A new method to diagnose the type and location of disturbances in Fars power distribution system

Fault detection and diagnosis (FDD) of power systems have become important issues due to the high power quality (PQ) demands for modern systems. For this purpose, wavelet transform is invoked to extract features of different transient disturbances. Then, an artificial neural network (ANN) as a powerful intelligent method is employed to automatically classify the disturbances based on their features. The energies of the features based on Parseval's theorem are used to train the ANN. The collected data of Fars power system is considered to evaluate the proposed FDD approach. Simulation results show the approach can diagnose different fault categories and detect the fault locations.