Using evidence combination for transformer defect diagnosis

This paper describes a number of methods of evidence combination, and their applicability to the domain of transformer defect diagnosis. It explains how evidence combination fits into an on-line and implemented agent-based condition monitoring system, and the benefits of giving selected agents reflective abilities. Reflection has not previously been deployed in an industrial setting, and theoretical work has been in domains other than power engineering. This paper presents the results of implementing five different methods of evidence combination, showing that reflective techniques give greater accuracy than non-reflective.

[1]  Martin D. Judd,et al.  Advances in intelligent condition monitoring and asset management of power transformers , 2003 .

[2]  Remo Pareschi,et al.  Reflective Agents for Adaptive Workflows , 1996, PAAM.

[3]  Stephen McArthur,et al.  Intelligent diagnosis of defects responsible for partial discharge activity detected in power transformers , 2003 .

[4]  Weiru Liu,et al.  Reinvestigating Dempster's Idea on Evidence Combination , 2000, Knowledge and Information Systems.

[5]  Martin D. Judd,et al.  Intelligent condition monitoring and asset management. Partial discharge monitoring for power transformers , 2002 .

[6]  Victoria M. Catterson,et al.  The Industrialisation of a Multi-Agent System for Power Transformer Condition Monitoring , 2004, SGAI Conf..

[7]  Shlomo Argamon,et al.  Arbitrating Among Competing Classifiers Using Learned Referees , 2001, Knowledge and Information Systems.

[8]  Jan Treur,et al.  Compositional modelling of reflective agents , 1999, Int. J. Hum. Comput. Stud..

[9]  Martin D. Judd,et al.  Condition monitoring of power transformers using UHF partial discharge sensors: operating principles and site testing , 2003 .

[10]  S.D.J. McArthur,et al.  The design of a multi-agent transformer condition monitoring system , 2004, IEEE Transactions on Power Systems.

[11]  J. R. McDonald,et al.  A multi-agent condition monitoring architecture to support transmission and distribution asset management , 2005 .

[12]  W.H. Tang,et al.  An evidential reasoning approach to transformer condition assessments , 2004, IEEE Transactions on Power Delivery.

[13]  S.D.J. McArthur,et al.  The practical implications of bringing a multi-agent transformer condition monitoring system on-line , 2004, IEEE PES Power Systems Conference and Exposition, 2004..

[14]  Bernadette Dorizzi,et al.  Neural networks and fuzzy data fusion. Application to an on-line and real time vehicle detection system , 1999, Pattern Recognit. Lett..