Study On Bayesian Network Based Traffic Risk Assessment Of Lock Transiting

In 2003, the world witnessed the accomplishment of Three Gorges Project. It began to operate with comprehensive functions, such as water storage, power generation and navigation, which indicates that China has created a number of the Worlds-First in of the field of hydraulic engineering. As an important part of the Project, the Three Gorges double-way and five-level shiplock was also put into use. The huge investment and construction scale of the Project, including the tremendous economic and livelihood impact on the upstream and downstream people after completion, demands a quite higher safety level and stability. It is well known that the ship lock system consists of complicated structure components and passing ships. With the character of complexity, uncertainty and dynamics, the Lock has been one of the weakest parts in the safety of The Three Gorges Project. Therefore, Using the advanced theoretical methods and technology to research the accident mechanism of ship transit lock, will play very important roles not only to navigation safety, but also the safety and stability of the Three Gorges Project. In this background, the author conducts the accident causation analysis, and study the risk of ship transit lock based on Bayesian Network. The content of the paper is organized in the following four sections.Section 1 presents the background, aims of the research. To tackle the lock system's complexity, uncertainty and dynamics, the advantage of Bayesian network inference techniques are competency due to its powerful functions in expressing the complex systems and reasoning uncertain knowledge,Section 2 fully reviewes the progress of accident causation theory. Afterward, the application in the safety of vessel transiting the Lock is carried out. According to the System Engineering Theory of Human-Machine- Environment and the characteristics of lock systems, based on "man-machine-environment-management", the accident causation factor set of the accident of vessel traffic in the lock was built.Based on the principle of Bayesian Network, the third section introduces in detail the theory of the Bayesian Network, such as the method of determining the prior probability of network nodes, the learning of network parameters, network inference algorithms for structural analysis and so on, they all provide technical support for the case study of the risk analysis of the vessel traffic in the lock.The last section takes the Three Gorges ship lock as case study for the risk assessment system of ship lockage. According to the historical data analysis, accident data processing, and the network topology construction of ship lockage risk assessment system has been developed in the Three Gorges ship lock. The risk assessment system of ship lockage based on Bayesian network has been developed in VS 2008 system. With running example through the system, this paper gets the conclusion as follows:When more than one cause of the factors interacting, together, the probability of an accident will increase dramatically.