Analysing power system waveforms using wavelet-transform approach

This paper presents the application of a wavelet transform-based algorithm to the harmonic analysis of nonstationary power system waveforms. The waveforms are first decomposed into sub-bands of smaller frequency spectra using modified discrete wavelet packet transform filter banks. Harmonic trends are obtained from the decomposed outputs. A continuous wavelet transform is then applied to sections of the decomposed sub-bands to evaluate the harmonic frequencies, amplitudes and phases. The results of analysis of field-test capacitor bank switching waveforms show that the approach is excellent for harmonic analysis of nonstationary power system waveforms. It gives both the harmonic trends and the harmonic amplitude, phase and frequency values.