A spectral-neuro modeling for fault location and diagnosis in distributed power system

In this paper an approach for the improvement of the neural training error and its improvement for real time system is developed. The approach of learning the neural network is developed using the advanced feature of wavelet transformation approaches and is designed to observe even the minor current variations, which result in variation in the observation result. In this paper a contribution towards the accurate estimation of the current disturbance is developed using DWT learning for a neuro classifier.