Dynamic Hazard Identification on SOFC system using Bayesian Network

Abstract

Accidents are expected when operating SOFC unit system in a plant due to its complexity and operating conditions. SOFC system which consists of risky components such as combustor, reformer, heaters and SOFC stack poses risk of fire and explosion especially due to its high operating temperature. In addition, other factors such as failure rate components, quantity materials, gas leakage and chemical characteristics involved further increase the risks to an alarming level. In reality, these conditions are evolving depending on the current situation which make it challenging in determining the actual risks. Since SOFC technology is still emerging and not widely used, the studies on hazard identification on SOFC system is very minimal. The present work develops a new hazard evolution framework for SOFC system which is mapped into Bayesian network model using open-source software programs, GeNie to bring dynamics in identifying risks and hazards. This allow all potential hazards to be updated in real-time to ensure safe implementation of SOFC unit system in a plant. With this, all factors and evolving conditions contributing to the risks can be estimated with higher precisions to reduce the accidents probability. Sensitivity analysis is also carried out to determine how input parameters influencing the identified hazards. Results showed the probability of fire and explosion occurring in SOFC system is approximately 21% and 7% respectively. Operating conditions (temperature and pressure) are identified as the main causes contributing to the risks. Higher temperature increases risks of Fire from 17% to 21% while higher pressure increases the risks of Explosion from 7% to 18%. The current work identified the occurrence of final hazards in SOFC system dynamically and can be served as guideline for safer implementation of SOFC.