Premium
Application of a combined decision model based on optimal weights in incipient faults diagnosis for power transformer
Author(s) -
Tingfang Yang,
Haifeng Liu,
Xiangjun Zeng,
Wei Qiu,
Wenbin Deng
Publication year - 2017
Publication title -
ieej transactions on electrical and electronic engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 30
eISSN - 1931-4981
pISSN - 1931-4973
DOI - 10.1002/tee.22363
Subject(s) - transformer , artificial neural network , backpropagation , computer science , fault (geology) , decision model , electric power system , artificial intelligence , data mining , algorithm , power (physics) , engineering , machine learning , voltage , physics , quantum mechanics , seismology , geology , electrical engineering
A new method for the diagnosis of incipient transformer faults based on a combined decision model using IEC 60599 three‐ratio method, grey system theory (GS) algorithm, Dempster–Shafer evidence (DS) theory, and backpropagation (BP) neural network diagnosis method is proposed in this paper. The new method will eliminate the influence of any one individual diagnosis method due to insufficient information or diagnostic prejudices. In order to minimize the sum‐of‐squared errors of the proposed method, the optimal weights of every individual method which comprises the combined decision model are calculated based on their own diagnosis error ratios. Thus a combined decision model based on optimal weights (CDMOW) is formed. Then the optimal diagnosis result is obtained based on CDMOW. Results of diagnosis instances show that the use of CDMOW not only gives a much higher diagnostic accuracy rate but also more stability in fault diagnosis than any single method composing it. © 2016 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.