The Research And The Application Of The Ship Fault Diagnosis System Based On The N- DFBN Method

Marine system is a high coupling, between each device to connect with each other, the main components are widely distributed system integration of complex system. Once appear, equipment failure, just rely on the crew according to their experience and professionalism to troubleshooting and maintenance, it is difficult to quickly and efficiently in the limited time to repair, This is likely to lead to shipment delayed, property loss and even threatens the safety of staff life on ship, especially in the ocean, the loss will be immeasurable. Ship in the event of equipment failure, therefore, use the right method to quickly find the cause of the problem and make relevant maintenance is particularly important.This article first studied the status and the basic method of the ship fault diagnosis, the basic principle of Bayesian networks, the inference method and the learning method of Bayesian network. In the study, this article concluded that the Bayesian network in fault diagnosis can effectively deal with many uncertainty factors of ship equipment failure, absorb a priori information, and improve the accuracy of fault diagnosis. The fault tree in the application of fault diagnosis of the ship is relatively perfect, and through the analysis of fault tree to build Bayesian network method is more convenient. In this paper, the innovation points and work content mainly embodied in the following content:1.In this paper, in combination the advantage of the Fault tree to the advantage of Bayesian networks in the ship fault diagnosis, this paper proposes a ships fault diagnosis method based on N- DFBN, and applies the method to reasoning the fault of the ship’s deck lighting system.The main idea of the fault diagnosis method based on N-DFBN is to use fault tree and Bayesian network combination way to establish a diagnosis fault DFBN Bayesian network.This kind of diagnostic method in Marine system such large equipment integration system not only improved the efficiency of recognition, but also can carry on the rational analysis to the system of specific performance.2.In order to apply the ship fault diagnosis based on N-DFBN method to the actual ship fault diagnosis,this paper design the ship fault diagnosis system based on N-DFBN method used as the core of the paper,and realizes the key part of the system.Design content mainly includes the process design system based on the actual demand function, system architecture design, the main subsystem design, implementation of part mainly includes the IETM troubleshooting and maintenance, remote maintenance support part and the interaction of remote data processing part. To make N- DFBN method based diagnosis system can in actual ship to play a bigger role in fault diagnosis, this paper also gives a composite system deployment scheme of the actual demand.