Analysis of power transformer dissolved gas data using the self-organizing map

Incipient faults in power transformers can degrade the oil and cellulose insulation, leading to the formation of dissolved gases. Even though established approaches that relate these dissolved gas information to the condition of power transformers are already developed, it is discussed in this paper that they still contain some limitations. In view of that, this paper introduces an alternative approach for the analysis of dissolved gas data, which can produce more convincing interpretation and fault diagnosis. The proposed approach, which is based on the data mining methodology and the self-organizing map, has been compared and validated using conventional interpretation schemes and real fault-cases, thereby proven to be capable of enhancing the condition monitoring of power transformers.

[1]  R. Rogers.  IEEE and IEC Codes to Interpret Incipient Faults in Transformers, Using Gas in Oil Analysis , 1978, IEEE Transactions on Electrical Insulation.

[2]  A. Ultsch,et al.  Self-Organizing Neural Networks for Visualisation and Classification , 1993 .

[3]  Yukio Mizuno,et al.  Diagnosis of oil-insulated power apparatus by using neural network simulation , 1997 .

[4]  Padhraic Smyth,et al.  From Data Mining to Knowledge Discovery: An Overview , 1996, Advances in Knowledge Discovery and Data Mining.

[5]  Hong-Tzer Yang,et al.  Developing a new transformer fault diagnosis system through evolutionary fuzzy logic , 1997 .

[6]  John W. Sammon,et al.  A Nonlinear Mapping for Data Structure Analysis , 1969, IEEE Transactions on Computers.

[7]  P. J. Griffin,et al.  A combined ANN and expert system tool for transformer fault diagnosis , 1998 .

[8]  Y. M. Tu,et al.  Transformer insulation diagnosis based on improved ANN analysis , 1997, Proceedings of 5th International Conference on Properties and Applications of Dielectric Materials.

[9]  D. G. Esp,et al.  Advances in data mining for dissolved gas analysis , 2000, Conference Record of the 2000 IEEE International Symposium on Electrical Insulation (Cat. No.00CH37075).

[10]  Kevin Tomsovic,et al.  A fuzzy information approach to integrating different transformer diagnostic methods , 1993 .

[11]  D. G. Esp,et al.  Data mining applied to transformer oil analysis data , 1998 .

[12]  Chin E. Lin,et al.  An expert system for transformer fault diagnosis using dissolved gas analysis , 1993 .