The rapid deployment of hydrogen technologies is creating growing demand for safety methodologies capable of supporting innovative designs while maintaining acceptable risk levels. Although originally developed to support hydrogen-fuelled seagoing vessels, the methodology described in the “Handbook for Hydrogen-Fuelled Vessels” provides a transferable framework for risk-informed design of both maritime and land-based hydrogen systems. Central to this approach is the integration of Quantitative Risk Assessment (QRA) with Computational Fluid Dynamics (CFD), enabling safety decisions to be based on the actual physical behaviour of hydrogen rather than on conservative assumptions and modelling alone.

 This presentation discusses practical experience gained from hydrogen safety assessments performed for actual projects, including the “Dhamma Sea” hydrogen vessel study and a hydrogen-fuelled bulk carrier risk assessment (termed as “ZeroCoaster”). These projects applied a structured methodology combining hazard identification, leak and ignition frequency analysis, CFD-based dispersion and explosion modelling, sensitivity and risk assessment techniques to evaluate design alternatives and demonstrate equivalence with conventional fuel systems. The work highlights how hydrogen's unique properties, including high diffusivity, wide flammability range, low ignition energy, and high explosion potential, require modifications to traditional risk assessment approaches while remaining fully aligned with established QRA practices.

 The presentation will demonstrate how CFD-derived consequences are integrated into QRA frameworks to quantify both individual and societal risk, support explosion risk assessments, evaluate mitigation measures, and optimize safety-critical design features such as ventilation, gas detection, emergency shutdown systems, isolation philosophy, and hazardous area management. Practical examples will illustrate how experimental hydrogen knowledge, validated modelling tools, and emerging regulatory requirements can be combined to support robust engineering decisions under uncertainty.