EXPERIMENTAL INVESTIGATION OF POWER TRANSFORMERS PROTECTION USING WAVELETS AND PATTERN RECOGNITION TECHNIQUES

Document Type : Original Article

Author

Helwan University, Faculty of Engineering, Dept. of Electrical Power & Machines Engineering, Cairo, Egypt.

Abstract

 
A few papers have been interested in the experimental investigation for the transformer protection; in this paper a real time laboratory investigation for power transformer protection is presented. A transformer of (5 KVA, 220/110 V) is used to represent a power transformer. The current transducers are used to transfer the primary and secondary current signals of the tested transformer to the data acquisition system connected to pc by sampling frequency of 10 kHz. Lab- View package is used as software. The protection algorithm based on wavelet transform and pattern recognition (support vector machine) as described in [21]. The experimental results indicate that it are compatible with the simulated results and the proposed technique is stable, reliable, and fast during the discrimination between internal and external faults, magnetizing inrush currents, and switching on internal faults, so, the proposed technique can be used in the real time applications.

Keywords