| Summary: | 碩士 === 國立中山大學 === 電機工程學系研究所 === 91 === Because of the voluminous distribution facilities, the distribution system becomes very complex and it is more and more difficult to manage the distribution system manually with the traditional methodology. With the advancement of computer technologies, it is feasible for the electric utility to use the workstations and Geographic Information System (GIS) to create and maintain distribution system database (DSDB) based on automated mapping and facilities management system (AM/FM). To advance the distribution system operation, it has been considered to develop outage management system based on DSDB to support the planning and operation of distribution systems, more efficiently.
In this thesis, a knowledge-based outage management system with a colored Petri net (CPN) inference model is designed to identify the most probable faulted device for trouble call analysis. The CPN model for inference engine considers two major parts: the traditional escalation method and the load comparison method. To achieve more accurate the estimation of faulted devices according to the outage reports by service customers, the searching procedures based on the improved escalation method are included in the best first search of the CPN. The most possible outage locations with the corresponding blown out protective device are determined by applying parallel-like reasoning in the CPN to identify multiple fault events simultaneously. A distribution system of Taipower is selected for computer simulation to demonstrate the effectiveness of the proposed methodology to enhance the function for distribution outage management.
|
|---|
