Detection and Estimation of Oscillatory Components in Power Networks

• Activity Number: 628 - Statistical Applications in the Physical Sciences
• Type: Contributed
• Date/Time: Thursday, August 3, 2017 : 8:30 AM to 10:20 AM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Charlotte Haley* and Christopher Geoga and Mihai Anitescu
• Companies: Argonne National Laboratory and Argonne National Laboratory and Argonne National Laboratory
• Keywords: Signal Detection ; Time Series ; Spectrum ; Power Grid ; Signal Estimation

Abstract

Methodology for the detection and estimation of the parameters of evolutive ringing artifacts in nonstationary time series is central to seismology, radar applications, and the monitoring of power grid stability. We present a likelihood ratio test for damped oscillations in noise based on a model equation having known Karhunen-Loève decomposition of its covariance matrix. The result is reduced computational complexity for near real-time estimation of signal amplitude, frequency, damping, and signal to noise ratio. Nonstationary multitaper spectral analysis of complex phasor measurement data reveals the temporal evolution of oscillatory components in interarea (< 1Hz) and local (1-2Hz) frequency ranges in the phasor data, and our detection estimation method is applied to both estimate the parameters of the process, and test for their signicance in noise. The result is compared against the classical Prony analysis for the estimation of autoregressive model oscillations in noise.