Power system transient signal analysis based on Prony algorithm and neural network

Abstract

A method of power system transient signal analysis based on improved Prony algorithm and neural network is used here to improve the accuracy of analysis of the transient signal with harmonic and inter-harmonic. The model of the Prony algorithm has the excellent characteristic of exactly describing transient signal and directly acquiring the frequency of signal. The frequency components of the signal were estimated assumably using Prony algorithm to confirm the amount of nerve cells and the beginning parameter of neural network firstly. Next, each frequency was treated as weight to be adjusted to estimate frequency values and amplitudes of the base wave, harmonics and inter-harmonics. Matlab simulation results demonstrated that the algorithm achieved high accuracy and rapid convergence.