STheBaN - system-theoretic bayesian approach for the evaluation of inspections workability in hydrogen operations

The demand for hydrogen is expected to grow significantly in the coming years. Despite its versatility and sustainability, hydrogen presents notable safety concerns due to its flammability, ignitability, and tendency to leak and embrittle metallic materials. To ensure the reliability of hydrogen technologies, regular inspection activities are essential. However, standards for designing, operating, and inspecting these components are still under development. Additionally, the complexity of the human-machine couplings involved can compromise the success of an inspection. In this context, this study proposes an innovative methodology to evaluate how human, technical, and environmental factors impact the workability of hydrogen inspections operations. The methodology combines the System-Theoretic Accident Model and Processes (STAMP) and Bayesian Networks to assess the complexity of human performance during acoustic emission testing of hydrogen storage tanks. Unlike traditional sequential methods, this approach employs STAMP to structure a Bayesian Network, enabling a comprehensive understanding of operator-process interactions. Prior probabilities, derived from Human Reliability Analysis, are incorporated into the Bayesian Network, facilitating quantitative assessments of inspection success probabilities. The study reveals that tasks may become unsafe under conditions that adversely affect the quality of the inspection. Furthermore, the operating environment, experience of operators, and their safety attitudes are critical for successful inspections. The adaptability of this methodology makes it a valuable tool for assessing human performance in their operations within complex socio-technical systems and safety-critical applications.