Reactive Power Compensation of Distribution Feeder by Neural Networks

• Main Authors: TZENG, BEN LI, 曾本立
• Other Authors: CHEN,CHAO SHUN ; MOO,CHAN SHINE
• Format: Others
• Language: zh-TW
• Published: 1996
• Online Access: http://ndltd.ncl.edu.tw/handle/40915510492887823986

碩士 === 國立中山大學 === 電機工程研究所 === 84 === Capacitors can provide reactive power compensation of a power system to reduce feeder load, improve low voltages and power factor,and result in less generation capacity. Because it is more cost effective than the other schemes, this thesis focuses on the development of systematic rules and methods for an effective reactive power compensation by considering the proper installation of capacitors. To achieve efficient reactive power compensation ,the reactive and real power profiles of the study feeder has to be investigated first. This thesis collected the feeder load data and meteorology information as the training set of the neural network to predict the feeder loading. According to the load composition and typical load curve of each type of customers, the hourly real and reactive power loading of each feeder section can be estimated. With the hourly loading information derived, the reactive power compensation is solved by considering the minimum unit capacity of distribution capacitors used in distribution system. The real and reactive power loss of line feeder section was also solved easily by neural networks. By the methodology proposed, the reactive power compensation can be obtained in a very efficient manner. Because the neural network is used in this study, the complicated process of conventional reactive power compensation by regression analysis, load flow analysis and optimization procedures can therefore be prevented.According to the computer simulation, it is concluded that the neural network can be applied to the proper capacitor installation to reduce feeder loss effectively.