Multi-fault diagnosis information fusion for transformer

It is the first time that the method of fault intelligent prediction and diagnosis of transformer in early stage has been put forward based on multi- physical effects in this thesis. The multi-signal, multi-parameter model was elaborated from the different angles when transformer is in faults. The parameters and signals can be found that indicate the state in faults based on the method of parameter estimation and the method of the signal analysis. Owing to the use of method of multi- physical information fusion, it is easy to detect early faults and separate a fault from others with the aid of powerful parallel processing and the non-linear reflective ability of intelligent measures as nerve network etc. So this realized earlier period faults intelligent diagnosis and prediction for transformer. Taking a typical fault as the example, author analyzed the availability of information fusion for fault diagnosis and multi-breakdown separation.