Safety

Hydrogen is expected to play a key role in the decarbonised future of energy. For hydrogen distribution pipelines are seen as the main method for mass transport of hydrogen gas. To support the evaluation of risk related to hydrogen pipelines a revised QRA methodology is presented based on currently available and industry accepted guidance related to natural gas. The QRA approach is primarily taken from HSE UK’s MISHAP methodology [1]. The base methodology is reviewed and modifications suggested to adapt it for use with hydrogen gas transport. Compared to natural gas it was found that the escape distances for hydrogen (based on the degree of heat flux) were lower. However as for the overall risk for both individual and societal the case with hydrogen was more severe close to the pipeline. This was driven by the increased ignition probability of hydrogen. The approach may be used as part of the review and appraisal process of hydrogen projects

In 2021 only 6 hydrogen fuelling station have been built in Italy of which 3 are not operational and only 1 is open to the public while the rest are built in private or industrial areas. While fuelling station which store more than 5000 kg of hydrogen are subjected to the “Seveso Directive” the permitting procedure for refuelling station which store less than the threshold is supervised by the fire brigade command of the province where the station is built. Recently in the effort to easy the permitting procedure to establish new stations a Ministerial Decree was published in the official gazette of the Italian Republic which lists minimum safety features and safety distances that if respected guarantee the approval by the authority. Nevertheless the imposed distances are such that the land required to build the station constitute a barrier rather than a facilitation. Exploiting the possibility introduced by the Decree to calculate safety distances following a Fire Safety Engineering approach a method is proposed for calculation of safety distances. The present paper presents the Italian regulation and describes an approach to calculate the safety distances including an example applied on the dispenser.

The use of hydrogen storage tanks at 100% of nominal working pressure NWP is expected only after refuelling. Driving between refuellings is characterised by the state of charge SoC<100%. There is experimental evidence that Type IV tanks tested in a fire at initial pressures below one-third of its NWP depending on a fire source were leaking without rupture. This paper aims at understanding this phenomenon and the development of a predictive model. The numerical research has demonstrated that the heat transfer from fire through the composite overwrap is sufficient to melt the polymer liner. This initiates hydrogen microleaks through the composite wall before it loses the load-bearing ability when the resin degrades deep enough to cause the tank to rupture. The dependence of tank fire-resistance rating (FRR) on the SoC is presented for tanks of volume in the range 36-244 L. The tank wall thickness non-uniformity i.e. thinner composite at the dome area is identified as a serious issue for tank’s fire resistance that must be addressed by tank manufacturers and OEMs. The effect of the burst pressure ratio on FRR is investigated. It is concluded that thermal parameters of the composite wall i.e. decomposition heat and temperatures play a vital role in simulations of tank failure and thus FRR.

Unexpected hydrogen leakage may occur in the production storage transportation and utilization of hydrogen. The lower flammability limit (LFL) for the hydrogen is 4% in air. The combustion and explosion of hydrogen-air mixture poses potential hazards to personnel and property. In this study unintended release of hydrogen from a hydrogen fuel cell forklift vehicle inside a enclosed warehouse is simulated by fireFoam which is an LES Navier-Stokes CFD solver. The simulation results are verified by experimental data. The variation of hydrogen concentration with time and the isosurface of hydrogen concentration of 4% vol. are given. Furthermore the leakage of hydrogen from a storage tanks in a hydrogen refueling station is simulated and the evolution of the isosurface of hydrogen concentration of 4% vol. is given which provides a quantitative guidence for determination the hazardous area after the leakage of hydrogen.

The aim of this study is to assess numerically the influence of terrain landscape on the distribution of probable harmful consequences to personnel of hydrogen fueling station caused by an accidentally released and exploded hydrogen. In order to extract damaging factors of the hydrogen explosion wave (maximum overpressure and impulse of pressure phase) a three-dimensional mathematical model of gas mixture dynamics with chemical interaction is used. It allows controlling current pressure in every local point of actual space taking into account complex terrain. This information is used locally in every computational cell to evaluate the conditional probability of such consequences on human beings as ear-drum rupture and lethal ones on the basis of probit analysis. In order to use this technique automatically during the computational process the tabular dependence ""probit-functionimpact probability"" is replaced by a piecewise cubic spline. To evaluate the influence of the landscape profile on the non-stationary three-dimensional overpressure distribution above the earth surface near an epicenter of accidental hydrogen explosion a series of computational experiments with different variants of the terrain is carried out. Each variant differs in the level of mutual arrangement of the explosion epicenter and the places of possible location of personnel. Two control points with different distances from the explosion epicenter are considered. Diagrams of lethal and ear-drum rupture conditional probabilities are build to compare different variants of landscape profile. It is found that the increase or decrease in the level of the location of the control points relative to the level of the epicenter of the explosion significantly changes the scale of the consequences in the actual zone around the working places and should be taken into account by the risk managing experts at the stage of deciding on the level of safety at hydrogen fueling stations.

The study aims at the development of a safety engineering methodology for the assessment of flame length after full-bore rupture of hydrogen pipeline. The methodology is validated using experimental data on hydrogen jet flame from full-bore pipeline rupture by Acton et al. (2010). The experimental pressure dynamics in the hydrogen pipeline system is simulated using previously developed adiabatic and “isothermal” blowdown models. The hydrogen release area is taken as equal similar to the experiment to doubled pipeline cross-section as hydrogen was coming out from both sides of the ruptured pipe. The agreement with the experimental pressure decay in the piping system was achieved using discharge coefficient CD=0.26 and CD=0.21 for adiabatic and “isothermal” blowdown model respectively that indicates significant friction and minor pressure losses. The hydrogen flame length was calculated using the dimensionless correlation by Molkov and Saffers (2013). The correlation relies on the density of hydrogen in the choked flow at the pipe exit. The maximum experimental flame length between 92 m and 111 m was recorded at 6 s after the pipe rupture under the ground. The calculated by the dimensionless correlation flame length is 110 m and 120 m for the “isothermal” and adiabatic blowdown model respectively. This is an acceptable accuracy for such a large-scale experiment. It is concluded that the methodology can be applied as an engineering tool to assess flame length resulting from ruptured hydrogen pipelines.

Hydrogen safety issues attract focuses increasingly as more and more hydrogen powered vehicles are going to be operated in traffic infrastructures of different kinds like tunnels. Due to the confinement feature of traffic tunnels hydrogen deflagration may pose a risk when a hydrogen leak event occurs in a tunnel e.g. failure of the hydrogen storage system caused by a car accident in a tunnel. A water injection system can be designed in tunnels as a mitigation measure to suppress the pressure and thermal loads of hydrogen combustion in accident scenarios. The COM3D is a fully verified three-dimensional finite-difference turbulent flow combustion code which models gas mixing hydrogen combustion and detonation in nuclear containment with mitigation device or other confined facilities like vacuum vessel of fusion and semi-confined hydrogen facilities in industry such as traffic tunnels hydrogen refueling station etc. Therefore by supporting of the European HyTunnel-CS project the COM3D is applied to simulate numerically the hydrogen deflagration accident in a tunnel model being suppressed by water mist injection. The suppression effect of water mist and the suppression mechanism is elaborated and discussed in the study.

Austenitic stainless steels are used for hydrogen containment of high-pressure hydrogen gas due to their ability to retain high fracture properties despite the degradation due to hydrogen. Forging and other strain-hardening processes are desirable for austenitic stainless steels to increase the material strength and thus accommodate higher stresses and reduce material costs. Welding is often necessary for assembling components but it represents an area of concern in pressure containment structures due to the potential for defects more environmentally susceptible microstructure and reduced strength. Electron beam (EB) welding represent an advanced joining process which has advantages over traditional arc welding techniques through reduced input heat and reduced heat-affected zone (HAZ) microstructure and thus present a means to maintain high strength and improve weld performance in hydrogen gas containment. In this study fracture coupons were extracted from EB welds in forged 304L and subjected to thermal gaseous hydrogen precharging at select pressures to introduce different levels of internal hydrogen content. Fracture tests were then performed on hydrogen precharged coupons at temperatures of both 293 K and 223 K. It was observed that fracture resistance (JH) was dependent on internal hydrogen concentration; higher hydrogen concentrations resulted in lower fracture resistance in both the forged 304L base material and the 304L EB welds. This trend was also apparent at both temperatures: 293 K and 223 K. EB weld samples however maintain high fracture resistance comparable to the forged 304L base material. The role of weld microstructure solidification on fracture is discussed.

The present work performed within the PRESLHY EC-project presents a simplified 1-d transient modelling methodology to account for discharge line effects during blowdown. The current formulation includes friction extra resistance area change and heat transfer through the discharge line walls and is able to calculate the mass flow rate and distribution of all physical variables along the discharge line. Choked flow at any time during the transient is calculated using the Possible Impossible Flow (PIF) algorithm. Hydrogen single phase physical properties and vapour-liquid equilibrium are calculated using the Helmholtz Free Energy (HFE) formulation. Homogeneous Equilibrium Mixture (HEM) model is used for two-phase physical properties. Validation is performed against the new experiments with compressed gaseous hydrogen performed at the DISCHA facility in the framework of PRESLHY (200 bar ambient and cryogenic initial tank temperature 77 K and 4 nozzle diameters 0.5 1 2 and 4 mm) and an older experiment at 900 bar ambient temperature and 2 mm nozzle. Predictions are compared against measured data from the experiments and the relative importance of line heat transfer compared to flow resistance is analysed.

In water-cooled nuclear power plants hydrogen gas can be generated by various mechanisms during an accident. In case combustion of the resulting hydrogen-air mixture within the facility occurs existing containment structures may be compromised and excessive radio-active material can be released to the environment. Thus an improved understanding of the propagation of lean hydrogen deflagrations within buildings and structures is essential for the development of appropriate accident management strategies associated with these scenarios. Following the accident in Fukushima Japan the application of three-dimensional computational fluid dynamics methods to high-fidelity detailed analysis of hydrogen combustion processes in both closed and vented vessels has become more widespread. In this study a recently developed large-eddy-simulation (LES) capability is applied to the prediction of lean premixed hydrogen deflagrations in vented vessels. The LES methodology makes use of a flamelet- or progress-variable-based combustion model coupled with an empirical burning velocity model (BVM) an anisotropic block-based adaptive mesh refinement (AMR) strategy an accurate finite-volume numerical scheme and a mesh independent subfilter-scale (SFS) model. Several different vessel and vent sizes and configurations are considered herein. The LES predictions are compared to experimental data obtained from the Large-Scale Vented Combustion Test Facility (LSVCTF) of the Canadian Nuclear Laboratories (CNL) with both quiescent and turbulent initial conditions. Following descriptions of the LES models LES results for both variable chamber sizes and single- and double-vent cases are presented to illustrate the capabilities of the proposed computational approach. In particular the predicted time histories of pressure as well as the maximum overpressure achieved within the vessels and combustion compartments are compared to those from the LSVCTF experiments. The influences of the modelled ignition process initial turbulence and mesh resolution on the LES results are also discussed. The findings highlight the potential and limitations of the proposed LES approach for accurately describing lean premixed hydrogen deflagrations within vented vessels.