Transformer Fault Diagnosis And Maintenance Decision Method Based On Rough Set Theory

Power transformers played the role of bridge and link as a network of power structure in the rapiddevelopment of the power industry, which stable operation is closely related to the safety and reliability ofthe entire electric power system. So, when the power transformer is in trouble, determine the type of fault iscritical for the safe and stable operation of the whole electric power system.Under normal circumstances, in the electric power system network, the operating environment of thepower transformer is relatively poor, and when a fault occurs, the sensor may occur erroneously detected orthe signal is lost during transmission, thus making the collected fault information is imprecise and notcomplete, which caused a lot of distress to the power transformer fault diagnosis. In view of the rough settheory has a strong fault tolerance, the paper uses rough set theory, with the incomplete power transformerfault information as a set of condition attributes of fault classification, and the decision-making table isestablished after taking into account possible failures; In order to achieve a power transformer faultdiagnosis, it need apply the identification of matrix theory to reduce the decision table for obtaining a moresimple diagnostic rules, and verified the accuracy of fault diagnosis with examples, the results show thecorrectness of the application of rough set theory in fault diagnosis. Taking into account that the coreattributes of diagnostic rules may loss, and give the Corresponding solution which BP neural network in theMATLAB neural network toolbox is applied to simulation.In the giant power station, the power distribution network is often complicated and malfunctions ofmore than one transformer become a very real possibility (the more of the transformers in the electricsystem, the higher possibility of malfunction of one or more transformers). When more than onetransformer is in trouble and the maintenance resources is deficient, it is very urgent to make a reasonableoverhaul strategic decision (conclude a scientific overhaul order of the voltage changers) for avoidingshutdown of the troubled transformer and advancing the stability and reliability of the power supply system.The overhaul strategy is a multi-attribute decision making actually, and Dominance Rough Sets could helpa good settlement of multi-attribute decision making with preference relations. Therefore, this paperproposes overhaul of management decisions based on dominance rough set theory.So, this article proposes a maintenance management strategy of transformers based on DominanceRough Sets theory. The transformer auxiliary power system and booster station of a power plant inXinjiang as an example in the paper, with the degree of influence on the whole plant protection and thesecurity of the power supply system and the auxiliary power distribution system as condition attributes inthe ordered information system which use transformer as the object of study when the transformer is introuble; Then, the ordered information system is reduced with the method of maintaining quality boundarydomain classification invariant, and it can get the decision rule sets of the transformer overhaul sequentialthrough the reduction decision table of the ordered information system. Finally, according to the rule set, itcan acquire failure overhaul sort when the few transformer in Figure5-1, the third plant service highertransformer under overhaul and6kV Ⅲ segment operate with the first plant service higher standbytransformer. The proposed method can optimize the overhaul asset management of the transformer, andincrease transparent and rationality of the transformer maintenance management, which have referentialsignificance to the substation equipment maintenance management.