The use of time-varying AR models for the characterization of voltage disturbances

Abstract

A signal-processing tool, autoregressive (AR) modeling, has been introduced to analyze disturbance measurement data from events, like voltage dips, capacitor switching and transformer energization. We show that time-varying AR models may be utilized for detecting sudden transitions of the signal by examining the models' residuals. We also show a potential way of detecting and describing signal transients, using the time-evolved AR model spectra. To examine the disturbances over the entire spectral range, a multiresolution disturbance analysis scheme is presented.