| Distribution system is a complicated large-scale distributed system with multi-layers, and recently, customers requirement for sustainable and reliable power supply have heightened. Consequently, research on multi-layer intelligent fault diagnosis method for distribution equipment based on SCADA/AM/FM/GIS is of great academic and practical significance for reducing equipment fault and improving running reliability of distribution system. Through consulting volumes of material at home and abroad, the structures and features of SCADA, AM/FM/GIS and insulation on-line monitoring system for distribution apparatus have been analyzed respectively, and a new integration system of insulation on-line monitoring system based on SCADA/AM/FM/GIS has been established. Moreover, on the basis of these efforts, various intelligent fault diagnosis methods are systematically analyzed among which the combinatorial fault diagnosis model, composed of neural network, expert system, wavelet theory and fuzzy logic etc, is deeply studied. Two models, namely, a distributed multi-layer fault diagnosis model and a coordinated fault diagnosis expert system model, are established which obtain satisfactory results and are justified efficiently. The main work is as follows: (1) Based on the analysis of SCADA/AM/FM/GIS structure and data organizing format, an integrated system of SCADA/AM/FM/GIS and insulation on-line monitoring system is designed, whose total structure and detailed integrating steps are also discussed. After the establishment of distribution network GIS data model and equipment coding, data model, organizing format and database structure of the integration system are proposed respectively. Testing results indicate the proposed integrating method is feasible and has good prospect of popularization and application. (2) According to researching on artificial neural network, expert system, wavelet theory, fuzzy logic, genetic algorithm, and rough set theory carefully, a combinatorial fault diagnosis method of diagnosing integration system equipment is put forward. Several combinatorial diagnosis models incorporated with neural network are established. Application reveal that the combinatorial models make full use of each single diagnosis method superiority and replenish mutually. The proposed models can improve fault diagnosis accuracy a lot, and are believed to be the direction of diagnosis in distribution system. (3) Distributed multi-layer fault diagnosis methods are implemented in fault diagnosis of integration system due to its distributed and multi-layer characteristic in function and structure. The design of internal information correlated database for substation, the intelligent fault diagnosis knowledge base, coding table of diagnosis program module are delineated, and a multi-layer intelligent diagnosis model is established. Moreover, the principle and method of establishing an intelligent diagnosis knowledge base is studied, meantime the construction of an object-oriented knowledge base and a fault diagnosis intelligent inference mechanism are designed, their ability is illustrated effectively and practicably by examples. (4) Based on the principle, structure and application effect of coordinated expert system, a multi-layer coordinated fault diagnosis expert system is used in fault diagnosis of equipment for integration system, so the problems with multi variables and multi faults have been resolved. Knowledge organizing format of this coordinated expert system, and a multilevel blackboard model are studied and established respectively. At the same time, a combinatorial model incorporated with advice of experts is designed. Combinatory algorithm integrated with various expert opinion is proposed which effectively solve the confliction between different expertise. Structure of coordinated diagnosis expert system for transformer is conceived, and examples indicate the coordinated expert system can diagnose multi faults from different aspects and obtained perfect results.
|
PDF Full Text Request