| With the implementation of the planning in the Guangdong,Hongkong and Macao grand bay area and the strategy of the Nansha free trade zone and the construction of the Guangzhou international shipping center,it becomes more and more important to ensure the safety of the ships in Guangzhou port.The occurrence of ship traffic accidents has a process,and there are various kinds of information in the process of appearing to the critical state.Through these information,the risk degree of ship navigation can be judged.If the risk early-warning is carried out in time,it can avoid the transformation from dangerous state to accident state.It is of great practical significance to reduce the incidence of ship traffic accidents and improve the safety level of ship traffic in port waters.In order to realize the objective of monitoring and timely prevention of the dangerous state of ship navigation in Guangzhou port and turn the ship safety supervision work in front of the accident,this paper has carried out the research on the risk early-warning of ship traffic in Guangzhou port for making the ship safety supervision work from passive to initiative and from experience to science.The main research contents include the following four parts:(1)The dynamic analysis index is used to analyze the traffic accident distribution law in Guangzhou port,and the appropriate indexes are extracted as the ship traffic risk early-warning indicators in Guangzhou port.Through the dynamic analysis of the time distribution,spatial distribution and morphological distribution of ship traffic accidents in Guangzhou port,the dynamic change characteristics of the accident distribution law with time are obtained.(2)The clustering method is applied to analyze the cause of the ship traffic accidents in Guangzhou port,and the typical factors are selected as the ship traffic risk early-warning indicators in Guangzhou port.Based on the actual accident statistics data of Guangzhou port over the past years,the formation of the causes of the ship traffic accidents is identified,and four kinds of influence factors,such as the factors of human,ship,environment and management,are extracted from the factors.The selection of typical factors is carried out by the system clustering method,which can not only consider the influence of various factors and save the cause of important influence,but also reduce the number of factors and eliminate some factors which are not typical,similar in character or high in collinearity.(3)The ship traffic risk early-warning index system in Guangzhou port is established.According to the establishment principle of early-warning index system and the research results of accident analysis,the risk early-warning index of ship traffic in Guangzhou port is determined as the two aspects of the analysis index of ship traffic accident distribution law and the analysis index of accident cause.The importance and validity of the early-warning index are confirmed by Delphy Fa.And the early-warning index system of ship traffic in Guangzhou port is established.(4)Back Propagation neural network method is used to build the ship traffic risk early-warning model in Guangzhou port.The statistical data of related early-warning indicators of typical ship traffic accidents case in Guangzhou port is selected to train the BP neural network in order to construct the early-warning model.The new sample data are selected to verify the model,and it is proved that the model has high prediction accuracy and can improve the accuracy and reliability of ship traffic risk early-warning in Guangzhou port.The ship traffic risk early-warning index system based on accident analysis and the ship traffic risk early-warning model based on BP neural network are practical and innovative.It can provide information for the ship traffic safety management in Guangzhou port and ensure the safe and smooth navigation of ships in the harbor waters.And it can improve ship traffic safety level and safety management capacity in port waters. |
