Rms and Power Measurements: A Wavelet Packet Transform Approach

Abstract

This study provides the theoretical basis for the use of wavelet packet transform (WPT) approach for root mean square (rms) and power/energy measurements. The proposed approach can simultaneously measure the distribution of the rms and power with respect to individual frequency bands directly from the wavelet transform coefficients (WTCs) associated with concurrent voltage current pair. Their dependent quantities such as power factor and total harmonic band distortion can be calculated as well. Uniform frequency bands are yielded from the WPT decomposition process of power system waveforms and can be used for identifi cation of harmonic components. The frequency bands also retain both the time and frequency relationship of the original waveforms, which is one of the major benefits provided by this approach. The approach is evaluated by its application to both analytical and actual power system waveforms.

Measurement of power systems harmonics using QMF filter banks

“…In order to investigate the influence of different wavelet filters on the harmonic analysis of power system waveform we have implemented Wavelet packet transform-based algorithm given in [3], using Wave Lab 802 software package [10]. The power system waveform is decomposed in frequency bands with a same width and the central frequency of the band is the frequency of the odd harmonics.…”

“…The number of output bands for L-level decomposition is 2 L . By adequately selecting the sampling frequency and the decomposition tree, obtained frequency bands can be used for measurement of the different harmonic components of the input signal [2], [3]. where ( )…”

“…Wavelet decomposition over 3 levelsThe rms of voltage can be computed directly from the wavelet coefficients[2],[3]:…”

Near Perfect Reconstruction Filter Banks for Power Quality Analysis

“…For instance, with the sampling frequency set to 1600Hz and with three levels of decomposition, eight frequency bands with 100 Hz interval are obtained and can be used for measurement of odd harmonics up to the 15-th harmonic. The rms of the signal can be computed directly from the wavelet coefficients in the last bands [7].…”

“…In order to investigate the influence of different wavelet filters on the high frequency harmonic analysis of power system waveforms, we have implemented the wavelet packet transform-based algorithm given in [7]. In this work we used 50Hz input waveforms sampled at 6400Hz.…”