Safety

GASFLOW is a finite-volume computer code that solves the time-dependent two-phase homogeneous equilibrium model compressible Navier–Stokes equations for multiple gas species with turbulence. The fluid-dynamics algorithm is coupled with conjugate heat and mass transfer models to represent walls floors ceilings and other internal structures to describe complex geometries such as those found in nuclear containments and facilities. Recent applications involve simulations of cryogenic hydrogen tanks at elevated pressures. These applications which often have thermodynamic conditions near the critical point require more accurate real-gas Equations-of-State (EoS) and transport properties than the standard ideal gas EoS and classical kinetic-theory transport properties. This paper describes the rigorous implementation of the generalized real-gas EoS into the GASFLOW CFD code as well as the specific implementation of respective real-gas models (Leachman's NIST hydrogen EoS a modified van der Waals EoS and a modified Nobel-Abel EoS); it also includes a logical testing procedure based upon a numerically exact benchmark problem. An example of GASFLOW simulations is presented for an ideal cryo-compressed hydrogen tank of the type utilized in fuel cell vehicles.

Computational fluid dynamics (CFD) is used to numerically solve the sudden release of hydrogen from a high pressure tank (up to 70MPa) into air. High pressure tanks increase the risk of failure of the joints and pipes connected to the tank which results in release of Hydrogen. The supersonic flow caused by high pressure ratio of reservoir to ambient generates a strong Mach disk. A three dimensional in-house code is developed to simulate the flow. High pressure Hydrogen requires a real gas law because it deviates from ideal gas law. Firstly Beattie-Bridgeman and Abel-Noble real gas equation of states are applied to simulate the release of hydrogen in hydrogen. Then Abel-Noble is implied to simulate the release of hydrogen in air. Beattie-Bridgeman has stability problems in the case of hydrogen in air. A transport equation is used to solve the concentration of Hydrogen-air mixture. The code is second order accurate in space and first order in time and uses a modified Van Leer limiter. The fast release of Hydrogen from a small rupture needs a very small mesh therefore parallel computation is applied to overcome memory problems and to decrease the solution time. The high pressure ratio of the reservoir to ambient causes a very fast release which is accurately modelled by the code and all the shocks and Mach disk happening are observed in the results. The results show that the difference between real gas and ideal gas models cannot be ignored.

The competition between chemical energy release rate and volumetric expansion related to shock wave’s dynamics is of primary importance for a number of situations relevant to explosion safety. While studies have been performed on this topic over the years they have been limited to mixtures with monotonous energy release profile. In the present study the ignition of H2-NO2/N2O4 mixtures which exhibit a single-step or a two-step energy release rate profile depending on the equivalence ratio has been investigated under volumetric expansion conditions. The rate of expansion has been calculated using the Taylor-Sedov solution and accounted for using 0-D numerical simulations with time-dependent specific volume. The results were analyzed in terms of a Damkohler number defined as the ratio of the expansion to ignition times. For mixtures with non-monotonous energy release rate profiles two critical Damkohler numbers can be identified one for each of the steps of energy release. It was also shown that the fluid element which is the most likely to ignite corresponds to the one behind a shock propagating at the Chapman-Jouguet velocity. The thermo-chemical dynamics have been analyzed about the critical conditions using energy release rate per reaction rate of production and sensitivity analyses.

The current study addresses the spontaneous ignition of hydrogen jets released into a confined oxidizer environment experimentally. The experiments are conducted in a shock tube where hydrogen gas is shock-accelerated into oxygen across a perforated plate. The operating conditions and hole dimension of the plate were varied in order to identify different flow field and ignition scenarios. Time resolved Schlieren visualization permitted to reconstruct the gasdynamic evolution of the release and different shock interactions. Time resolved self-luminosity records permitted us to record whether ignition was achieved and also to record the dimension of the turbulent mixing layer. The ignition limits determined experimentally in good agreement with the 1D diffusion ignition model proposed by Maxwell and Radulescu. Nevertheless the experiments demonstrated that the mixing layer is two to three orders of magnitude thicker than predicted by molecular diffusion which can be attributed to the observed mixing layer instabilities and shock-mixing layer interactions which provide a much more intense mixing rate than anticipated from previous and current numerical predictions. These observations further clarify why releases through partly confined geometries are more conducive to jet ignition of the jets.

The paper will present results of hydrogen/oxygen mixtures ignited by using electric sparks electrostatic discharges a heating element and a flame. Measurements of the lower flammability limit (LFL) was done for each ignition method. The hydrogen mixtures of different concentrations were ignited at the bottom of a combustion chamber leading to an upward propagation of the resulting flame. At some level of concentration the combustion was partial due to the limited upward propagation. The complete combustion of the whole mixture was observed at concentration limits higher than the known LFL of 4% vol. for hydrogen in air. The paper will describe the test facilities and the resulting ignition probabilities for different ignition methods.

The development of a set of safety codes and standards for hydrogen facilities is necessary to ensure they are designed and operated safely. To help ensure that a hydrogen facility meets an acceptable level of risk code and standard development organizations (SDOs) are utilizing risk-informed concepts in developing hydrogen codes and standards. Two SDOs the National Fire Protection Association (NFPA) and the International Organization for Standardization (ISO) through its Technical Committee (TC) 197 on hydrogen technologies have been developing standards for gaseous hydrogen facilities that specify the facilities have certain safety features use equipment made of material suitable for a hydrogen environment and have specified separation distances. Under Department of Energy funding Sandia National Laboratories (SNL) has been supporting efforts by both of these SDOs to develop the separation distances included in their respective standards. Important goals in these efforts are to use a defensible science-based approach to establish these requirements and to the extent possible harmonize the requirements. International harmonization of regulations codes and standards is critical for enabling global market penetration of hydrogen and fuel cell technologies.

In order to improve risk analyses and influence the design of the future H2 systems an experimental study on “real” leaks qualification and quantification was performed. In H2 energy applications fittings appeared as a significant leakage potential and subsequently explosion and flame hazards. Thus as a part of the “Horizon Hydrogène Energie” French program four kinds of commercial fittings usually employed on H2 systems were tested thanks to a new high pressure test bench – designed setup and operated by INERIS – allowing experiments to be led for H2 pressures until 700 bar. The fittings underwent defined stresses representative of H2 systems lifetime and beyond. The associated leaks – when existing – are characterized in terms of flow rate.

Sandia National Laboratories is working with stakeholders to develop scientific data for use by standards development organizations to create hydrogen codes and standards for the safe use of liquid hydrogen. Knowledge of the concentration field and flammability envelope for high-pressure hydrogen leaks is an issue of importance for the safe use of liquid hydrogen. Sandia National Laboratories is engaged in an experimental and analytical program to characterize and predict the behaviour of liquid hydrogen releases. This paper presents a model for computing hydrogen dilution distances for cold hydrogen releases. Model validation is presented for leaks of room temperature and 80 K high-pressure hydrogen gas. The model accounts for a series of transitions that occurs from a stagnate location in the tank to a point in the leak jet where the concentration of hydrogen in air at the jet centerline has dropped to 4% by volume. The leaking hydrogen is assumed to be a simple compressible substance with thermodynamic equilibrium between hydrogen vapor hydrogen liquid and air. For the multi-phase portions of the jet near the leak location the REFPROP equation of state models developed by NIST are used to account for the thermodynamics. Further downstream the jet develops into an atmospheric gas jet where the thermodynamics are described as a mixture of ideal gases (hydrogen–air mixture). Simulations are presented for dilution distances in under-expanded high-pressure leaks from the saturated vapor and saturated liquid portions of a liquid hydrogen storage tank at 10.34 barg (150 PSIG).

Conditions are examined under which mechanical stimuli produced by striking controlled blows can result in sparking and ignition of hydrogen in air mixtures. The investigation principally concerns magnesium thermite reaction as the ignition source and focuses on the conditions and thermomechanical parameters that are involved in determining the probability of ignition. It is concluded that the notion of using the kinetic energy of impact as the main criterion in determining whether an ignition event is likely or not is much less useful than considering the parameters which determine the maximum temperature produced in a mechanical stimuli event. The most influential parameter in determining ignition frequency or probability is the velocity of sliding movement during mechanical stimuli. It is also clear that the kinetic energy of a moving hammer head is of lesser importance than the normal force which is applied during contact. This explains the apparent discrepancy in previous studies between the minimum kinetic energy thought to be necessary to allow thermite sparking and gas ignition to occur with drop weight impacts and glancing blow impacts. In any analysis of the likelihood of mechanical stimuli to cause ignition the maximum surface temperature generated should be determined and considered in relation to the temperatures that would be required to initiate hot surface reactions sufficient to cause sparking and ignition.

The U.S. Department of Energy supports the implementation of hydrogen fuel cell technologies by providing hydrogen safety and emergency response training to first responders. A collaboration was formed to develop and deliver a one-day course that uses a mobile fuel cell vehicle (FCV) burn prop designed and built by Kidde Fire Trainers. This paper describes the development of the training curriculum including the design and operation of the FCV prop; describes the successful delivery of this course to over 300 participants at three training centers in California; and discusses feedback and observations received on the course. Photographs and video clips of the training sessions will be presented.

This paper presents the current results of the theoretical and experimental activity carried out by the Italian Working Group on the fire prevention safety issues in the field of the hydrogen transport in pipelines. From the theoretical point of view a draft document has been produced beginning from the regulations in force on the natural gas pipelines; these have been reviewed corrected and integrated with the instructions suitable to the use with hydrogen gas. From the experimental point of view a suitable apparatus has been designed and installed at the University of Pisa; this apparatus will allow the simulation of hydrogen releases from a pipeline with or without ignition of the hydrogen-air mixture. The experimental data will help the completion of the above-mentioned draft document with the instructions about the safety distances. However in the opinion of the Group the work on the text contents is concluded and the document is ready to be discussed with the Italian stakeholders involved in the hydrogen applications.

The main objective of this study is an insight into physical phenomena underlying spontaneous ignition of hydrogen at sudden release from high pressure storage and its transition into the sustained jet fire. This paper describes modelling and large eddy simulation (LES) of spontaneous ignition dynamics in a tube with a rupture disk separating high pressure hydrogen storage and the atmosphere. Numerical experiments carried out by a LES model have provided an insight into the physics of the spontaneous ignition phenomenon. It is demonstrated that a chemical reaction commences in a boundary layer within the tube and propagates throughout the tube cross-section after that. Simulated by the LES model dynamics of flame formation outside the tube has reproduced experimental observation of combustion by high-speed photography including vortex induced “flame separation". It is concluded that the model developed can be applied for hydrogen safety engineering in particular for development of innovative pressure relief devices.

Because of hydrogen low volumetric energy content under its gaseous form transport and storage of liquid hydrogen will certainly play a major role in any future hydrogen economy. One of the obstacles to the expected development use of hydrogen is the poor state of knowledge on explosion risks in the event of an extensive spillage. INERIS set up a large-scale experiment to study the mechanisms of the formation of the gas cloud resulting from such a spillage and the associated mixing process and turbulence effects. Dispersion tests have been performed with cryogenic helium presenting similar dispersion characteristics than liquid hydrogen (buoyancy). Flowrates up to 3 kg/s have been investigated and the instrumentation allowed the observation and quantification of bouyancy effects including internal turbulence. Those results constitute an originals et of data which can be used as a basis for the development of dispersion software and reinterpretation of other existing databases ([10 11])

In the present paper the results of experiments on study of high-speed deflagrations in flat layer of hydrogen-air mixtures unconfined from below are presented. The experiments were performed in two different rectangular channels: small-scale with mixture volume up to 0.4 m3 and large-scale with volume up to 5.5 m3. The main goal of the experiments was to examine the possibility of the layer geometries to maintain high-speed deflagration and detonation. With the aim to study a range of combustion regimes the experiments were performed varying degree of channel obstruction hydrogen concentration and thickness of the layer. Depending on the experimental conditions all major combustion regimes were observed: slow flame fast – ‘choked’ flame and steady-state detonation. It was found that minimum layer layer thickness in the range of 8 to 15 detonation cell widths is required for sustainable detonations.

Virtually all major automotive companies are currently developing hydrogen-powered vehicles. The vast majority of them employ compressed hydrogen tanks and components as a means of storing the fuel onboard. Compressed hydrogen vehicle fuel systems are designed in the same way as compressed natural gas vehicles (NGV) i.e. the high pressure (up to 70 MPa) fuel is always contained within the system under all conditions with the exception of vehicular fire. In the event of a vehicle fire the fuel system is protected using a non-reclosing thermally activated pressure relief device (PRD) which safely vents the contents. Hydrogen fuel system PRDs are presently qualified to the performance requirements specified in draft hydrogen standards such ANSI/CSA HPRD 1 and EIHP Rev. 12b. They are also qualified with individual fuel tank designs in accordance with the engulfing bonfire requirements in various published/draft tank standards such as CSA B51 Part 2 JARI S001 SAE TIR J2579 ANSI/CSA HGV 2 ISO DIS 15869.2 and EIHP Rev. 12b. Since 2000 there have been over 20 documented NGV tank failures in service 11 of which have been attributed to vehicle fires. This paper will examine whether currently proposed hydrogen performance standards and installation requirements offer suitable fuel system protection in the event of vehicular fires. A number of alternative fire protection strategies will be discussed including:

• The requirement of an engulfing and/or localized fire test for individual tanks fuel systems and complete vehicles;
• The advantages/disadvantages of point source- surface area- and/or fuse-based PRDs
• The use of thermal insulating coatings/blankets for fire protection resulting in the NONventing of the fuel
• The specification of appropriate fuel system installation requirements to mitigate the effect of vehicular fires.

Extensive infrastructure exists for the transport of natural gas and it is an obvious step to assess its use for the movement of hydrogen. The Naturalhy project’s objective is to prepare the European natural gas industry for the introduction of hydrogen by assessing the capability of the natural gas infrastructure to accept mixtures of hydrogen and natural gas. This paper presents the ongoing work within both Durability and Integrity Work Packages of the Naturalhy project. This work covers a gap in knowledge on risk assessment required for delivering H2+natural gas blends by means of the existing natural gas grids in safe operation.<br/>Experiments involving several parts of the existing infrastructure will be described that are being carried out to re-examine the major risks previously studied for natural gas including: effect of H2 on failure behaviour and corrosion of transmission pipes and their burst resistance (link to the Work Package Safety) on permeability and ageing of distribution pipes on reliability and ageing of domestic gas meters tightness to H2 of domestic appliances and their connexions. The information will be integrated into existing Durability assessment methodologies originally developed for natural gas.<br/>An Integrity Management Tool will be developed taking account of the effect of hydrogen on the materials properties. The tool should enable a cost effective selection of appropriate measures to control the structural integrity and maintaining equipment. The main measures considered are monitoring non destructive examination (pigging and non pigging) and repair strategies. The tool will cover a number of parameters e.g.: percentage of hydrogen in the gas mixture material of construction operating conditions and condition of cathodic protection. Thus the Integrity Management Tool will yield an inspection and maintenance plan based on the specific circumstances.

If the general public is to use hydrogen as a vehicle fuel customers must be able to handle hydrogen with the same degree of confidence and with comparable risk as conventional liquid and gaseous fuels. The hazards associated with jet releases from leaks in a vehicle-refuelling environment must be considered if hydrogen is stored and used as a high-pressure gas since a jet release in a confined or congested area could result in an explosion. As there was insufficient knowledge of the explosion hazards a study was initiated to gain a better understanding of the potential explosion hazard consequences associated with high-pressure leaks from refuelling systems. This paper describes two experiments with a dummy vehicle and dispenser units to represent refuelling station congestion. The first represents a ‘worst-case’ scenario where the vehicle and dispensers are enveloped by a 5.4 m x 6.0 m x 2.5 m high pre-mixed hydrogen-air cloud. The second is an actual high-pressure leak from storage at 40 MPa (400 bar) representing an uncontrolled full-bore failure of a vehicle refuelling hose. In both cases an electric spark ignited the flammable cloud. Measurements were made of the explosion overpressure generated its evolution with time and its decay with distance. The results reported provide a direct demonstration of the explosion hazard from an uncontrolled leak; they will also be valuable for validating explosion models that will be needed to assess configurations and conditions beyond those studied experimentally.

To safely remove from its fire accident site a hydrogen fuel cell vehicle equipped with a carbon fiber reinforced plastic composite cylinder for compressed hydrogen (CFRP cylinder) and to safely keep the burnt vehicle in a storage facility it is necessary to verify whether the thermally-activated pressure relief device (TPRD) of the CFRP cylinder has already been activated releasing the hydrogen gas from the cylinder. To develop a simple post-fire verification method on TPRD activation the present study was conducted on the using hydrogen densitometer and Type III and Type IV CFRP cylinders having different linings. As the results TPRD activation status can be determined by measuring hydrogen concentrations with a catalytic combustion hydrogen densitometer at the cylinder's TPRD gas release port.

Explosion venting is commonly used in the process industry as a prevention solution to protect equipment or buildings against excessive internal pressure caused by an explosion. This article is dedicated to the validation of FLACS CFD code for the modelling of vented explosions. Analytical engineering models fail when complex cases are considered for instance in the presence of obstacles or H2 stratified mixtures. CFD is an alternative solution but has to be carefully validated. In this study FLACS simulations are compared to published experimental results and recommendations are suggested for their application.

This paper presents results of experimental investigations on spherical and cylindrical flame propagation in pre-mixed H2/air-mixtures in unconfined and semi-confined geometries. The experiments were performed in a facility consisting of two transparent solid walls with 1 m2 area and four weak side walls made from thin plastic film. The gap size between the solid walls was varied stepwise from thin layer geometry (6 mm) to cube geometry (1 m). A wide range of H2/air-mixtures with volumetric hydrogen concentrations from 10% to 45% H2 was ignited between the transparent solid walls. The propagating flame front and its structure was observed with a large scale high speed shadow system. Results of spherical and cylindrical flame propagation up to a radius of 0.5 m were analyzed. The presented spherical burning velocity model is used to discuss the self-acceleration phenomena in unconfined and unobstructed pre-mixed H2/air flames.

The safe decommissioning as well as decontamination of the radioactive waste resulting from the nuclear accident in Fukushima Daiichi represents a huge task for the next decade. At present research and development on long-term safe storage containers has become an urgent task with international cooperation in Japan. One challenge is the generation of hydrogen and oxygen in significant amounts by means of radiolysis inside the containers as the nuclear waste contains a large portion of sea water. The generation of radiolysis gases may lead to a significant pressure build-up inside the containers and to the formation of flammable gases with the risk of ignition and the loss of integrity.

In the framework of the project “R&D on technology for reducing concentration of flammable gases generated in long-term waste storage containers” funded by the Japanese Ministry of Education Culture Sports Science and Technology of Japan (MEXT) the potential application of catalytic recombiner devices inside the storage containers is investigated. In this context a suitable catalyst based on the so-called intelligent automotive catalyst for use in a recombiner is under consideration. The catalyst is originally developed and mass-produced for automotive exhaust gas purification and is characterized by having a self-healing function of precious metals (Pd Pt and Rh) dissolved as a solid solution in the perovskite type oxides. The basic features of this catalyst have been tested in an experimental program. The test series in the REKO-4 facility has revealed the basic characteristics of the catalyst required for designing the recombiner system.

In the present study the self-ignition of high-pressure hydrogen released in atmospheric air through a diaphragm is visualized under various test conditions. The experimental results indicate that the hydrogen that jets through the rupturing diaphragm is mixed with the heated air near the tube wall. The self-ignition event originated from this mixing. The self-ignition was strongly dependent on the strength of an incident shock wave generated at the diaphragm rupture. As a result a cylindrical flame that formed after the self-ignition shows a tendency to become longer as it propagates in the downstream direction. The head velocities of the hydrogen-air mixture and the cylindrical flame are consistent with that of a contact surface calculated from the measured shock speed. A modified self-ignition mechanism is proposed based on the experimental observations.

The paper presents an overview of the main achievements of the internal project InsHyde of the HySafe NoE. The scope of InsHyde was to investigate realistic small-medium indoor hydrogen leaks and provide recommendations for the safe use/storage of indoor hydrogen systems. Additionally InsHyde served to integrate proposals from HySafe work packages and existing external research projects towards a common effort. Following a state of the art review InsHyde activities expanded into experimental and simulation work. Dispersion experiments were performed using hydrogen and helium at the INERIS gallery facility to evaluate short and long term dispersion patterns in garage like settings. A new facility (GARAGE) was built at CEA and dispersion experiments were performed there using helium to evaluate hydrogen dispersion under highly controlled conditions. In parallel combustion experiments were performed by FZK to evaluate the maximum amount of hydrogen that could be safely ignited indoors. The combustion experiments were extended later on by KI at their test site by considering the ignition of larger amounts of hydrogen in obstructed environments outdoors. An evaluation of the performance of commercial hydrogen detectors as well as inter-lab calibration work was jointly performed by JRC INERIS and BAM. Simulation work was as intensive as the experimental work with participation from most of the partners. It included pre-test simulations validation of the available CFD codes against previously performed experiments with significant CFD code inter-comparisons as well as CFD application to investigate specific realistic scenarios. Additionally an evaluation of permeation issues was performed by VOLVO CEA NCSRD and UU by combining theoretical computational and experimental approaches with the results being presented to key automotive regulations and standards groups. Finally the InsHyde project concluded with a public document providing initial guidance on the use of hydrogen in confined spaces.

We have investigated hydrogen explosion risk and its mitigation focusing on compact hydrogen refuelling stations in urban areas. In this study numerical analyses were performed of hydrogen blast propagation and the structural behaviour of barrier walls. Parametric numerical simulations of explosions were carried out to discover effective shapes for blast-mitigating barrier walls. The explosive source was a prismatic 5.27 m3 volume that contained 30% hydrogen and 70% air. A reinforced concrete wall 2 m tall by 10 m wide and 0.15 m thick was set 2 or 4 m away from the front surface of the source. The source was ignited at the bottom centre by a spark for the deflagration case and 10 g of C-4 high explosive for two detonation cases. Each of the tests measured overpressures on the surfaces of the wall and on the ground displacements of the wall and strains of the rebar inside the wall. The blast simulations were carried out with an in-house CFD code based on the compressive Euler equation. The initial energy estimated from the volume of hydrogen was a time-dependent function for the deflagration and was released instantaneously for the detonations. The simulated overpressures were in good agreement with test results for all three test cases. DIANA a finite element analysis code released by TNO was used for the structural simulations of the barrier wall. The overpressures obtained by the blast simulations were used as external forces. The analyses simulated the displacements well but not the rebar strains. The many shrinkage cracks that were observed on the walls some of which penetrated the wall could make it difficult to simulate the local behaviour of a wall with high accuracy and could cause strain gages to provide low-accuracy data. A parametric study of the blast simulation was conducted with several cross-sectional shapes of barrier wall. A T-shape and a Y-shape were found to be more effective in mitigating the blast.

The release of gases heavier than air like propane at ground level or lighter than air like hydrogen close to a ceiling can both lead to fire and explosion hazards that must be carefully considered in safety analyses. Even if the simulation of accident scenarios in complex installations and long transients often appears feasible only using lumped parameter computer codes the phenomenon of denser or lighter gas dispersion is not implicitly accounted by these kind of tools. In the aim to set up an ad hoc model to be used in the computer code ECART fluid-dynamic simulations by the commercial FLUENT 6.0 CFD code are used. The reference geometry is related to cavities having variable depth (2 to 4 m) inside long tunnels filled with a gas heavier or lighter than air (propane or hydrogen). Three different geometrical configurations with a cavity width of 3 6 and 9 m are considered imposing different horizontal air stream velocities ranging from 1 to 5 m/s. A stably-stratified flow region is observed inside the cavity during gas shearing. In particular it is found that the density gradient tends to inhibit turbulent mixing thus reducing the dispersion rate. The obtained data are correlated in terms of main dimensionless groups by means of a least squares method. In particular the Sherwood number is correlated as a function of Reynolds a density ratio modified Froude numbers and in terms of the geometrical parameter obtained as a ratio between the depth of the air-dense gas interface and the length of the cavity. This correlation is implemented in the ECART code to add the possibility to simulate large installations during complex transients lasting many hours with reasonable computation time. An example of application to a typical case is presented.

Reliable and effective sensors for the accurate detection of hydrogen concentrations in air are essential for the safe operation of fuel cells hydrogen fuelled systems (e.g. vehicles) and hydrogen production distribution and storage facilities. The present paper describes the activity on-going at JRC for the establishment of a facility that can be used for testing and validating the performance of hydrogen sensors under a range of conditions representative of those to be encountered in service. Potential aspects to be investigated in relation to the sensors performances are the influence of temperature humidity and pressure (simulating variations in altitude) the sensitivity to target gas and the cross sensitivity to other gases/vapours the reaction and recovery time and the sensors’ lifetime. The facility set up at JRC for the execution of these tests is described including the program for its commissioning. The results of a preliminary test are presented and discussed as an example.

The method of the numerical solution of a three-dimensional problem of atmospheric release dispersion and explosion of gaseous admixtures is presented. It can be equally applied for gases of different densities including hydrogen. The system of simplified Navie-Stocks equations received by truncation of viscous members (Euler equations with source members) is used to obtain a numerical solution. The algorithm is based on explicit finite-difference Godunov scheme of arbitrary parameters breakup disintegration. To verify the developed model and computer system comparisons of numerical calculations with the published experimental data on the dispersion of methane and hydrocarbons explosions have been carried out. Computational experiments on evaporation and dispersion of spilled liquid hydrogen and released gaseous hydrogen at different wind speeds have been conducted. The largest mass concentrations of hydrogen between the bottom and top limits of flame propagation and cloud borders have been determined. The problem of the explosion of a hydrogen-air cloud of the complex form generated by large-scale spillage of liquid hydrogen and instant release of gaseous hydrogen has been numerically solved at low wind speed. Shock-wave loadings affecting the buildings located on a distance of 52 m from a hydrogen release place have been shown.

The paper presents the results of the CFD inter-comparison exercise SBEP-V3 performed within the activity InsHyde internal project of the HYSAFE network of excellence in the framework of evaluating the capability of various CFD tools and modelling approaches in predicting the physical phenomena associated to the short and long term mixing and distribution of hydrogen releases in confined spaces. The experiment simulated was INERIS-TEST-6C performed within the InsHyde project by INERIS consisting of a 1 g/s vertical hydrogen release for 240 s from an orifice of 20 mm diameter into a rectangular room (garage) of dimensions 3.78x7.2x2.88 m in width length and height respectively. Two small openings at the front and bottom side of the room assured constant pressure conditions. During the test hydrogen concentration time histories were measured at 12 positions in the room for a period up to 5160 s after the end of release covering both the release and the subsequent diffusion phases. The benchmark was organized in two phases. The first phase consisted of blind simulations performed prior to the execution of the tests. The second phase consisted of post calculations performed after the tests were concluded and the experimental results made available. The participation in the benchmark was high: 12 different organizations (2 non-HYSAFE partners) 10 different CFD codes and 8 different turbulence models. Large variation in predicted results was found in the first phase of the benchmark between the various modelling approaches. This was attributed mainly to differences in turbulence models and numerical accuracy options (time/space resolution and discretization schemes). During the second phase of the benchmark the variation between predicted results was reduced.

This paper discusses the rationale structure and contents of the Canadian Hydrogen Safety Program developed by the Codes & Standards Working Group of the Canadian Transportation Fuel Cell Alliance consisting of representatives from industry academia government and regulators. The overall program objective is to facilitate acceptance of the products services and systems of the Canadian Hydrogen Industry by the Canadian Hydrogen Stakeholder Community to facilitate trade ensure fair insurance policies and rates ensure effective and efficient regulatory approval procedures and to ensure that the interests of the general public are accommodated. The Program consists of four projects including Comparative Quantitative Risk Assessment of Hydrogen and Compressed Natural Gas (CNG) Refuelling Stations; Computational Fluid Dynamics (CFD) Modelling Validation Calibration and Enhancement; Enhancement of Frequency and Probability Analysis and Consequence Analysis of Key Component Failures of Hydrogen Systems; and Fuel Cell Oxidant Outlet Hydrogen Sensor Project. The Program projects are tightly linked with the content of the IEA Task 19 Hydrogen Safety. The Program also includes extensive (destructive and non-destructive) testing of hydrogen components.

The large eddy simulation (LES) model developed at the University of Ulster has been applied to simulate releases of 5.11 m3 liquefied hydrogen (LH2) in open atmosphere and gaseous hydrogen (GH2) in 20-m3 closed vessel. The simulations of a spill of liquefied hydrogen confirmed the advantage of LES application to reproduce experimentally observed eddy structure of hydrogen-air cloud. The inclination angle of simulated cloud is close to experimentally reported 300. The processes of two phase hydrogen release and heat transfer were simplified by inflow of gaseous hydrogen with temperature 20 K equal to boiling point. It is shown that difference in inflow conditions geometry and grid resolution affects simulation results. It is suggested that phenomenon of air condensationevaporation in the cloud in temperature range 20-90 K should be accounted for in future. The simulations reproduced well experimental data on GH2 release and transport in 20-m3 vessel during 250 min including a phenomenon of hydrogen concentration growth at the bottom of the vessel. Higher experimental hydrogen concentration at the bottom is assumed to be due to non-uniformity of temperature of vessel walls generating additional convection. The comparison of convective and diffusion terms in Navie-Stokes equations has revealed that a value of convective term is more than order of magnitude prevail over a value of turbulent diffusion term. It is assumed that the hydrogen transport to the bottom of the vessel is driven by the remaining chaotic flow velocities superimposed on stratified hydrogen concentration field. Further experiments and simulations with higher accuracy have to be performed to confirm this phenomenon. It has been demonstrated that hydrogen-air mixture became stratified in about 1 min after release was completed. However one-dimensional models are seen not capable to reproduce slow transport of hydrogen during long period of time characteristic for scenarios such as leakage in a garage.

Experimental studies were carried out to examine the lift-off and blow-out stability of H2/CO2 H2/Ar H2/C3H8 and H2/CH4 jet flames. The experiments were carried out using a burner with a 2mm inner diameter. The flame structures were recorded by direct filming and also by a schlieren apparatus. The experiments showed that the four gases affected the lift-off and blow-out stability of the hydrogen differently. The experiments showed that propane addition to an initially attached flame always produced lifted flame and the flame was blown out at higher jet velocity. The blow-out velocity decreased as the increasing in propane concentration. Direct blow-off of hydrogen/propane was never observed. Methane addition resulted in a relatively stable flame comparing with the carbon dioxide and propane addition. Comparisons of the stability of H2/C3H8 H2/CH4 and H2/CO2 flames showed that H2/C3H8 produced the highest lift-off height. Propane is much more effective in lift-off and blow out hydrogen flames. The study carried out a chemical kinetic analysis of H2/CO2 H2/Ar H2/C3H8 and H2/CH4 flames for a comparison of effect of chemical kinetics on flame stability.

Several today’s and future applications in energy technology bear the risk of the formation of flammable hydrogen/air mixtures either due to the direct use of hydrogen or due to hydrogen appearing as a by-product. If there’s the possibility of hydrogen being released accidentally into closed areas countermeasures have to be implemented in order to mitigate the threat of an explosion. In the field of nuclear safety passive auto-catalytic recombiners (PAR) are well-known devices for reducing the risk of a hydrogen detonation in a nuclear power plant in the course of a severe accident. Hydrogen and oxygen react on catalyst materials like platinum or palladium already far below conventional flammability limits. The most important concern with regard to the utilization of hydrogen  recombiners is the adequate removal of the reaction heat. Already low hydrogen concentrations may increase the system temperature beyond the self-ignition limit of hydrogen/air mixtures and may lead to an unintended ignition on hot parts of the PAR.<br/>Starting from the nuclear application since several years IEF-6 and LRST perform joint research in the field of passive auto-catalytic recombiners including experimental studies modelling and development of new design concepts. Recently approaches on specifically designed catalysts and on passive cooling devices have been successfully tested. In a design study both approaches are combined in order to provide means for efficient and safe removal of hydrogen. The paper summarizes results achieved so far and possible designs for future applications.

Previous experiments demonstrated that the accidental release of high pressure hydrogen into air can lead to the possibility of spontaneous ignition. It is believed that this ignition is due to the heating of the mixing layer between hydrogen and air that is caused by the shock wave driven by the pressurized hydrogen during the release. Currently this problem is poorly understood and not amenable to direct numerical simulation. This is due to the presence of a wide range of scales between the sizes of the blast wave driven and the very thin mixing layer. The present study addresses this fundamental ignition problem and develops a solution framework in order to predict the ignition event for given hydrogen storage pressures and dimension of the release hole. In this problem only the mixing layer between the hydrogen and air is considered. This permits us to use much higher resolution than previous studies. This mixing layer at the jet head is advected as a Lagrangian fluid particle. The key physical processes in the problem are identified to be the mixing of the two gases at the mixing layer the initial heating by the shock wave and a cooling effect due to the expansion of the mixing layer. The results of the simulations indicate that for every storage pressure there exists a critical hole size below which ignition is prevented during the release process. Close inspection of the results indicate that this limit is due to the competition between the heating provided by the shock wave and the cooling due to expansion. Furthermore the results also indicate that the details of the mixing process do not play a significant role to leading order. The limiting ignition criteria were found to be well approximated by the Homogeneous Ignition Model of Cuenot and Poinsot supplemented by a heat loss term due to expansion. Therefore turbulent mixing occurring in reality is not likely to affect the ignition limits derived in the present study. Comparison with existing experiments showed very good agreement.

This study addresses one of knowledge gaps in hydrogen safety science and engineering i.e. a predictive model for calculation of deterministic separation distances defined by the parameters of a blast wave generated by a high-pressure gas storage tank rupture in a fire. An overview of existing methods to calculate stored in a tank internal (mechanical) energy and a blast wave decay is presented. Predictions by the existing technique and an original model developed in this study which accounts for the real gas effects and combustion of the flammable gas released into the air (chemical energy) are compared against experimental data on high-pressure hydrogen tank rupture in the bonfire test. The main reason for a poor predictive capability of the existing models is the absence of combustion contribution to the blast wave strength. The developed methodology is able to reproduce experimental data on a blast wave decay after rupture of a stand-alone hydrogen tank and a tank under a vehicle. In this study the chemical energy is dynamically added to the mechanical energy and is accounted for in the energy-scaled non-dimensional distance. The fraction of the total chemical energy of combustion released to feed the blast wave is 5% and 9% however it is 1.4 and 30 times larger than the mechanical energy in the stand-alone tank test and the under-vehicle tank test respectively. The model is applied as a safety engineering tool to four typical hydrogen storage applications including onboard vehicle storage tanks and a stand-alone refuelling station storage tank. Harm criteria to people and damage criteria for buildings from a blast wave are selected by the authors from literature to demonstrate the calculation of deterministic separation distances. Safety strategies should exclude effects of fire on stationary storage vessels and require thermal protection of on-board storage to prevent dangerous consequences of high-pressure tank rupture in a fire.

Numerical studies were conducted with the objective of gaining a better understanding of the consequences of potential explosion that could be associated with release of hydrogen in a confined and unconfined environment. This paper describes the work done by us in modelling explosion of accidental releases of hydrogen using our Fire Explosion Release Dispersion (FRED) software. CAM and SCOPE models in FRED is used for validation of congested/uncongested unconfined and congested/uncongested confined vapour cloud explosion respectively. In the first step CAM is validated against experiments of varying gas cloud size blockage ratio equivalence ratio of the mixture and blockage configuration. The model predictions of explosion overpressure are in good agreement with experiments. The results obtained from FRED i.e. overpressure as a function of distance match well in comparison to the experiments. In the second step SCOPE is validated against vented explosion experiments available in open literature. In general SCOPE reproduces the maximum overpressure within the factor of 2. Moreover it well predicts the trends of increase in overpressure with change in type of the fuel increase in number of obstacles blockage ratio and decrease in the vent size.

Experiments were carried out inside a 25 m3 vented combustion test facility (CVE) with a fixed vent area sealed by a plastic sheet vent. Inside the CVE a 0.64 m3 open vent box called RED-CVE was placed. The vent of the RED-CVE was left open and three different vent area were tested. Two different mixing fans one for each compartment were used to establish homogeneous H2 concentrations. This study examined H2 concentrations in the range between 8.5% vol. to 12.5% vol. and three different ignition locations (1) far vent ignition (2) inside the RED-CVE box ignition and (3) near vent ignition (the vent refers to the CVE vent). Peak overpressures generated inside the test facility and the smaller compartment were measured. The results indicate that the near vent ignition generates negligible peak overpressures inside the test facility as compared to those originated by far vent ignition and ignition inside the RED-CVE box. The experiments with far vent ignition showed a pressure increase with increasing hydrogen concentration which reached a peak value at 11% vol. concentration and then decreased showing a non-monotonic behaviour. The overpressure measured inside the RED-CVE was higher when the ignition was outside the box whereas the flame entered the box through the small vent.

Over the last century there have been reports of high pressure hydrogen leaks igniting for no apparent reason and several ignition mechanisms have been proposed. Although many leaks have ignited there are also reported leaks where no ignition has occurred. Investigations of ignitions where no apparent ignition source was present have often been superficial with a mechanism postulated which whilst appearing to satisfy the conditions prevailing at the time of the release simply does not stand up to rigorous scientific analysis. Some of these proposed mechanisms have been simulated in a laboratory under superficially identical conditions and appear to be rigorous and scientific but the simulated conditions often do not have the same large release rates or quantities mainly because of physical constraints of a laboratory. Also some of the release scenarios carried out or simulated in laboratories are totally divorced from the realistic situation of most actual leaks. Clearly there are gaps in the knowledge of the exact ignition mechanism for releases of hydrogen particularly at the high pressures likely to be involved in future storage and use. Mechanisms which have been proposed in the past are the reverse Joule-Thomson effect; electrostatic charge generation; diffusion ignition; sudden adiabatic compression; and hot surface ignition. Of these some have been characterized by means of computer simulation rather than by actual experiment and hence are not validated. Consequently there are discrepancies between the theories releases known to have ignited and releases which are known to have not ignited. From this postulated ignition mechanisms which are worthy of further study have been identified and the gaps in information have been highlighted. As a result the direction for future research into the potential for ignition of hydrogen escapes has been identified.

The anticipated upscaling of hydrogen energy applications will involve the storage and transport of hydrogen in a cryogenic state. Understanding the potential hazard arising from small leaks in pressurized storage and transport systems is needed to assist safety analysis and development of mitigation measures. The current knowledge of the ignited pressurized cryogenic hydrogen jet flame is limited. Large eddy simulation (LES) with detailed hydrogen chemistry is applied for the reacting flow. The effects of ignition locations are considered and the initial development of the transient flame kernel from the ignition hot spots is analysed. The flame structures namely side flames and envelop flames are observed in the study which are due to the complex interactions between turbulence fuel-air mixing at cryogenic temperature and chemical reactions.

Hydrogen is one of the most suitable solutions to replace hydrocarbons in the future. Hydrogen consumption is expected to grow in the next years. Hydrogen liquefaction is one of the processes that allows for increase of hydrogen density and it is suggested when a large amount of substance must be stored or transported. Despite being a clean fuel its chemical and physical properties often arise concerns about the safety of the hydrogen technologies. A potentially critical scenario for the liquid hydrogen (LH2) tanks is the catastrophic rupture causing a consequent boiling liquid expanding vapour explosion (BLEVE) with consequent overpressure fragments projection and eventually a fireball. In this work all the BLEVE consequence typologies are evaluated through theoretical and analytical models. These models are validated with the experimental results provided by the BMW care manufacturer safety tests conducted during the 1990’s. After the validation the most suitable methods are selected to perform a blind prediction study of the forthcoming LH2 BLEVE experiments of the Safe Hydrogen fuel handling and Use for Efficient Implementation (SH2IFT) project. The models drawbacks together with the uncertainties and the knowledge gap in LH2 physical explosions are highlighted. Finally future works on the modelling activity of the LH2 BLEVE are suggested.

Concentration and velocity measurements are crucial for developing and validating hydrogen jet models which  provide  scientific  bases  for hydrogen  safety  analyses.  The  concentration   fields  have  been visualized and accurately measured using laser diagnostic methods based on lase Rayleigh and Raman scattering  techniques.  However  the  velocity  measurements  are  more   challenging.  Particle  image velocimetry (PIV) has been commonly used for measuring velocities in turbulent flows by seeding tracer particles into the flow and assuming the particles intimately following the flow.  However sometimes the  particle  seeding  is  difficult  or  disturbs  the   flow.  Moreover simultaneously  concentration  and velocity measurements are very difficult when using PIV systems to measure the velocities. Therefore the optical flow velocimetry (OFV) method was used to resolve the velocity fields from the scalar fields or  particle  images  of  hydrogen   jets.  In the  present  work  the  velocity  field  and  particle  images  of hydrogen  jets   were  simulated  using  FLUENT  with  the  large  eddy simulation (LES)  model  and  the particle images were then used to resolve the velocity field by the OFV method. The OFV results were compared with the CFD simulations to verify their accuracy. The results show that the OFC method was an  efficient  low-cost  way  to  extract the  velocity  fields  from  particle  images.  The OFV  method accurately  located  the  large  vortices  in  the  flow  and  the  velocity distribution of  the  high-velocity gradients regions was consistent with the CFD results. The present study lays a foundation for using the OFV  method  to directly  resolve  the  velocity fields from  the  concentration  fields  of  hydrogen  jets measured by laser diagnostics.

Purpose- To present an overview of the research and development carried out in a European funded framework 7 (FP7) project called SafeHPower for the implementation of neutron radiography to inspect fatigue cracks in vehicle and storage hydrogen fuel tanks. Project background– Hydrogen (H2) is the most promising replacement fuel for road transport due to its abundance efficiency low carbon footprint and the absence of harmful emissions. For the mass market of hydrogen to take off the safety issue surrounding the vehicle and storage hydrogen tanks needs to be addressed. The problem is the residual and additional stresses experienced by the tanks during the continuous cyclic loading between ambient and storage pressure which can result in the development of fatigue cracks. Steel tanks used as storage containers at service stations and depots and/or the composite tanks lined with steel are known to suffer from hydrogen embrittlement (HE). Another issue is the explosive nature of hydrogen (when it is present in the 18-59% range) where it is mixed with oxygen which can lead to catastrophic consequences including loss of life. Monitoring systems that currently exist in the market impose visual examination tests pressure tests and hydrostatic tests after the tank installation [1] [2]. Three inspection systems have been developed under this project to provide continuous monitoring solutions. Approach and scope- One of the inspection systems based on the neutron radiography (NR) technology that was developed in different phases with the application of varied strategies has been presented here. Monte Carlo (MCNP) simulation results to design and develop a bespoke collimator have been presented. A limitation of using an inertial electrostatic Deuterium-Tritium (D-T) pulsed neutron generator for fast neutron radiography has been discussed. Radiographs from the hydrogen tank samples obtained using thermal neutrons from a spallation neutron source at ISIS Rutherford laboratory UK have been presented. Furthermore radiograph obtained using thermal neutrons from a portable D-T neutron generator has been presented. In conclusion a proof in principle has been made to show that the defects in the hydrogen fuel tank can be detected using thermal neutron radiography.

A non-equilibrium two-phase single-component critical (choked) flow model for cryogenic fluids is developed from first principle thermodynamics. Modern equations-of-state (EOS) based upon the Helmholtz free energy concepts are incorporated into the methodology. Extensive validation of the model is provided with the NASA cryogenic data tabulated for hydrogen methane nitrogen and oxygen critical flow experiments performed with four different nozzles. The model is used to develop a hydrogen critical flow map for stagnation states in the liquid and supercritical regions.

The CEP (Clean Energy Partnership) – Berlin is one of the most diversified hydrogen demonstration projects at present. The first hydrogen fuelling station serving 16 cars is fully integrated in an ordinary highly frequented Aral service station centrally located at Messedamm in Berlin. Hydro has supplied and is the owner of the electrolyser with ancillary systems. This unit produces gaseous hydrogen at 12 bar with use of renewable energy presently serving 13 of the cars involved. The CEP project is planned to run for a period of five years and is supported by the German Federal Government and is part of the German sustainability strategy. During the planning and design phase there have been done several safety related assessments and analyses:

• Hydro has carried out a HAZOP (HAZard and OPerability) analysis of the electrolyser and ancillary systems delivered by Hydro Electrolysers.
• Hydro arranged with support from the partners a HAZOP analysis of the interface between the electrolyser and the compressor an interface with two different suppliers on each side.
• A QRA (Quantitative Risk Assessment) of the entire fuelling station has been carried out.
• Hydro has carried out a quantitative explosion risk analysis of the electrolyser container supplied by Hydro Electrolysers.

The aim in the CEP-Berlin project was to follow the guidelines from the European Integrated Hydrogen Project phase 2 (EIHP2) [1] when planning and designing a new hydrogen fuelling station. These guidelines focus on installation and operation of hydrogen refuelling stations and were based upon best practice and knowledge from the industrial partners in EIHP 2. The explosion risk analysis for the CEP-Berlin facility demonstrated that the safety risk was acceptable if certain design modifications were carried out. Both Aral/BP and Hydro accepted the results from the quantitative risk assessments. These companies have also their own standards regarding safety acceptance criteria which have to be complied with. The fuelling station was opened November 12 2004 and was put in normal operation during winter 2005.

We present results from hydrogen dispersion simulations from a pressurized reservoir at constant flow rate in the presence and absence of a wall. The dispersion simulations are performed using a commercial finite volume solver. Validation of the approach is discussed. Constant concentration envelopes corresponding to the 2% 4% and 15% hydrogen concentration in air are calculated for a subcritical vertical jet and for an equivalent subcritical horizontal jet from a high pressure reservoir. The consequences of ignition and the resulting overpressure are calculated for subcritical horizontal and vertical hydrogen jets and in the latter case compared to available experimental data.

The manuscript firstly describes the data collection and validation process for the European Hydrogen Incidents and Accidents Database (HIAD 2.0) a public repository tool collecting systematic data on hydrogen-related incidents and near-misses. This is followed by an overview of HIAD 2.0 which currently contains 706 events. Subsequently the approaches and procedures followed by the authors to derive lessons learned and formulate recommendations from the events are described. The lessons learned have been divided into four categories including system design; system manufacturing installation and modification; human factors and emergency response. An overarching lesson learned is that minor events which occurred simultaneously could still result in serious consequences echoing James Reason's Swiss Cheese theory. Recommendations were formulated in relation to the established safety principles adapted for hydrogen by the European Hydrogen Safety Panel considering operational modes industrial sectors and human factors. This work provide an important contribution to the safety of systems involving hydrogen benefitting technical safety engineers emergency responders and emergency services. The lesson learned and the discussion derived from the statistics can also be used in training and risk assessment studies being of equal importance to promote and assist the development of sound safety culture in organisations.

Distributed hydrogen optical fiber sensor with Brillouin scattering is an innovative solution to measure hydrogen in harsh environment as nuclear industry. Glass composition is the key point to enhance the sensing parameter of the fiber in the target application. Several optical fiber with different doping element were used for measuring hydrogen saturation. Permeability of optical plays a major role to the kinetic of hydrogen diffusion. Fluorine doped fiber increase the sorption and the desorption of hydrogen.

Although many studies have looked at safety issues relating to hydrogen fuelling stations few studies have analyzed the security risks such as deliberate attack of the station by threats such as terrorists and disgruntled employees. The purpose of this study is to analyze security risks for a hydrogen fuelling station with an on-site production of hydrogen from methylcyclohexane. We qualitatively conducted a security risk analysis using American Petroleum Institute Standard 780 as a reference for the analysis. The analysis identified 93 scenarios including pool fires. We quantitatively simulated a pool fire scenario unique to the station to analyze attack consequences. Based on the analysis and the simulation we recommend countermeasures to prevent and mitigate deliberate attacks.

Hydrogen has potential applications that require larger-scale storage use and handling systems than currently are employed in emerging-market fuel cell applications. These potential applications include hydrogen generation and storage systems that would support electrical grid systems. There has been extensive work evaluating regulations codes and standards (RCS) for the emerging fuel cell market such as the infrastructure required to support fuel cell electric vehicles. However there has not been a similar RCS evaluation and development process for these larger systems. This paper presents an evaluation of the existing RCS in the United States for large-scale systems and identifies potential RCS gaps. This analysis considers large-scale hydrogen technologies that are currently being employed in limited use but may be more widely used as large-scale applications expand. The paper also identifies areas of potential safety research that would need to be conducted to fill the RCS gaps. U.S. codes define bulk hydrogen storage systems but do not define large-scale systems. This paper evaluates potential applications to define a large-scale hydrogen system relative to the systems employed in emerging technologies such as hydrogen fuelling stations. These large-scale systems would likely be of similar size to or larger than industrial hydrogen systems.

A benchmark exercise on vented explosion engineering model was carried out against the maximum overpressures (one or two peaks) of published experiments. The models evaluated are Bauwens et al. (2012-1 and 2012-2) [4 7] models Molkov Vent Sizing Technology 1999 2001 and 2008 models [12 13 6]. The experiments in consideration are Pasman et al. experiments (1974) (30% H2 - 1m3) [1] Bauwens et al. (2012) experiments (64m3) [4] Daubech et al. (2011) experiments (10 to 30% H2 - 1 and 10 m3) [2] and Daubech et al. (2013) [5] experiments (4 m3 – H2 10 to 30%). On this basis recommendations and limits of use of these models are proposed.

Experiments in an obstructed semi-confined vertical combustion channel with a height of 6 m (cross-section 0.4 × 0.4 m) inside a safety vessel of the hydrogen test center HYKA at the Karlsruhe Institute of Technology (KIT) are reported. In the work homogeneous hydrogen-air-mixtures as well as mixtures with different well-defined H2-concentration gradients were ignited either at the top or at the bottom end of the channel. The combustion characteristics were recorded using pressure sensors and sensors for the detection of the flame front that were distributed along the complete channel length. In the tests slow subsonic and fast sonic deflagrations as well as detonations were observed and the conditions for the flame acceleration (FA) to speed of sound and deflagration-to-detonation transition (DDT) are compared with the results of similar experiments performed earlier in a larger semi-confined horizontal channel.