Research on Typical Artificial Neural Networks in Power Transformer Fault Diagnosis
Zhao Hai-jun · Gaoya dianqi · 2008
In order to examine transformer momentarily to exclude possible fault in time, three kinds of artificial neural network were applied in the transformer fault diagnosis, such as the BP network, the GA-BP network, and RBF network. The structures and principles of the three networks were analyzed, and fault diagnosis was realized by MATLAB programming. The analysis of the massive experimental data indicate that the RBF network is better than the others in diagnostic accuracy.