Application of Hilbert Spectrum based on SWT to Low-Frequency Oscillation Analysis

Abstract

Abstract Low-frequency oscillation (LFO) appears in power systems with fast excitation of a unit or large-scale connections. An LFO signal exhibits typical nonstationary and time-varying characteristics. The paper proposed an LFO analyzing technique through the Hilbert spectrum of stationary wavelet transform (SWT). The SWT-based Hilbert spectrum contains time-varying and frequency information of LFO. The analytic framework is introduced, SWT is formulated to decompose an LFO signal into the sum of several coefficient components, the instantaneous parameters of each coefficient are then individually calculated, the complete time-frequency distribution is combined, and finally, the time-varying oscillations characteristics parameters are extracted. The simulation shows that the proposed method effectively reveals time-frequency features of the LFO signal.