Energy Loss Analysis and Improvement of Fault Isolation Strategy for Distribution Feeders

• Main Authors: SHEN-WEN HSIAO, 蕭勝文
• Other Authors: MING-YUAN CHO
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/57gr69

博士 === 國立高雄應用科技大學 === 電機工程系博碩士班 === 103 === This dissertation proposes both complete energy loss calculation models and fault isolation strategy for feeders analysis and operation of distribution system. The inaccurate factors of manual metering and load transfer are first calibrated to efficiently calculate the monthly energy loss for all feeders in distribution system. The customer load patterns are also applied to estimate the same customer’s daily electricity to calibrate the mismatch by manual metering. Both the Outage Management System / Conduit Mapping Management System (OMS/CMMS) and the New Billing System (NBS) support complete feeder information to improve the accuracy of feeder energy loss calculation. The regional energy loss model is first developed by subtracting the electricity sold from the electricity supplied, which has been calibrated by the customer load pattern for the business district. A practical feeder energy loss model is derived by calculating monthly individual feeder energy losses, including the high and low voltage feeders, copper and core losses of the distribution transformer, and non-technical losses using exact feeder data extracted from both OMS/CMMS and NBS. Besides, mathematical optimization technique was applied to improve the fault interval assessment of Feeder Terminal Units (FTUs). Because of the insufficient hardware computing power in such units, using a time coordination curve to calculate fault curves leads to the occurrence of jagged curves. Such jagged curves lead to considerably high errors, ultimately resulting in malfunction or erroneous nonoperation. A Multiple Particle Swarm Optimization (MPSO) algorithm was proposed for resolving the malfunction caused by protection curves with considerably high errors. This algorithm improves the early convergence problem encountered in the original standard PSO algorithm and the long calculation time observed in the area method. When the proposed MPSO algorithm is applied in the optimization process, the arrangement of the nodes on the jagged curves improved; this results in the jagged curves exhibiting relatively smooth and curved outlines, thereby eliminating malfunctions and achieving upstream and downstream protection coordination as well as completing fault interval identification procedures. Finally, a Web-based information system which integrates the proposed energy loss models has been developed and operated over two years for distribution system loss analysis in the business department and district. The practicality and effectiveness of the proposed models and associated information system can be justified. Moreover, the improvement of the erroneous assessment of distribution line faults observed in FTU fault flags as well as the success rate of Fault detection isolation restoration (FDIR) in feeder automation systems can be justified.