Fault Diagnosis of Oil-Immersed Power Transformers by Using Back-Propagation Neural Networks

• Main Authors: Tsai,Cheng-Yu, 蔡承佑
• Other Authors: Wang,Chau-Shing
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/a35j63

碩士 === 國立彰化師範大學 === 電機工程學系 === 106 === The large power transformer is one of the most important and expensive equipment in the power system. The operating situation of transformer will directly affect the safety of the power system, and the fault of the transformer may cause the power interruption and the profit loss. Therefore, early detection of the initial fault of transformers, reduction and prevention of the fault can improve the reliability of power system.In this paper, we use the dissolved gas in transformer oil for fault diagnosis. The data of dissolved gas will provide an indirect basis for the internal hidden trouble of transformers. Dissolved Gas Analysis (DGA) is the most popular and effective method for diagnosing transformer’s initial failure; however, due to the variability of gas data and operation, it is not easy to identify the characteristic of transformer’s fault by traditional methods. Therefore, in this paper, we used the algorithm of BP (Back-Propagation BP) neural network to train and test to improve the efficiency and accuracy of system diagnosis. The experimental results show that, using BP neural network to identify the transformer fault can improve the accuracy of about 40% than the traditional IEC ratio method, which can not only make up the error of traditional IEC ratio method, but also prove the reliability of this method in the initial fault detection of the transformer.