The main theme of this paper is to protect the transformer from unnecessary tripping due to inrush current and to overcome drawbacks in traditional frequency transform based protection schemes. In this method, Inrush and Internal fault currents are simulated and Protection of power transformer is presented using a time-frequency transform. Pre-processing is done using Continuous Wavelet Transform for decomposition of signals. Preprocessed signals are used to train Artificial Neural Network architecture using Multi Layer Perceptron, Particle Swarm Optimization Techniques. Results are compared and better classification combination is chosen.
Poornima et al. (Mon,) studied this question.