Safety

In recent years the use and development of hydrogen as a carbon-free energy carrier have grown. However as hydrogen is flammable with air safety issues are raised. In the case of ignition especially in confined space the flame can accelerate and reach the detonation regime causing severe structural damage [1].<br/>To assess these safety issues it is required to understand the fluid-structure interaction in the case of a detonation impacting a deformable structure and to quantify and model this interaction [2]. At the CEA (Commissariat à l’énergie atomique et aux energies alternatives) a combustion tube experimental facility [3] for studying the fluid-structure interaction in the case of hydrogen combustion has been developed. Several Photomultipliers and Pressure sensors are placed along the tube to monitor the flame acceleration and the detonation location. A fluid-structure interaction (FSI) module or a non-deformable flange can be placed at the end of the tube. Post-processing of the sensor’s signal will provide insight into the occurring phenomena inside the tube.<br/>Several experimental campaigns have been conducted with various initial conditions and configurations at the end of the tube. In this contribution the experiments resulting in a detonation are presented. First the recorded pressure and velocities will be compared to theoretical values coming from combustion models [4] [5]. Secondly the impulse before and after reflection for thin plate and non-deformable flange will be compared to quantify the energy transmitted to the plate and the influence of the fluid-structure interaction on the reflected shock.

During a severe accident in a nuclear power plant one of the potential threats to the containment is the occurrence of energetic combustion events. In modern plants Severe Accident Management Guidelines (SAMG) as well as dedicated mitigation hardware are in place to minimize/mitigate this combustion risk and thus avoid the release of radioactive material into the environment. Advancements in SAMGs are in the focus of AMHYCO an EU-funded Horizon 2020 project officially launched on October 1st 2020. The project consortium consists of 12 organizations (from six European countries and one from Canada) and is coordinated by the Universidad Politécnica de Madrid (UPM). The progress made in the first two years of the AMHYCO project is here presented. A comprehensive bibliographic review has been conducted providing a common foundation to build the knowledge gained during the project. After an extensive set of accident transients simulated both for phases occurring inside and outside the reactor pressure vessel a set of challenging sequences from the combustion risk perspective for different power plant types were identified. At the same time three generic containment models for the three considered reactor designs have been created to provide the full containment analysis simulations with lumped parameter models 3-dimensional containment codes and CFD codes. In order to further consolidate the model base combustion experiments and performance tests on passive auto-catalytic recombiners under explosion prone H2/CO atmospheres were performed at CNRS (France) and FZJ (Germany). Finally it is worth saying that the experimental data and engineering models generated from the AMHYCO project are useful for other industries outside the nuclear one.

The maritime industry is gaining momentum towards a more decarbonized and sustainable path. However most of the worldwide fleet still relies on fossil fuels for power producing harmful environmental emissions. Hydrogen as a clean fuel is a promising alternative but its unique properties pose significant safety challenges. For instance hydrogen has a wide flammability range inherently increasing the risk of ignition. Moreover its comparatively low volumetric energy density necessitates faster filling rates and larger volumes for bunkering and onboard storage leading to higher risk rates. Therefore the use of hydrogen for maritime applications requires the development of specialized riskbased approaches according to safety engineering principles and techniques. The key safety implications are discussed and reviewed with focus on onboard hydrogen storage handling and refueling while a priority-based Failure Mode and Effects Analysis (FMEA) method for risk assessment is proposed based on the revised guidelines of Automotive Industry Action Group (AIAG) and German Association of the Automotive Industry (VDA). The revised AIAG-VDA FMEA method replaces the conventional Risk Priority Number (RPN) with a new Action Priority (AP) rating enabling the prioritization of recommended actions for risk reduction. The paper aims to a more profound understanding of the safety risks associated with hydrogen as a maritime fuel and to provide an effective risk assessment method for hydrogen applications onboard maritime vessels.

CFD modelling of liquified hydrogen boiling and evaporation during the pressurised tank venting is presented. The model is based on the volume-of-fluid method for tracking liquid and gas phases and Lee’s model for phase change. The simulation results are compared against the liquid hydrogen evaporation experiment performed by Tani et al. (2021) in a large-scale pressurised storage tank using experimental pressure dynamics and temperatures measured in gas and liquid phases. The study focuses on tank pressure decrease and recovery phenomena during the first 15 s of the venting process. The model sensitivity have been studied applying different Lee’s model evaporisation-condensation coefficients. The CFD model provided reasonable agreement with the observed pressure and gas phase temperature dynamics during the liquid hydrogen storage depressurisation using Lee’s model coefficient =0.05 s-1. Experimentalists’ hypothesis about particularly intensive boiling in the proximity of thermocouples was supported by close agreement between simulated and experimental saturation temperatures obtained from pressure dynamics.

The widespread production and use of hydrogen (H2) requires safe handling due to its wide range of flammability and low ignition energy. In confined and semi-confined areas such as garages and tunnels a hydrogen leak will create a potential accumulation of flammable gases. Hence forced ventilation is required in such confined spaces to prevent hydrogen hazards. However this practice may incur higher operating costs and could become ineffective during a power outage. Passive Autocatalytic Recombiners (PARs) are defined as safety devices for preventing hydrogen accumulation in confined spaces. PARs have been widely adopted for hydrogen mitigation in nuclear containment buildings in worst case accident scenarios where forced ventilation is not feasible. PARs are equipped with catalyst plates that self-start due to hydrogen reacting with oxygen at relatively low concentrations (<2 vol. % H2 in air). The heat generated from the reaction creates a self-sustained flow continuously supplying the catalyst surface with fresh hydrogen and oxygen. In this study a 2D transient numerical model has been developed in COMSOL Multiphysics to simulate the operation of PARs. The model was used to analyze the effect of surface reactions on the catalyst temperature flow dynamics self-start behaviour forced versus natural convective flow and steady-state hydrogen recombination rates. The model was also used to simulate carbon monoxide poisoning and its influence on the catalyst performance. Experimental data were used for model calibration and validation showing good agreement for different conditions. Overall the model provides novel insights into PARs operation such as radiation and poisoning effects on the catalyst plate. As a next step assessment of the effectiveness of PARs is underway to mitigate hydrogen hazards in selected confined and semi-confined areas including nuclear and non-nuclear applications.

Shock-flame interactions (SFI) occur in a variety of combustion scenarios of scientific and engineering interest which can distort the flame extend the flame surface area and subsequently enhance heat release. This process is dominated by Richtmyer-Meshkov instability (RMI) that features the perturbation growth of a density-difference interface (flame) after the shock passage. The main mechanism of RMI is the vorticity deposition results from a misalignment between pressure and density gradients. This paper focuses on the multi-dimensional interactions between shock wave and flame in a hydrogen-air mixture. The simulations of this work were conducted by solving three-dimensional fully-compressible reactive Navier-Stokes equations using a high-order numerical method on a dynamically adapting mesh. The effect of wall friction on the SFI was examined by varying wall boundary condition (free-slip/no-slip) on sidewall. The results show that the global flame perturbation grows faster with the effect of wall friction in the no-slip case than that in the free-slip case in the process of SFI. Two effects of wall friction on SFI were found: (1) flame stretching close to the no-slip wall and (2) damping of local flame perturbation at the no-slip wall. The flame stretch effect leads to a significantly higher growth rate in the global flame perturbation. By contrast the damping effect locally moderates the flame perturbation induced by RMI in close proximity to the no-slip wall because less vorticity is deposited on this part of flame during SFI.

In August 2022 Shell started construction of Holland Hydrogen I (HH I) a 200 MW electrolyser plant in the port of Rotterdam’s industrial zone on Maasvlakte II in the Netherlands. HH I will produce up to 60000 kg of renewable hydrogen per day. The development and demonstration of a safe layout and plant design had been challenging due to ambitious HH I project premises many technical novelties common uncertainties in hydrogen leak effect prediction a lack of large-scale water electrolyzer operating history and limited standardization in this industry sector. This paper provides an industry perspective of the major challenges in commercial electrolyzer plant HSSE risk assessment and risk mitigation work processes required to develop and demonstrate a safe design and it describes lessons learned in this area during the HH I project. Furthermore the paper lists major common gaps in relevant knowledge engineering tools standards and OEM deliverables that need closure to enable future commercial electrolyzer plant projects to develop an economically viable and plant design and layout more efficiently and cost-effectively.

Cryogenic liquid is often stored in a vacuum insulated Dewar vessel for a high efficiency of thermal insulation. Multi-layer insulation (MLI) can be further applied in the double-walled vacuum space to reduce the heat transfer from the environment to the stored cryogenic fluid. However in loss-of-vacuum accident (LOVA) scenarios heat flux across the MLI will raise to orders of magnitudes larger than with an intact vacuum shield. The cryogenic liquid will boil intensively and pressurize the vessel due to the heat ingress. The pressurization endangers the integrity of the vessel and poses an extra catastrophic risk if the vapor is flammable e.g. hydrogen. Therefore safety valves have to be designed and installed appropriately to make sure the pressure is limited to acceptable levels. In this work the dynamic process of the heat and mass transfers in the LOVA scenarios is studied theoretically. The mass deposition - desublimation of gaseous nitrogen on cryogenic surfaces is modeled as it provides the dominant contribution of the thermal load to the cryogenic fluid. The conventional heat convection and radiation are modeled too although they play only secondary roles as realized in the course of the study. The temperature dependent thermal properties of e.g. gaseous and solid nitrogen and stainless steel are used to improve the accuracy of calculation in the cryogenic temperature range. Presented methodology enabling the computation of thermodynamic parameters in the cryogenic storage system during LOVA scenarios provides further support for the future risk assessment and safety system design.

In international regulations on explosion protection mechanical friction impact or abrasion is usually named as one of 13 ignition sources that must be avoided in hazardous zones with explosive atmospheres. In different studies it is even identified as one of the most frequent ignition sources in practice. The effectiveness of mechanical impacts as ignition source is dependent from several parameters including the minimum ignition energy of the explosive atmosphere the properties of the material pairing the kinetic impact energy or the impact velocity. By now there is no standard procedure to determine the effectiveness of mechanical impacts as ignition source. In some previous works test procedures with poor reproducibility or undefined kinetic impact energy were applied for this purpose. In other works only homogeneous material pairings were considered. In this work the effectiveness of mechanical impacts with defined and reproducible kinetic impact energy as ignition source for hydrogen containing atmospheres was studied systematically in dependence from the inhomogeneous material pairing considering materials with practical relevance like stainless steel low alloy steel concrete and non-iron-metals. It was found that ignition can be avoided if non-iron metals are used in combination with different metallic materials but in combination with concrete even the impact of non-iron-metals can be an effective ignition source if the kinetic impact energy is not further limited. Moreover the consequence of hydrogen admixture to natural gas on the effectiveness of mechanical impacts as ignition source was studied. In many cases ignition of atmospheres containing natural gas by mechanical impacts is rather unlikely. No influence could be observed for admixtures up to 25% hydrogen and even more. The results are mainly relevant in the context of repurposing the natural gas grid or adding hydrogen to the natural gas grid. Based on the test results it can be evaluated under which circumstances the use of tools made of non-iron-metals or other non-sparking materials can be an effective measure to avoid ignition sources in hazardous zones containing hydrogen for example during maintenance work.

Due to the small characteristic molecular size of hydrogen small leaks are more common in hydrogen systems compared to similar systems with hydrocarbons. This together with the high reactivity makes an efficient ventilation system very important in hydrogen applications. There are several models available for ventilation sizing that are based on either a well-mixed assumption or a fully stratified situation. However experiments show that many realistic releases will be neither and therefore additional models are needed. One possibility is to use CFD-models but the small release sizes for pinhole releases (<<1 mm) make it difficult to find an appropriate mesh without excessive computational time (especially since the simulations need to be iterated to find the optimum ventilation size). An alternative approach which is described and benchmarked in the current paper is to use a multi-zone model where the domain is divided into several large cells where the mass exchange is simplified compared to CFD and thus simulation time is reduced. The flow in the model is governed by mass conservation and density differences due to concentration gradients using the Bernoulli equation. The release of gas generates a plume which is modelled based on an empirical plume model which gives the entrainment and hydrogen source term for each cell. The model has a short run time and will therefore allow optimization in a short time frame. The model is benchmarked against five experiments with helium at the Canadian Nuclear Laboratories (CNL) in Canada and one hydrogen experiment performed at Lodz University of Technology in Poland. The result shows that the model can reasonably well reproduce accumulation in the experiments with small release without ventilation but appears to slightly underestimate the level of stratification and the interface height for ventilated cases where the source is elevated from the floor level.