Use of Bayesian networks for emergency assistance assessment of ship capsizing

Paper in proceeding, 2017

Accident caused by a ship’s capsize is one of the hottest spots in the public security due to their sudden and severe consequences, such as Eastern Star and Sewol accidents happened in the past few years. To improve the ability of emergency handling, e.g. Search And Rescue (SAR) operations in inland waters to waterborne transport accidents, a SAR evaluation model for inland ships is proposed based on the Bayesian network technique and an expert survey. The model starts with the statistical analysis of historical ship capsize data from 1998 to 2010. The statistics and expert knowledge are synthesized in the Bayesian network model to obtain the probability distribution of the consequences. Then, several indicators that influence the consequences are identified based on case study and sensitivity analysis, e.g. navigation environment, self-rescue ability and emergency disposal etc. Finally, the proposed Bayesian network model is used as a tool for emergency assistance assessment aiming to mitigate the consequences due to ship capsizing.