Research Of Fault Diagnosis Technology To The Marine Electric Propulsion System

With the coming practical stage of permanent magnet synchronous motor, marineelectric propulsion system technology is having been developed fast. The electric propulsionsystem is one of the most important part of a ship, so the normal operation of the system isvery significant for the ship’s safety and fight security. Therefore, how to diagnose and detectthe fault of the marine electric propulsion quickly and accurately becomes one of theimportant contents of ship maintenance. Although fault diagnosis can reduce the loss andincrease the economic benefits and has been widely used in various fields, the fault diagnosisof the marine electric propulsion system is few both at home and abroad. So it’s verysignificant to research the marine permanent magnet synchronous motor electric propulsionsystem fault diagnosis technology. The research work of this paper mainly includes thefollowing aspects:(1) The structure of the marine electric propulsion system is analyzed and each modulesand working principles are introduced briefly. The fault characteristics of the main parts ofthe system are analyzed. The fault tree and expert system are combined to diagnose thesystem fault. The general structure of the marine electric propulsion system fault diagnosissystem is built.(2) The basic concept, the building principles and steps of the fault tree are introducedbriefly. The fault tree model of the marine electric propulsion system is built. According tothe characteristics of the system, the fault tree and the analytic hierarchy process are used tocalculate the weighting of the fault tree. The fault tree with weighting is formed. The order ofthe fault removal is decided by the weighting.(3) The marine electric propulsion system fault diagnosis system based on fault tree-expert system is built. The production representation is used to extract rules from the faulttree of the marine electric propulsion system. The knowledge base of the fault diagnosissystem is built. The hybrid reasoning way and the depth and width search strategy are used tobuild the reasoning machine of the fault diagnosis expert system.(4) According to the shortcoming that traditional fault diagnosis expert system is lack oflearning ability, the BP neural network technology is used to diagnose the system fault. TheBP neural network built through the fault diagnosis samples training and simulation is used todiagnose the fault. (5)An electric propulsion ship is used as the diagnosis project to test the diagnosis abilityof two kinds of fault diagnosis system. The result of simulation indicates that the systemshave the good fault diagnosis abilities and are well designed to meet requirements.