Research Of Fault Diagnosis Expert System Based On Agent And Fault Tree

The fault diagnosis expert system is a development direction of the fault diagnosis technology. The design of the system, technology which is used, knowledge base and the design of inference engine will be the primary problems to be solved in the fault diagnosis expert system.Combined with the Provincial Natural Science Funds Project:Research on Key Technology of Train Fault Diagnosis Expert System for Passenger Dedicated Lines (Item Number:20102014), depended on the characteristics of electrical system failure of electric locomotive, some related technologies of developing fault diagnosis expert system are researched, and architecture design and developing method of fault diagnosis expert system are analyzed. According to the failure characteristics of the electrical system of electric locomotive, the status of the current fault diagnosis technology is chosen for the background, and the primary problems which are solved by fault diagnosis expert system are the final goal. Combining with the knowledge in the field of artificial intelligence technology, computer technology and fault diagnosis technology, a highly versatile and well reliable fault diagnosis expert system is designed. The functions and tasks of all the Agents which are the component units of the system are designed. The traditional expert system is transformed, so that it has the characteristics of Agent. The access to the knowledge based on fault tree model in this system is elaborated. The conventional stored way of rules is changed into three-level storage mechanism:diagnosis phenomena, diagnosis reason and diagnosis conclusion. The system is changed on the basis of fault tree model, combining with the knowledge of data structures, the binary fault tree model is proposed and on this basis, the inference mechanism of fault diagnosis expert system is designed and realized.The fault diagnosis expert system is applied to electrical system of the electric locomotive, as the example of fault trees in the electrical system, the design details and complete process of system establishment of fault diagnosis expert system based on Agent and fault tree are elaborated, and the simulation results of system are analyzed. The results prove that the Agent and fault tree used in the field of fault diagnosis can effectively improve the work efficiency of the diagnosis tasks. Good scalability of the system can be easily upgraded to accommodate to the new demands, and the high versatility of system allows the system to apply to other fields and equipment fault diagnosis without too many changes.