Identification of Transformer Internal Faults by Using an RBF Network Based on Dynamical Principle Component Analysis

Abstract

In this paper; a method is proposed to detect and identify parameter faults in nonlinear dynamical systems. The approach is based on the principal component analysis (PCA) and artificial neural networks (ANNs) based on radial basis functions (RBFs). A nonlinear system's input and output data is manipulated without taking consideration any model in the approach. The method is applied to a three phase custom built transformer in order to detect and identify internal short circuit faults. It is observed through various application examples that the proposed method leads to satisfactory results in terms of detecting parameter faults in non-linear dynamical systems. © 2007 IEEE.