Pollution prevention of vessels in the greater bay area: A practical contribution of port state control inspection system towards carbon neutralisation using a tree augmented naive bayes approach

Abstract

Vessel emissions are one of the major emissions sources affecting carbon neutralisation achievement in the Greater Bay Area (GBA). The vessel pollution prevention ability is effective in preventing vessels from discharging mass emissions. Thus, it is of great value to figure out the vessel's performance in pollution prevention in this area. This study develops a data-driven Bayesian network model focusing on vessel performance in pollution prevention in port state control (PSC) inspections in the GBA. The Tree Augmented Naïve and Expectation Maximisation approaches are applied to learn structures and parameters based on the inspection records obtained from January 2015 to September 2022. An analysis can identify key variables with significant effects on vessel performance in pollution prevention and clarify the positive influence and contributions of PSC inspections toward carbon neutralisation achievement. Practical implications are provided to foster port authorities and local governments in the GBA to minimize the vessels with poor pollution prevention performance sailing in this area. The study also identifies the characteristics of vessels with poor performance and proposes valuable suggestions to restrict the inattentive actions of ship owners and stimulate ship owners to put more effort into the improvement of pollution prevention ability. Further, the study reveals the significance of executing efficient and reliable PSC inspections in promoting carbon neutralisation progress.