Intelligent Fault Types Diagnosis System For Dissolved Gas Analysis Of Power Transformer

• Main Authors: Hu, Li-Siang, 胡立翔
• Other Authors: Yang, Ming-Ta
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/71967416753666815466

碩士 === 聖約翰科技大學 === 電機工程系碩士班 === 100 === Power transformers are crucial devices in power systems with the primary function of stepping voltages up or down to an appropriate level at which the system can operate. Transformer malfunction is often caused by the aging of internal insulation. Dissolved gas analysis is important for performing condition-based maintenance of transformers. This study proposes a fault diagnosis system for transformers that references the content of dissolved gas in the transformer. Using multinomial logistic regression in conjunction with a backpropagation neural network (BPNN) for diagnosis, the system facilitates the arrangement of various maintenance procedures to enhance transformer reliability. We employed Matlab to construct neural networks for various fault types and selected various gas compositions and amounts that influence malfunction using multinomial logistic regression. We trained the various neural networks with the selected parameters, after which transformer fault diagnosis was implemented.