Research On Fault Diagnosis Approach Based On Signed Directed Graph For Thermal System

There are many problems in fault diagnosis process of complex thermal system, such as multiple influencing factors of causing fault, multiplicity of symptom, strong coupling among faults and so on, some facility abnormality of sub-system may lead to the parameters abnormality of the correlative facility. So in the diagnosis process, the fault sources can be located effectively only when the fault propagation characteristic among the different sub-system and thermal facility can be sufficient considered.The deep knowledge model of Signed Directed Graph (SDG) can express complicated cause-effect relationship, has the capacity of containing large-scale potential information and good completeness in the process industrial system. On the basis of summarizing the existing literatures of fault diagnosis based on SDG method, This paper mainly researches some difficulties and key technologies, such as SDG modeling, SDG simplification, SDG reasoning rules, threshold value modification of SDG, integration and transform of qualitative and quantitative model of SDG and so on. The above research works improve the system of graph theory and fault diagnosis method. And according to the characteristic of thermal system of power plant, SDG is first presented to apply in fault diagnosis of thermal system. The mainly research results are the followings:Aiming at the localization of the above traditional modeling method in actual modeling process, considering that the simulator can not only provide high-fidelity dynamic simulation model of the thermal system, but also provide the over-all analysis environment using the method of system engineering, a SDG grading modeling method aided by simulation technology is studied,which is an effective method to solve the problem of"information explode"and provides a very effective path to solve the difficult problem of SDG modeling in complex system.To improve the diagnostic speed of SDG diagnosis process, the idea of combining SDG with expert systems and the way to produce the diagnostic rules from SDG are described. Furthermore, to improve the SDG diagnosis resolution, SDGs are simplified by the following ways, (1) adopting discrete state to depict the continuous process to deal with the inconsistency in the branch during transient; (2) adopting the expert experience and simulation testing to find the dominant effect branches of SDG in the fault propagation; (3) adopting the variable transformation of the controller output to...