| Zhoushan sea area is located at the intersection of China’s North-South shipping routes and the Yangtze River estuary.The scattered distribution of thousands of islands in the Zhoushan sea area not only provides many excellent ports,but also has complex hydrological and navigation conditions.As the starting point for shipping in the Yangtze River Delta region,the shipping industry and port industry are thriving,with a wide range of ships passing by.Under the comprehensive influence of ships,environment and other factors,maritime traffic accidents occur frequently in the sea area,resulting in huge personal and property losses.In order to help maritime traffic participants understand the basic situation of maritime traffic safety in Zhoushan sea area scientifically and provide reference for the formulation of maritime traffic management policies,this thesis,based on historical accident examples,uses mathematical models to study the causes and prediction of maritime traffic accidents in Zhoushan sea area.Through sorting and analyzing the historical data of maritime traffic accidents in Zhoushan sea area,the main characteristics of historical accidents in the sea area are explored,and then combined with previous research conclusions,four types of factors affecting maritime traffic accidents in Zhoushan sea area are obtained by qualitative analysis.On the premise that the data can be quantified,six factors that are likely to lead to maritime traffic accidents in the sea area are selected according to the above analysis,and the correlation degree between the six factors and the number of accidents is calculated by using the grey correlation analysis method.Finally,the three factors with the highest correlation degree are selected as the impact indicators of accident prediction.Since the occurrence of maritime traffic accidents is the result of the interaction of multiple factors,the OBGM(1,N)model which can take into account multiple factors is selected to predict the number of accidents.Then,in order to further improve the prediction accuracy,the OBGM(1,N)model is optimized by using particle swarm optimization algorithm and the residual correction method based on buffer operator.By calculating the prediction accuracy under different background values and comparing with other commonly used accident number prediction models,the advantages and practicability of the model are shown.The results show that:(1)There are obvious characteristics of maritime traffic accidents in Zhoushan sea area.In terms of accident types,collision of ships is the main type of accident.Geographically,accidents occur in the vicinity of "three vertical and four horizontal" waterways and ports.Timely,accidents often occur in spring and autumn,which are susceptible to fog.In addition,accidents often occur at night due to fatigue,sleepiness and poor vision.From the point of ship type,the main ship types of accidents are Zhoushan registered ship,dry goods ship and fishing vessel.(2)The causes of maritime traffic accidents in the Zhoushan Sea mainly include such factors as inadequate crew manning,illegal operation of ships on the sea,poor meteorological conditions and inadequate safety supervision and control by ship operating units.(3)The correlation between the number of ships,registered crew members,and productive berths and the number of accidents indicates that they have a significant impact on the occurrence of accidents and can be used as accident impact indicators to predict the number of accidents.(4)The prediction accuracy of the OBGM(1,N)model optimized by the particle swarm optimization algorithm is significantly improved,and the residual correction method based on the buffer operator is also useful to improve the accuracy of the model.The prediction method used can be effectively used to predict maritime traffic accidents in Zhoushan sea area. |
