To improve the estimation of active power, the possibility of estimating the amplitude square of a signal component using the interpolation of the squared amplitude discrete Fourier transform (DFT) coefficients is presented. As with an energy-based approach, the amplitude square can be estimated with the squared amplitude DFT coefficients around the component peak and a suitable interpolation algorithm. The use of the Hann window, for which the frequency spectrum is well known, and the three largest local amplitude DFT coefficients gives lower systematic errors in squared interpolated approach or in better interpolated squared approach than the energy-based approach, although the frequency has to be estimated in the first step. All investigated algorithms have almost the same noise propagation and the standard deviations are about two times larger than the Cramér-Rao lower bound.

JO - Metrology and Measurement Systems L1 - http://rhis.czasopisma.pan.pl/Content/89823/PDF/Journal10178-VolumeXVIII+Issue4_06paper.pdf L2 - http://rhis.czasopisma.pan.pl/Content/89823 IS - No 4 EP - 596 KW - estimation KW - amplitude square KW - spectral data processing KW - interpolated DFT KW - Hann window ER - A1 - Lušin, Tomaž A1 - Agrež, Dušan PB - Polish Academy of Sciences Committee on Metrology and Scientific Instrumentation JF - Metrology and Measurement Systems SP - 583 T1 - Estimation of the Amplitude Square Using the Interpolated Discrete Fourier Transform UR - http://rhis.czasopisma.pan.pl/dlibra/docmetadata?id=89823 DOI - 10.2478/v10178-011-0056-6