Evaluation of the factors causing container lost at sea through fuzzy-based Bayesian network

Abstract

To date, many high-profile accidents involving the loss of large numbers of containers overboard have occurred during ocean voyages. It is estimated that an average of 1629 containers are lost at sea each year, and the statistics also highlight that the number of lost shipping containers has increased by two-thirds in the last five years. It is important to consider that falling containers are not only a threat to the safety of shipping but also a potential health and environmental hazard. In this direction, there are ongoing initiatives to improve container handling and reduce the risk of container loss. This paper focuses on container loss at sea, which is a serious risk in terms of marine pollution and financial loss, and seeks to evaluate this problem by identifying its causes. A Fuzzy Bayesian Network (FBN) model is therefore created to provide an approach for determining the factors and weights to be considered by seafarers during container transportation when assessing the causes of a container falling overboard. As a result of the research process, a model for assessing the risk of container falls is developed and the probabilistic relationships between the causes of container falls overboard are revealed. According to the analysis of the model, improper stuffing (15%), misdeclaration of container weight (12.6%), and container structural resistance (11.1%), are the three most risky root causes of container falls. In addition, the most obvious finding to emerge from this study is that the causes of container losses overboard are strictly related to both the process of the lashing and securing of the cargo and to the stability of the ship. Taken together, the results indicate that to prevent container falls at sea, it is crucial to avoid poor stability and ensure proper cargo securing. Consequently, FBN model developed for the study is expected to help improve the safety of container transport and reduce the risk of container loss by providing a comprehensive probabilistic risk analysis of container operations and predicting the risk of container loss if undesirable factors arise.