13. Scenario-based accident data review and analysis

Hydrogen as an energy carrier provides a feasible solution to reduce greenhouse gas emissions, and hence achieve the goal of controlling global average temperature rise to no more than two degrees Celsius according to the Paris agreement. Successful transition to a hydrogen economy can provide a link between the generation of renewable electricity and sectors where carbon emissions are hard to abate (International Renewable Energy Agency, 2018[1]).

However, as a new energy form, most countries do not yet have a defined hydrogen strategy nor have hydrogen included in their regulatory policy frameworks. This report develops and presents a detailed, scenario-based review of two hydrogen incident & accident databases, HIAD 2.0 and H2tools, to provide more insights into the nature of potential safety consequences from the following scenarios indicated as of particular interest by the Dutch Ministry of Economy and Climate Policy:

  • Scenario 1 – Production: Leakage from pipes connected to electrolysers

  • Scenario 2 – Transport pipelines: Leakage from high-pressure pipeline

  • Scenario 3 – Road Transport: Hydrogen leakage in confined spaces/ built environments

  • Scenario 5 – Mobility and partially confined spaces: accidents at a hydrogen refuelling stations

An additional section on storage is also included, as it potentially relates to Scenarios 1, 3 or 5.

The report does not map Scenario 4 (Mobility and partially confined spaces: e.g., a hydrogen city bus driving in a tunnel involved in a collision and Scenario 6 (Domestic use: safety of hydrogen in buildings with focus on hydrogen boilers), as there are no recorded accidents related to these scenarios.

The analyses in this report are performed in a scenario-specific manner, allowing us to draw scenario-specific conclusions about the risks involved with each technology. This goes a step further than more general hydrogen accident analyses available in the literature. Furthermore, normalised accident rates are calculated for each scenario, which are then compared to accident rates in similar hydrocarbon-based industries, in order to determine whether hydrogen is more hazardous than currently used fuels. The following comparisons are made:

  • Scenario 1 – Normalised fatality rates caused by hydrogen production compared to fatality rates caused by energy production from other sources, such as coal, oil and natural gas

  • Scenario 2 – Leakage rate from hydrogen pipelines compared to natural gas pipelines

  • Scenario 3 – Hydrogen leakage rates from hydrogen-powered vehicles compared to liquefied petroleum gas (LPG) vehicles

  • Scenario 5 – Accident rates at hydrogen refuelling stations compared to LPG refuelling stations

In general, the reported hydrogen accidents are typical of the types and range of accidents occurring in conventional hydrocarbon-based industry sectors. The calculated normalised accident rates suggest that the use of hydrogen in each of the four scenarios is as safe, and even safer, than the fuel currently used in comparable industries, when the proper precautions are taken. As the accident causes are the same or broadly similar it can be concluded that: (a) Existing knowledge and good practices and safety measures can be applied to the technology deployed in the four scenarios analysed and (b) No new or special solutions are required to make the use of hydrogen sufficiently safe for the intended use.

The report reviews 2 incident databases, the Hydrogen Incidents and Accidents Database (HIAD) and H2tools. The incidents recorded in both HIAD 2.0 and the H2Tools database are hydrogen-related incidents that either resulted in, or had the potential to result in hydrogen leakage.

The Hydrogen Incidents and Accidents Database (HIAD) was initially created as part of the HySafe project (2004-09), a research project supported by the European Commission, and was populated by the members of the HySafe network. The network consisted of 25 partners across 12 countries who shared their respective expertise in the automotive, chemical, gas and oil and nuclear fields with the aim of developing mitigation methods that facilitate the safe and efficient introduction of hydrogen technologies (HySafe, 2007[2]). Since the end of the project, HIAD has been populated by the Joint Research Centre of the European Commission (EC-JRC). It is an open communication platform that compiles publicly available data on international hydrogen-related incidents and accidents. A new, more streamlined, version of the database, HIAD 2.0, has been in development since 2016. This new version is more focused on identifying and sharing lessons learnt from hydrogen-related incidents, as well as other useful information about hydrogen system safety. HIAD 2.0 contains 628 entries as of December 2021. The entries are largely compiled from 10 smaller incident databases.1

The H2Tools database was created with the support of the US Department of Energy, with the aim of assisting in the spread of important information and lessons from incidents during the use of hydrogen (H2Tools, n.d.[3]). The database is a hydrogen incident reporting tool which any individual can use to provide information on incidents or near-misses involving hydrogen. The events are anonymised to encourage the reporting of any incidents. As of May 2022, the database contains 221 entries, 67 of the most relevant are examined for this report.

Chapter 14 – Results and discussion presents an overview of the databases in terms of the number of accidents recorded and human & social consequences in terms of the number of injuries & deaths. The following subsections provide in-depth analysis of the recorded accidents mapped to 4 key scenarios – Scenarios 1, 2, 3 and 5.

In addition to the scenarios, a separate subsection on hydrogen storage is presented, since the safe storage of hydrogen is an important consideration in many applications. Hydrogen storage can be potentially linked to Scenarios 1, 3 and 5, since in most instances storage is required during its production, transportation and at hydrogen refuelling stations before dispensing. Finally, the knowledge drawn from the database review is summarised and scenario-specific analyses to evaluate the risk and consequences associated with the use of hydrogen are presented.

A total of 266 incidents and near-misses, reported in HIAD 2.0 and the H2tools database, were studied for the purpose of this report, as they were relevant to applications of interest. These accidents were sorted out based on their relevance to scenarios 1, 2, 3 and 5 out of the six scenarios of interest, with no relevant accidents having been reported for scenarios 4 and 6. Normalised incident rates were calculated for the different scenarios which were then compared to incident rates in similar hydrocarbon-based industries. Through this comparison, it was determined that, based on the information provided by the databases, the incident rate caused by the use of hydrogen in scenarios 1, 2 and 5 is lower than in the cases where a comparable hydrocarbon fuel is used. In the case of scenario 3, data suggests that LPG vehicles are currently safer than hydrogen-powered vehicles, likely due to the difference in the maturity of the two technologies. However, most of the recorded incidents involving hydrogen vehicles were traffic accidents caused by external factors and therefore there were no novel causes that resulted in these incidents.

The hydrogen incidents were further analysed based on their physical consequences, as well as their reported root causes. For the majority of the incidents, there were multiple interconnected contributing causes. For example, equipment failure was frequently caused by inadequate maintenance and/or deficiency in procedure. In such cases, only the main cause of the incident was considered. Overall, 153 of the accidents studied (58%) were more severe, resulting in a fire or explosion, while the other 113 incidents (42%) resulted in unignited hydrogen release or no hydrogen release at all. It may be worth noting that especially in the case of voluntary self-reporting, severe hydrogen specific incidents may be over represented.

The decline in the number of casualties and injuries caused by hydrogen accidents in recent years demonstrates that the production and use of hydrogen is relatively safer than before, although there is more room for improvement.

  • In accidents connected to hydrogen production (scenario 1), the hydrogen compressor was identified as the most risky component.

  • Most accidents are related to equipment failure, it is therefore recommended to evaluate guidance on the design life of critical components.

The use of high-pressure pipelines for hydrogen transport (scenario 2) is not yet wide-spread, a fact that is reflected by the small number of hydrogen incidents involving hydrogen pipelines. Furthermore, small hydrogen leaks from pipes might not be detected and thus not reported.

It is important that hydrogen pipelines are maintained by competent personnel who are aware of the proper maintenance procedures.

Warning labels at regular space intervals to indicate the presence of underground hydrogen pipelines is highly recommended to avoid excavation works that could lead to pipe damage.

  • Most accidents related to scenario 3 are traffic accidents involving either vehicles transporting hydrogen or vehicles powered by hydrogen. As vehicles powered by hydrogen are not yet very common, most accidents relate to hydrogen transportation.

  • The high number of traffic accidents highlights the importance of proper training for the drivers. It also highlights how necessary it is that the drivers are always alert and in good physical conditions.

  • Most incidents at hydrogen refuelling stations (scenario 5) involved equipment failure, with the most common being dispenser & compressor failure.

  • The compressors as well as stand-by machines should be maintained regularly. Specifically, a research piece on the common faults of hydrogen compressors (Han et al., 2020[4]) suggests regular checks & cleaning of the lubricating oil system, air valves, cylinder blocks and crankshafts (ranked by fault frequency).

  • Accidents involving hydrogen storage have become less common in recent years, as the safety regulations have become stricter, however, accidents can still occur if the proper safety procedures are not enforced.

  • Ensuring that the correct protocols for hydrogen storage and handling are followed is vital, as well as ensuring that all relevant personnel are suitably trained.

The main causes of process safety loss of containment incidents in the chemical industry are a combination of technical, organisational and human failures which are well-documented and understood and taken into account when designing new process equipment and installations (COMAH competent authority, 2011[5]), (HSE, 2017[6]), (Lisbona, 2022[7]), (Wishart, Chowdhury and Ayeni, 2022[8]). As this report has shown, the causes of the hydrogen accidents recorded in the databases are typical of the types and range of accidents, which occur in the conventional hydrocarbon-based industry sectors. In other words, there are no novel hydrogen accidents when it comes to causation. Based on this observation we can conclude that hydrogen installations and equipment will suffer from the same types of failure and at similar frequencies as has been previously observed in the industry. The consequences of the accidents may vary slightly but not significantly.

As such, it can be concluded that existing knowledge and good practice to safeguards that is currently in place can be applied to the technology deployed in the four example scenarios and no new or uncommon solutions are required to make the use of hydrogen sufficiently safe for the intended use.

HIAD 2.0 is a database collecting systematic data on hydrogen-related incidents, accidents or near misses. The database combines information on accidents from a range of sources that collect data at the national or regional level (Table 13.1).


[5] COMAH competent authority (2011), Annual Operational Intelligence Report 2010, CDOIF Meeting on 2 November 2011.

[3] H2Tools (n.d.), About H2 Tools, https://h2tools.org/about (accessed on 1 June 2023).

[4] Han, Y. et al. (2020), “The Common Faults and Analysis of Hydrogen Compressor”, IOP Conference Series: Earth and Environmental Science, Vol. 440/3, https://doi.org/10.1088/1755-1315/440/3/032055.

[9] HIAD 2.0 (2018), , https://hysafe.info/hiad-2-0-free-access-to-the-renewed-hydrogen-incident-and-accident-database/.

[6] HSE (2017), Failure Rate and Event Data for use within Risk Assessments, https://www.hse.gov.uk/landuseplanning/failure-rates.pdf.

[2] HySafe (2007), The EC Network of Excellence for Hydrogen Safety “HySafe”, http://www.hysafe.org/.

[1] International Renewable Energy Agency (2018), “Hydrogen from renewable power: technology outlook for the energy transition”, https://www.irena.org/-/media/files/irena/agency/publication/2018/sep/irena_hydrogen_from_renewable_power_2018.pdf.

[7] Lisbona, D. (2022), Analysis of a loss of containment incident dataset for major hazards intelligence using Storybuilder, Draft.

[8] Wishart, J., A. Chowdhury and K. Ayeni (2022), 100 largest losses in the hydrocarbon industry, 27th edition, Marsh Specialty.


← 1. HySafe (28 June 2018), “HIAD 2.0- free access to the renewed hydrogen incident and accident database” https://hysafe.info/hiad-2-0-free-access-to-the-renewed-hydrogen-incident-and-accident-database/ (accessed on 16th May 2023).

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

© OECD 2023

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at https://www.oecd.org/termsandconditions.