A novel approach for spectral analysis of monitored power systems

Abstract

This paper presents a novel approach that combines the adaptive and nonadaptive notch filters, multilayer perceptron neural network, and warped discrete Fourier transform (WDFT) to estimate amplitudes, frequencies and phases of the fundamental and harmonic components in power systems. Simulation results show that only 1.25 cycles of the fundamental component are enough to provide small frequency, phase and amplitude error estimations.