Study On Application Of Power Transformer Insulation Fault Diagnosing Expert System Of Yibin Electric Power Bureau

Considering the characteristics of insulation fault diagnosis of the power transformer of Yibin electric power bureau, this paper studies the theory and the implement scheme of the blackboard-style expert system based on the multi-expert cooperating diagnosis methods. This paper also realizes the application of the expert system.This paper studies methods of the constitution of the Knowledge-base using the Object Oriented technology. The expert knowledge is characteristic of layers and distribution. The results indicate that the methods are more efficient and reliable.This paper studies the adaptive conjunction gradient BPNN (ACG-BPNN) used in dissolved gas analysis (DGA). This method overcomes the defects in slow convergence and complex debugging. It can properly distinguish the different faults and is more accurate.This paper studies fuzzy theory used in the character gas analysis. Considering the fuzzy semantic of the diagnosis, this paper studies the subject function of ascending mountain shape and the fuzzy semantic interval to describe the experts' experience. The paper also studies the method of distinguishing the overheat position using the fuzzy relation function. The examples show that it is effective to use the fuzzy theory to diagnose the insulation faults, and it possesses high operability and precision.The results of the application of power transformer insulation fault diagnosing expert system of Yibin electric power bureau show that it is more effective to fault diagnosis of the insulation of the transformer. It can synthetically analyze the insulation using many methods, properly distinguish the different faults and point out the fault positions, the severity and the trend of the faults. The expert system possesses high precision, clear flow and operability.