220kV Transformer Failure Analysis And Detection Investigation

By means of elaborating the causes of 220kv transformer's failure, this thesis introduced the basic idea of information fusion to the transformer fault detection, made full use of dissolved gas and a variety of failure related information such as electrical test, running condition, maintenance records and expert experience etc., under the basic framework of information integration, directly used the judge conclusion of modified three-ratio method, combined with artificial neural networks and evidential reasoning technique presented a hierarchical, reliable, open transformer integrated fault diagnosis model-multi-level decision fusion model.This model combined with the conclusions of conventional electrical testing, and took lessons from on-site operation, diagnostics and maintenance experience, had strong knowledge representation and uncertainty handling capacity, could stimulate the process of uncertain reasoning more intuitive and realistic.