Greece

In the present work CFD simulations of a large scale open deflagration experiment are performed. Stoichiometric hydrogen–air mixture occupies a 20 m hemisphere. Two combustion models are compared and evaluated against the experiment: the Eddy Dissipation Concept model and a multi-physics combustion model which calculates turbulent burning velocity based on Yakhot's equation. Sensitivity analysis on the value of fractal dimension of the latter model is performed. A semi-empirical relation which estimates the fractal dimension is also tested. The effect of the turbulence model on the results is examined. LES approach and k-ε models are used. The multi-physics combustion model with constant fractal dimension value equal to 2.3 using the RNG LES turbulence model achieves the best agreement with the experiment.

Hydrogen energy applications often require that systems are used indoors (e.g. industrial trucks for materials handling in a warehouse facility fuel cells located in a room or hydrogen stored and distributed from a gas cabinet). It may also be necessary or desirable to locate some hydrogen system components/equipment inside indoor or outdoor enclosures for security or safety reasons to isolate them from the end-user and the public or from weather conditions.<br/>Using of hydrogen in confined environments requires detailed assessments of hazards and associated risks including potential risk prevention and mitigation features. The release of hydrogen can potentially lead to the accumulation of hydrogen and the formation of a flammable hydrogen-air mixture or can result in jet-fires. Within Hyindoor European Project carried out for the EU Fuel Cells and Hydrogen Joint Undertaking safety design guidelines and engineering tools have been developed to prevent and mitigate hazardous consequences of hydrogen release in confined environments. Three main areas are considered: Hydrogen release conditions and accumulation vented deflagrations jet fires and including under-ventilated flame regimes (e.g. extinguishment or oscillating flames and steady burns). Potential RCS recommendations are also identified.

Correct use of Computational Fluid Dynamics (CFD) tools is essential in order to have confidence in the results. A comprehensive set of Best Practice Guidelines (BPG) in numerical simulations for Fuel Cells and Hydrogen applications has been one of the main outputs of the SUSANA project. These BPG focus on the practical needs of engineers in consultancies and industry undertaking CFD simulations or evaluating CFD simulation results in support of hazard/risk assessments of hydrogen facilities as well as on the needs of regulatory authorities. This contribution presents a summary of the BPG document. All crucial aspects of numerical simulations are addressed such as selection of the physical models domain design meshing boundary conditions and selection of numerical parameters. BPG cover all hydrogen safety relative phenomena i.e. release and dispersion ignition jet fire deflagration and detonation. A series of CFD benchmarking exercises are also presented serving as examples of appropriate modelling strategies.

In the course towards a safe future hydrogen based society one of the tasks to be considered is the investigation of the conditions under which the use or storage of hydrogen systems inside buildings becomes too dangerous to be accepted. One of the relevant scenarios which is expected to have a relatively high risk is a slow (and long lasting) hydrogen release from a vehicle stored in a closed private garage without any forced ventilation i.e. only with natural ventilation. This scenario has been earlier investigated experimentally (by M. Swain) using He (helium) to simulate the hydrogen behavior. In the present work the CFD code ADREA-HF is used to simulate three of the abovementioned experiments using the standard k-  turbulence model. For each case modeled the predicted concentration (by vol.) time series are compared against the experimental at the given sensor locations. In addition the structure of the flow is investigated by presenting the helium concentration field.

This paper summarises the results from a blind-prediction study for models developed for estimating the consequences of vented hydrogen deflagrations. The work is part of the project Improving hydrogen safety for energy applications through pre-normative research on vented deflagrations (HySEA). The scenarios selected for the blind-prediction entailed vented explosions with homogeneous hydrogen-air mixtures in a 20-foot ISO container. The test program included two configurations and six experiments i.e. three repeated tests for each scenario. The comparison between experimental results and model predictions reveals reasonable agreement for some of the models and significant discrepancies for others. It is foreseen that the first blind-prediction study in the HySEA project will motivate developers to improve their models and to update guidelines for users of the models.

The use of stationary H2 and fuel cell systems is expected to increase rapidly in the future. In order to facilitate the safe introduction of this new technology the HyPer project funded by the EC developed a public harmonized Installation Permitting Guidance (IPG) document for the installation of small stationary H2 and fuel cell systems for use in various environments. The present contribution focuses on the safety assessment of a facility inside which a small H2 fuel cell system (4.8 kWe) is installed and operated. Dispersion experiments were designed and performed by partner UNIPI. The scenarios considered cover releases occurring inside the fuel cell at the valve of the inlet gas pipeline just before the pressure regulator which controls the H2 flow to the fuel cell system. H2 was expected to leak out of the fuel cell into the facility and then outdoors through the ventilation system. The initial leakage diameter was chosen based on the Italian technical guidelines for the enforcement of the ATEX European directive. Several natural ventilation configurations were examined. The performed tests were simulated by NCSRD using the ADREA-HF code. The numerical analysis took into account the full interior of the fuel cell in order to investigate for any potential accumulation effects. Comparisons between predicted and experimental H2 concentrations at 4 sensor locations inside the facility are reported. Finally an overall assessment of the ventilation efficiency was made based on the simulations and experiments.

The different CFD tools used by the NoE HySafe partners are applied to a series of integral complex Standard Benchmark Exercise Problems (SBEPs). All benchmarks cover complementarily physical phenomena addressing application relevant scenarios and refer to associated experiments with an explicit usage of hydrogen. After the blind benchmark SBEPV1 and SBEPV3 with subsonic vertical release in a large vessel and in a garage like facility SBEPV4 with a horizontal under-expanded jet release through a small nozzle SBEPV5 covers the scenario of a subsonic horizontal jet release in a multi-compartment room.<br/>As the associated dispersion experiments conducted by GEXCON Norsk Hydro and STATOIL were disclosed to the participants the whole benchmark was conducted openly. For the purpose of validation only the low momentum test D27 had to be simulated.<br/>The experimental rig consists of a 1.20 m x 0.20 m x 0.90 m (Z vertical) vessel divided into 12 compartments partially even physically by four baffle plates. In each compartment a hydrogen concentration sensor is mounted. There is one vent opening at the wall opposite the release location centrally located about 1 cm above floor with dimensions 0.10 m (Y) times 0.20 m (Z). The first upper baffle plate close to the release point is on a sensitive location as it lies nearly perfectly in the centre of the buoyant jet and thus separates the flow into the two compartments. The actual release was a nominally constant flow of 1.15 norm liters for 60 seconds. With a 12mm nozzle diameter this corresponds to an average exit velocity of 10.17 m/s.<br/>6 CFD packages have been applied by 7 HySafe partners to simulate this experiment: ADREAHF by NCSRD FLACS by GexCon and DNV KFX by DNV FLUENT by UPM and UU CFX by HSE/HSL and GASFLOW by FZK. The results of the different participants are compared against the experimental data. Sensitivity studies were conducted by FZK using GASFLOW and by DNV applying KFX.<br/>Conclusions based on the comparisons and the sensitivity studies related to the performance of the applied turbulence models and discretisation schemes in the release and diffusion phase are proposed. These are compared to the findings of the previous benchmark exercises.

Computational Fluid Dynamics (CFD) has already proven to be a powerful tool to study the hydrogen dispersion and help in the hydrogen safety assessment. In this work the Large Eddy Simulation (LES) recently incorporated into the ADREA-HF CFD code is evaluated against the INERIS-6C experiment of hydrogen leakage in a supposed garage which provides detailed experimental measurements visualization of the flow and availability of previous CFD results from various institutions (HySafe SBEP-V3). The short-term evolution of the hydrogen concentrations in this confined space is examined and comparison with experimental data is provided along with comments about the ability of LES to capture the transient phenomena occurring during hydrogen dispersion. The influence of the value of the Smagorinsky constant on the resolved and on the unresolved turbulence is also presented. Furthermore the renormalization group (RNG) LES methodology is also tested and its behaviour in both highly-turbulent and less-turbulent parts of the flow is highlighted.

Hydrogen impinging jets can be formed in the case of an accidental release indoors or outdoors. The CFD simulation of hydrogen impinging jets suffers from numerical errors resulting in a non-physical velocity and hydrogen concentration field with a butterfly like structure. In order to minimize the numerical errors and to avoid the butterfly effect high order schemes need to be used. The aim of this work is to give best practices guidelines for hydrogen impinging jet simulations. A number of different numerical schemes is evaluated. The number of cells which discretize the source is also examined.

The time and space evolution of the distribution of hydrogen in confined settings was investigated computationally and experimentally for permeation from typical compressed gaseous hydrogen storage systems for buses or cars. The work was performed within the framework of the InsHyde internal project of the HySafe NoE funded by EC. The main goal was to examine whether hydrogen is distributed homogeneously within a garage like facility or whether stratified conditions are developed under certain conditions. The nominal hydrogen flow rate considered was 1.087 NL/min based on the then current SAE standard for composite hydrogen containers with a non-metallic liner (type 4) at simulated end of life and maximum material temperature in a bus facility with a volume of 681m3. The release was assumed to be directed upwards from a 0.15m diameter hole located at the middle part of the bus cylinders casing. Ventilation rates up to 0.03 ACH were considered. Simulated time periods extended up to 20 days. The CFD simulations performed with the ADREA-HF code showed that fully homogeneous conditions exist for low ventilation rates while stratified conditions prevail for higher ventilation rates. Regarding flow structure it was found that the vertical concentration profiles can be considered as the superposition of the concentration at the floor (driven by laminar diffusion) plus a concentration difference between floor and ceiling (driven by buoyancy forces). In all cases considered this concentration difference was found to be less than 0.5%. The dispersion experiments were performed at the GARAGE facility using Helium. Comparison between CFD simulations and experiments showed that the predicted concentrations were in good agreement with the experimental data. Finally simulations were performed using two integral models: the fully homogeneous model and the two-layer model proposed by Lowesmith et al. (ICHS-2 2007) and the results were compared both against CFD and the experimental data.