JOURNAL OF ETA MARITIME SCIENCE, cilt.5, sa.4, ss.396-413, 2017 (Hakemli Dergi)
The safety and risk assessment of dangerous cargo operations in oil and chemical tankers is a necessary process to prevent possible accidents during these operations. Fire and explosion are the major accidents encountered in tanker operations. In this study, a model was constructed through the Fuzzy Bayes Network Method for the probabilistic relationships between the causes of fire and explosion accidents that could occur during the tank cleaning process. The study is composed of two stages. Firstly, the variables that are the subject of the problem and that constitute the graphical structure of the Bayes Networks are identified. Then, expert opinion was sought as the statistical data on accident reports were not recorded properly while identifying the conditional probability of the relationships between the variables. Linguistic variables whose fuzzy membership functions were identified were used in detecting the probabilities. The findings of the sensitivity test revealed that the major reasons that could lead to fire and explosion during the tank cleaning process are ignition sources, reaction and safety culture.