Parameters Selection Criteria of Prony Method for Accurate Harmonics and Inter-Harmonic Component Identification.

A pseudo-perfect power supply is today a serious problem for researchers and power engineers and producers. The use of modern electronic devices for developing power systems leads unfortunately to presence of nonlinear loads and causes the apparition of harmonics and inter-harmonics superimposed to the fundamental sine wave. To reduce and mitigate harmonic propagation along the power network, filters are used. To design such filters an accurate method is needed to measure the harmonic levels in the network. Signal Processing has contributed amazingly in the identification of harmonics and inter-harmonics during last decades. Earlier, methods based on Fourier Transform (DFT, FFT, STFT, etc.) were used for spectrum estimation, but they suffer from problems of resolution and some invalid assumptions. On the other side, parametric methods based on mathematical models are widely used because of their accuracy and the ability to extract all harmonics and inter-harmonics without knowledge of the fundamental. This paper devotes to a detailed study of Prony method for harmonics and Inter-harmonics identification in power systems. The influence of selection of each parameter is investigated as well as the interaction between them. Experimental simulation using MATLAB confirms the criteria of parameters selection for accurate identification and the limits of Prony algorithm performance. [ABSTRACT FROM AUTHOR]