stoichiometric methane-air mixture. The overpressures of the blast waves were generated by venting high-pressure gas in the enclosure and volumetric expansion with a combustion reaction. The measured intensities of the blast waves were greater than the calculated values resulting from high-pressure bursting without a combustion reaction. The intensities of the blast waves resulting from the explosions of hydrogen-air mixtures were much greater than those of the methane-air mixture.

Since the rapid development of hydrogen stationary and vehicle fuel cells the last decade it is of importance to improve the prediction of overpressure generated during an accidental explosion which could occur in a confined part of the system. To this end small-scale vented hydrogen–air explosions were performed in a transparent cubic enclosure with a volume of 3375 cm3. The flame propagation was followed with a high speed camera and the overpressure inside the enclosure was recorded using high frequency piezoelectric transmitters. The effects of vent area and ignition location on the amplitude of pressure peaks in the enclosed volume were investigated. Indeed vented deflagration generates several pressures peaks according to the configuration and each peak can be the dominating pressure. The parametric study concerned three ignition locations and five square vent sizes.