Discrimination between inrush and fault currents of transformers using Artificial Neural Network Tools

Document Type : Original Article

Author

Maha A. Elmohallawya, Sameh I. Selemb And Amal F. Abdel-Gawada,b,* a Department of electrical Engineering, Zagazig Higher Institute of Engineering and Technology, Zagazig, Egypt. b Electrical Power & Machines Department, Faculty

Abstract

This paper presents a technique for modeling of transformer inrush and fault currents using fitting tools of artificial neural network (ANN) as a type of new security techniques. Inrush and fault currents at different winding connection, initial flux and fault type are simulated. The proposed method is simulated using MATLAB neural network tool and Simulation package. ANN trained to recognize the inrush current and the type of fault, based on several methods of statistical inference on three phase transformer signals such as the mean value, the standard deviation and the product moment correlation coefficient factor are qualified in this paper, and using 2nd harmonic Fourier analysis and recording the maximum 2nd harmonic current for three phase signals at different operating conditions maximum current of second harmonic signals. Using this data to train and test artificial neural network. Comparison between different algorithms of training is made to fit the inputs and targets.

Keywords

Main Subjects

The Egyptian International Journal of Engineering Sciences and Technology
Volume 37, Issue 2
Electrical Engineering
March 2022
Pages 34-38

Files

Share

How to cite

Statistics