A risk assessment framework for water electrolysis systems: Mapping System Theoretic Process Analysis (STPA) and Event Tree Analysis (ETA) into Fuzzy Bayesian Networks (FBN)

• Published in: Process safety and environmental protection 2025-02, Vol.194, p.306-323
• Main Authors: Zhu, Taolin, Meng, Chaoyue, Han, Xuzeng, Wang, Yaqi, Dang, Jing, Chen, Hui, Qi, Meng, Zhao, Dongfeng
• Format: Article
• Language: English
• Subjects: Fuzzy Bayesian Network (FBN); Hydrogen leakage; Proton exchange membrane water electrolysis (PEMWE); System Theoretic Process Analysis (STPA)
• ISSN: 0957-5820
• DOI: 10.1016/j.psep.2024.11.117

Risk assessment plays a vital role in facilitating the safety and sustainability of green hydrogen production. This study presents a risk assessment method incorporating causal qualitative analysis of System Theoretic Process Analysis (STPA) and Event Tree Analysis (ETA), and quantitative analysis of FBN. Based on the flexible adaptation of the typical STPA analysis process, the topology of STPA-ETA-based BN with hydrogen leakage as the critical event is established. It shows systematization, modularity and comprehensibility. The fuzzy Set Theory (FST) and expert judgment were adopted to solve the problem of inaccessible probabilities. The results of the prediction analysis indicate that the probability of hydrogen leakage is 12.63 %. The diagnostic analysis is capable of assessing the event's contribution to the hydrogen leakage accident, rating factors such as pump failures, flow control valve failures, hydrogen embrittlement, valve group manipulation risks and management risks, and environmental factors. The consequence analysis indicates that a timely manual emergency shutdown barrier plays a vital role in preventing the escalation of incidents. Finally, addressing the current diversity of sensitivity analysis methods in BN, the article summarizes six sensitivity analysis methods and explores their interrelationships in depth. Two categories of sensitivity analysis dimensions were found. The results of this study provide valuable insights for risk assessment, incident prevention, and emergency management in the PEMWE system. Also, the framework proposed in this study can be extended to other areas of reliability analysis.