Application of wavelet multi-resolution analysis & perceptron neural networks for classification of transients on transmission line

This paper proposes a technique that uses Wavelet Multiresolution Analysis (MRA) and Neural Networks for the detection and classification of transients in a power system. Daubechies eight (db 8) wavelet transforms of the phase current on a transmission line fed from both ends are used. The 5th level output of MRA detail signal of phase current is used to train a perceptron neural network. After training, the perceptron neural network is able to classify all three types of power system transients correctly. All the work is carried out in MATLAB Power System Block set program. The simulation results show that the proposed method is simple, accurate and reliable to automate the procedure of classification of power system transients. This paper is focused on identification of transients but can also be easily extended to other power system solutions such as fault location and so forth.