| 英文摘要 | Fault Detection And Diagnosis is very important for a maintainable system to operate. The purpose of a fault detection and diagnosis is: when a maintainable system can not operate normally because of faults occurred, determine the fault position and cause, SUPPLY the treatment measurement as soon as possible. Improve the availability of the system by means of reducing the maintain time. Automatic control system plays a considerable role in industry. In the production process, the fault will effect the control object, sometime this effect is critical. The task of fault detection and diagnosis is difficult, but it is significant for the system operate reliability, safety and benefit. Up to now, there are no reports of effective research and application of fault detection and diagnosis. The author combines the automatic control system reliability engineering and the artificial intelligence knowledge engineering, synthesizes the theory and the technology of these two engineering courses, research the theory and the application of computerizing fault detection and diagnosis of the automatic control system, and proposes the design and the application method. The content of this thesis is: system reliability analysis and synthesis, fault model determination, fault data base and knowledge base building method, fast inference method, imperfect information for fault determination, diagnosis and treatment, etc. According to these results, the power plant water supply control system fault detection and diagnosis expert system is designed. in this thesis, two kinds of expert system method are introduced: one is based on the mathematic and physical model, the other is based on the system reliability engineering model. In view of the second method, the author proposed three ways to build knowledge base: fault chain inductive method, logical tree deductive method, decision table method. By means of reliability analysis, master the fault mode, cause and consequence, get the fault qualitative and quantitative description, depend on the experience of the experts and the technicists, make the treatment. The corresponding inference method is Bayes Formula -- Vector Analysis two steps inference method. The first step is to build a sub-base from the knowledge base, the second step is to ~ain conclusion from the sub-base. Based on the 300MW power plant automatic control system, the author proposed the basic principle and the design process of the fault detection and diagnosis expert system. The author combines the reliability engineering and the knowledge engineering, researches the computerizing fault detection and diagnosis, advances the reliability engineering, and initiate the new field to applicate the expert system. |
修改评论