Scholarlink

<p>Hydrogen accident databases contain rich descriptions of incident scenarios and consequences, but they often lack explicit physical metrics needed for quantitative fire and explosion modelling. In this work, we present a framework to reconstruct key physical parameters - such as fuel mass, heat release rate (HRR), flame length, explosion-energy potential, and notional overpressure indicators - from the qualitative and semi-quantitative descriptions in the Joint Research Centre (JRC) Hydrogen Incident and Accident Database (HIAD 2.1). By merging multiple HIAD tables and applying transparent first-order engineering relationships, we derive approximate physical descriptors that can be used for screening, comparison, and uncertainty-aware interpretation of accident records. The reconstructed dataset of 954 hydrogen-related incidents is analyzed to reveal order-of-magnitude distributions of fire sizes and explosion-energy potential. The reconstructed quantities are compared with available observations where possible, but they are not treated as high-fidelity validation data because several key parameters, including leak diameter, participating cloud mass, confinement, and ignition timing, are often missing. Monte Carlo uncertainty propagation is performed to account for data uncertainties, yielding confidence bounds on the estimated parameters. The derived physical parameters are further compared with reported incident outcomes (fatalities, damage) to explore their predictive value. Results show that the incident scenarios span several orders of magnitude in energy release, and while the framework can estimate physically reasonable order-of-magnitude parameters, direct correlations with consequences are weak - underscoring the influence of exposure, mitigating factors, and safety measures. The estimates should therefore be interpreted as reconstruction-based hazard descriptors rather than measured incident parameters; their main value is to expose the physical scale of reported events and the consequences of missing accident-reporting fields. Nevertheless, this approach provides a useful bridge between empirical accident records and physics-based fire safety interpretation. The outcomes can support screening-level comparison with engineering tools, improve risk assessments with transparent assumptions, and guide the development of safety guidelines grounded in both data and physics.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Linking Accident Data and Physics-Based Modeling for Hydrogen Fire Safety

  • 期刊论文

  • 发表日期：

    2026年7月18日

查看全文
  • Abderrahim Zermane

摘要

Hydrogen accident databases contain rich descriptions of incident scenarios and consequences, but they often lack explicit physical metrics needed for quantitative fire and explosion modelling. In this work, we present a framework to reconstruct key physical parameters - such as fuel mass, heat release rate (HRR), flame length, explosion-energy potential, and notional overpressure indicators - from the qualitative and semi-quantitative descriptions in the Joint Research Centre (JRC) Hydrogen Incident and Accident Database (HIAD 2.1). By merging multiple HIAD tables and applying transparent first-order engineering relationships, we derive approximate physical descriptors that can be used for screening, comparison, and uncertainty-aware interpretation of accident records. The reconstructed dataset of 954 hydrogen-related incidents is analyzed to reveal order-of-magnitude distributions of fire sizes and explosion-energy potential. The reconstructed quantities are compared with available observations where possible, but they are not treated as high-fidelity validation data because several key parameters, including leak diameter, participating cloud mass, confinement, and ignition timing, are often missing. Monte Carlo uncertainty propagation is performed to account for data uncertainties, yielding confidence bounds on the estimated parameters. The derived physical parameters are further compared with reported incident outcomes (fatalities, damage) to explore their predictive value. Results show that the incident scenarios span several orders of magnitude in energy release, and while the framework can estimate physically reasonable order-of-magnitude parameters, direct correlations with consequences are weak - underscoring the influence of exposure, mitigating factors, and safety measures. The estimates should therefore be interpreted as reconstruction-based hazard descriptors rather than measured incident parameters; their main value is to expose the physical scale of reported events and the consequences of missing accident-reporting fields. Nevertheless, this approach provides a useful bridge between empirical accident records and physics-based fire safety interpretation. The outcomes can support screening-level comparison with engineering tools, improve risk assessments with transparent assumptions, and guide the development of safety guidelines grounded in both data and physics.