The Research On Fault Diagnosis And Maintenance Decision Support System Of Electric Power Equipment

In this thesis, the evolutive status about electric power equipment maintenance and condition based maintenance strategy at home and abroad is summarized. Whereafter,taking transformer for the typical equipment, the intelligent fault diagnosis and condition based maintenance decision support system is consummated on the basis of thesis 1 and thesis 2. The main contents are as follows:1,The evolutive status about electric power equipment fault diagnosis and Condition Based Maintenance technique at home and abroad is introduced. The selective method about power equipment maintenance strategy is expounded, which based on foult mode features. The fundamental and implementary process about condition Based maintenance are discussed. Then the ameliorative system structure about transformer intelligent fault diagnosis and condition based maintenance decision support system is proposed in this thesis.2,ID3 decision tree is applied to the transformer fault diagnosis model in thesis 2, which only constructs univariance decision tree. The most demerit of this method is that only single attribute would be verified in one node. So relevancy of attributes is neglected and many problems have occured. For instance, some child trees have appeared and some attributes have be verified at a certain branch time after time. Rough Set is a math method which can describe imperfection and indeterminacy, can effectively analyse and dispose diversified imprecise, conflicting, non holonomic informations, find connotative knowledge, prompt potential rule. In this thesis a multivariate decision tree algorithm based on roughness of rough set theory in transformer fault diagnosis is presented. This method selects generalization relative to decision attribute in transformer fault decision table as root of the transformer fault diagnosis decision tree. Then this condition attribute will be the branching node if which has the smallest roughness. A more compact and reasonable multivariate decision tree of transformer fault diagnosis is formed ultimately. Results of instance comparison test verify the effectiveness of the method and conquer the most demerit of ID3 decision tree.3,Analytic Hierarchy Process (AHP) is applied to the transformer condition based maintenance decision support system in thesis 2. AHP is generally applied to decision system which has clear hierarchies and elements. Further more AHP is not direct using fuzzy specialities to describe fuzziness but indirectly constructing math model by using some simple integers such as 1,2,…, 9 to form a judgement matrix in decision problem. The elements are the reciprocals each other. So this method has defect. Moreover on account of jamming, subjective judgement and different understanding of estimator would influence result. Fuzzy AHP based on entropy weight to maintenance decision support system of power transformer is put forward in this thesis. The method uses triangle fuzzy numbers to establish judgment matrix and improves the unbalance problem of AHP's judgment matrix. Through fuzzy interval arithmetic based on level sets and optimistic index the entropy weight is given, then the thesis analyses and evaluates maintenance project according to entropy weight. This method makes the evaluation more scientific, reasonable and practical.4,a substation equipment maintenance manage system is completed. This system adopts C/S and B/S mixed architecture, applies ASP technique and BDE database access technique. By adequately utilizing Internet/Intranet technological superiority, this system can be expediently shared informations with else power automatic systems.