The Research Of Knowledge Acquisition And Representation In Fault Diagnosis Expert System Of Signal Devices Of Railway

It is urgent to study on the important problem, in order to make sure driving safely, that realize supervising the state of signal devices in railway and diagnosing faults, by new technique and method. Because of the complicated structure of signal control devices in railway station and the effect of the devices usage and their running environment, it makes that the reason of devices fault is complicated, random and ambiguous. Therefore, it is significant to develop the corresponding fault diagnosis applied expert system.Knowledge acquisition and representation are bottleneck problems in building expert system. They are also hot spot in developing applied system. According to the Railway Ministry project: Intelligent System of Fault Diagnosis and Analysis of the Railway Signal Devices, upheld by my tutor professor Zhangxi, this paper study mainly on the method of knowledge acquisition and representation based on circuit simulation and object-oriented technique.Based on the comprehensive analysis of internal and external present research, this paper firstly studies on the model structure and the method of knowledge acquisition based on circuit simulation technique, including customer demanding analysis, collecting and dealing with original material, testability analysis, fixing fault test pool, establishing simulation model and its analysis, quantifying signs and confidence analysis, then put forward a scheme of knowledge acquisition based on fault mechanism simulation. Secondly, this paper makes researches on knowledge representation model and its method based on object-oriented technique, including the principle and model structure of object-oriented technique, the representation way of objects, rules and methods. Finally, according to the requirement of DES, this paper makes a test on a concrete example simulation analysis, design and its database realization. The test results indicate that this method can realize knowledge auto-acquisition. By its characters of module, block, inheritance and multi-state, it is also able to make complex knowledge present in a simple way, keeping the solution space and the question space having a concurrent structure. So that it can enhance the system capability in knowledge acquisition, representation and organization.