A robust TLS-ESPIRIT method using covariance approach for identification of low-frequency oscillatory mode in power systems

Abstract

A robust spectral estimation technique for identifying low frequency oscillatory mode is presented in this paper. The effect of outliers is taken into account in the proposed method which occurs due to temporary fault in network or measuring instruments. In order to reduce the effect of such bad data the paper presents a robust covariance approach based TLS-ESPIRIT estimation using multivariate M (MM) estimator. Comparison of the modified TLS-ESPIRIT method with the proposed method are done on two test signals for mode identification at different signal-to-noise ratio (SNR). Efficiency of the proposed method is obtained in presence of outliers through Monte Carlo simulations. Comparison of the proposed method with the TLS-ESPIRIT method is also carried out on real time measurement data of Western Electricity Coordinating Council (WECC) network.