An Automated On-line Monitoring and Fault Diagnosis System for Power Transformers

A combined artificial neural network and expert system tool (ANNEPS) was developed over the years as an off-line diagnosis tool for power transformers based on dissolved gas analysis (DGA). ANNEPS takes advantage of the inherent positive features of each method and offers a further refinement of present techniques. This tool has been confirmed to have high performance of diagnosing multiple fault types in power transformers. An automated diagnosis system, ANNEPS v4.0, is presented in this paper. The new system extends the existing expert system and artificial neural network diagnostic engine with new interface, automated database interactions, and alarm notification functions. The system receives DGA data from an on-line DGA monitor and then stores all information into a database. The combined neural network and expert system engine validates data, detects the faults, and recommends appropriate action. When the result indicates an "abnormal" condition, a notification of the diagnosis results is sent to transformer maintenance personnel through email. It also performs an automatic daily backup of both the input database and the output file. The developed ANNEPS system allows users automatic processing of on-line dissolved gas in oil data with a comprehensive diagnosis algorithm