Quantitative risk assessment for maritime safety management

Abstract

There are over 130,000 ships sailing around the world, these playing an important role in the world's economic development. However, shipping has always been characterized as a relatively risky business. With an increasing awareness of environmental protection and safety issues, research into maritime risk assessment (MRA) has become an important research domain. In this research, the following tasks are accomplished in order to develop a quantitative risk assessment for maritime safety management. First, this research summarizes previous studies into risk assessment by way of a literature review. It was found that the traditional and simplest way to estimate the probability of marine accidents is to consider accident statistics or expert estimation. However, both of these methods have certain limitations. Secondly, this research is based on the safety performance of global vessels and has found various risk indicators that can be used to indicate the probability of an accident. Following this, multivariate logistic regression was used to measure the probability of the occurrence of an accident through historical data on safety indicators, such as vessel age, type, registration and classification. Thirdly, a comprehensive database was built for this analysis. The availability of suitable data necessary for each step of the FSA process is very important. Fourthly, this study presents an innovative approach toward integrating logistic regression and a Bayesian Network together into risk assessment. This approach has been developed and applied to a case study in the maritime industry. It can apply to other industries as well. Finally, a case study, which applying this risk assessment approach in the port state control program, is presented. The optimal inspection policy with regard to different vessel types, as well as the effects of changing parameters on the optimal inspection rate, is examined in the light of the risk assessment results.