Risk assessment for berthing of hazardous cargo vessels using Bayesian networks

Abstract

Life and property losses along environmental pollution caused by accidents in the transportation of hazardous goods have raised serious concerns over the efficient handling of hazardous cargo vessels in a port environment. Very few studies have explicitly aimed the dynamic risk associated with hazardous cargos in a ship berthing scenario. Corresponding to these concerns, this study is aimed to assess the risks concomitant to the berthing of hazardous cargo vessels. The methodology adopted for this study is the combination of binary logistics regression & expert judgment for the identification of various influencing factors, and the Bayesian networks environment for the inference and analysis of these factors. The results indicate that under normal conditions the hazardous cargo vessel risk has a comparatively lower probability of 3.97, but its risk score indicates that it cannot be ignored and requires attention to be solved. Setting evidence at hazardous cargo accident, human and environment proves to be the most significant contributors. Environmental factor is the most prominent factor, which at its full effect, raises the accident probability to 14.91 with a risk score of 9.17 which is a highly dangerous class II risk requiring immediate solution. This study indicates that training along the mental & psychological state of ship crew, wind force, water velocity, channel width, berth layout and port location are the significant factors that the authorities need to focus on for invulnerable berthing of a hazardous cargo vessel. This study has prominent practical viability for conducting risk management and planning for hazardous cargo vessels.