Research On The Risk Mechanism Of Maritime Accidents

Maritime traffic safety has always been the most important part of maritime transportation,and it is also the focus of maritime shcolars at home and abroad.Accident causation mechanism and accident severity are two important contents in the study of maritime accidents.At present,studies on maritime accidents mainly focus on accident risk assessment,accident causation analysis,accident probability prediction,etc.,but studies on the coupling effect of risk factors and spatial distribution characteristics of maritime accidents are obviously insufficient,and there are few studies on accident severity and injury severity.In the above context,this thesis mainly carried out the following aspects of work.(1)Study on the spatial distribution characteristics of global maritime accidents.Based on the global maritime accident data and geographic information system,a general framework for the the spatial analysis of maritime accidents was constructed.In this framework,spatial density analysis was adopted to study the spatial distribution of maritime accidents in terms of frequency,and accident-prone sea areas were identified.Sea areas with high relative accident density were also identified after considering the influence of ship traffic density.Hotspot analysis and outlier analysis were used to study the spatial distribution of maritime accidents in terms of severity,and high-severity accident clustering areas and low-severity accident clustering areas were identified.In addition,buffer analysis was used to get the distribution proportion of maritime accidents with various severity in buffer zones.(2)Study on the risk coupling effect of maritime accidents.On the basis of systematically sorting out the influencing factors of maritime accidents and their coupling relationship,the NK model was applied to construct the risk coupling model of maritime accidents.Based on the risk coupling model,the coupling risk values of human factors,ship factors and environment factors under different coupling forms were calculated,and the risk coupling law of maritime accident was obtained.On this basis,the coupling degree model and N-K model were combined according to the principle of LEC method to calculate the total coupling risk of maritime traffic system.The influence of risk factors on the total coupling risk was calculated through sensitivity analysis of single index,and the key risk factors of marine accidents were identified.(3)Study on the severity of maritime accidents.Based on the global maritime accident data,an analysis model for the severity of maritime accidents was established by using the ordered logistic regression method,and the relationship between the severity of maritime accidents and influencing factors was analyzed.At the same time,based on the large amount of data obtained from the questionnaire survey,the relationship between the severity of maritime accidents and the influencing factors were evaluated by using fuzzy rule-based Bayesian network and evidence reasoning.Both methods identified the factors that are significantly related to the severity of mairitime accidents,and the results of the two methods are highly consistent.(4)Study on the injury severity of maritime accidents.In view of the zero inflated phenomenon and excessive dispersion characteristics of injury data in maritime accidents,the ordered probit model and zero-inflated ordered probit model were respectively used to construct the injury severity analysis model of maritime accidents,and the injury data of global maritime accidents were analyzed.Factors significantly associated with the injury severity of maritime accidents were identified.The comparison between the two models shows that the zero-inflated order probit model has better fitting performance in dealing with zero-inflated and over-discrete injury data.Through the systematic study of maritime accidents,this thesis revealed the spatial distribution characteristics and risk coupling rules of maritime accidents,and identified the influencing factors significantly associated with maritime accident severity and injury severity.The results of this thesis are of great significance for better control and management of the overall risk of maritime accidents.