Fault Diagnosis In Transformer Substation Based On Petri Net Theory And Rough Set

Substation, is the site to transform, collect and distribute electrical energy in the power system, plays an extremely important role in transmission and distribution system. So, the relevant operator should quickly and accurately locate the fault zone, identify and isolate the real fault for restoring normal operation of the non-fault zone as quickly as possible, when the substation occur fault. However, due to the complex multiple faults, incorrect action of switch or protection, the loss of information caused by channel interference and other factors, the alarm information may be uncertainty and incompleteness. Fault diagnosis becomes complicated. Therefore, it is significant to develop a fast and accurate substation fault diagnosis method with small interference, good fault-tolerant.In this thesis, the characteristics and the information sources of the substation fault diagnosis are analyzed, the substation fault diagnosis methods at home and abroad are summed up systematically, a substation fault diagnosis method based on rough set and Petri nets is established. The combination of Rough set and Petri nets can supplement mutually in the diagnostic reasoning using rule knowledge. Meanwhile, it overcomes a large number of calculations of Rough set in rules search and redundancy of Petri Nets.Since most of the substation will reserve the expansion space in the design and the expansion of the substation usually aim at only one voltage level, the substation is divided into several relatively independent regions according to the different voltage level. When wiring structure or protection configuration is changed, only rebuild the diagnosis model of this region, so the other regions are not affected. Then, for each region, fault decision table is established respectively, attributes and attribute value reduction of Rough set decision table are produced, the minimal diagnosis rule is got. Because of Petri nets with a quick reasoning search, mathematical diagnosis process, the graphical structure, by using Petri nets to describe the diagnosis rules, the diagnosis models of each region are established, which achieves efficient substation fault diagnosis.Finally, applying sub-regional fault diagnosis method based on Petri nets and rough sets to the actual 110kV substation, the fault diagnosis models of various regions are established. Then, the sample test and algorithms comparison of the model are conducted. The results showed that this method is indeed a fast and accurate substation fault diagnosis method with good fault-tolerant and adaptability, and it has great significance to achieve efficient online fault diagnosis of substation.