SAFETY ASSESSMENT OF DOUBLE-WALLED LPG TANKS USING FUZZY BAYESIAN NETWORK AND IMPROVED SIMILARITY AGGREGATION METHOD
Storage tanks are crucial in safety within industrial complexes. They can lead to harm to people and financial losses during accidents. Developing a more accurate method to identify pivotal accident causes can greatly enhance the prevention strategies.
Initially, a preliminary hazard identification from the storage tank was conducted, and tank leakage event was chosen as the top event. Next, bowtie and then a Bayesian network were constructed, and relationships between nodes were input into the software. The probabilities of nodes were estimated using expert opinions and an improved similarity aggregation method. Finally, considering the top event as the evidence node, RoV values were computed to prioritize the most important events
This study identified 53 basic events. The probability of a leakage event from the LPG storage system was 1.5E-02, with the most likely hazardous outcome being pool fire.
Using RoV calculation, some basic events like inadequate connection leaks, natural disasters, structural defects in valves and concrete wall were identified as the most critical nodes. Finally, preventive and mitigative recommendations were suggested for them.
Bayesian network is known as an effective tool for modeling cause-effect and integrating this approach with fuzzy logic and improved similarity aggregation method can reduce uncertainty in studies. Additionally, their ability for probability updating enables prioritization of important events. The results of such prioritization can serve as a valuable guide for optimizing preventive maintenance activities.
-
Presenting a Business Intelligence Dashboard to Analyze the State of Safety Risks by the Predictive Risk Index (PRI) in a Pharmaceutical Company
Atefeh Ahmadabadi, Shokooh Khaloo, Reza Saeedi, *
Journal of Health and Safety at Work, -
SAFETY ASSESSMENT OF DOUBLE-WALLED LPG TANKS USING FUZZY BAYESIAN NETWORK AND IMPROVED SIMILARITY AGGREGATION METHOD
Mohamadabadi, , Reza Saeedi*
Iran Occupational Health,