Research Of Intelligent Fault Diagnosis Of Transformer Substation Based On Petri Net

Transformer substation that is hub of power and user is an important part of power system, whose function is to convert and distribute electricity power. So that the reliability and safety of transformer substation affect safe running of power net. With the development of power system, the developing direction of transformer substation whose structure is more and more complicated turns towards the high-capacity, super-capacity and long distance recently. Fault cannot be avoided because of the factory of nature, human and equipment, so that operators must locate fault area and identify fault unit to isolate the fault and provide reliable foundation for fault restoration as quickly as possible, when the fault of transformer substation appears. As the same time, the standard of transformer substation integrated automation is increasing, more and more relay protection and automatic sets are applied to transformer substation. All sorts of alarm information appear with the occurrence of fault of transformer substation, such as the action of fault oscillogram alarm of protection set, protection action. The alarm information run into SCADA so immediately that Fault diagnosis become complicated. It is so difficult for operator to identify fault quickly and accurately that incorrect judgment and incorrect process will lead to more serious accident. Therefore, the research of intelligent fault diagnosis of transformer substation has important theory and practical meaning. A lot of technique and theory of fault diagnosis in the power system have been proposed, such as expert system, artificial nerve net, Petri Net and Fuzzy theory. But the above theory has limiting in the fault diagnosis of transformer, therefore it is urgent to study new theory and method for fault diagnosis model, analysis and appraisal of system characteristic. The thesis adopted patulous Petri Net to model and analyze fault diagnosis system of transformer substation, then designed fault monitor software of transformer substation. The main work of the thesis is showed as followed:(1) The paper introduced main achievement of fault diagnosis theory and analyzed their advantage and disadvantage.(2) Secondly, the thesis analyzed fault and fault cause of transformer substation, and then introduced an effect fault diagnosis method.(3) Petri Net is an effect model tool, which is applied to many fields such as computer science, industry control and power system. The thesis introduced the concept and characteristic of Petri net in the point of power system, and then proposed HCETPN.(4) Petri net and Bayers net model have been studied and applied in many fields as a tool of model and analysis, but they have their limiting. The thesis proposed BPN which combined Petri net and Bayers net to diagnose fault of transformer substation.(5) The fault diagnosis work and behaviour of unit of transformer substation emerge hybrid characteristic. And favorable math model is the key of fault diagnosis. The thesis modeled fault diagnosis of important unit of transformer substation, and studied the reliability and feasibility of the model with simulation.(6) Transformer substation monitor software that is the key of transformer substation integrated automation system gets fault information to provide foundation of fault diagnosis to operator, the monitor software is also precondition of fault diagnosis. The thesis used VC to design transformer monitor software.