A New Low-Frequency Oscillatory Modes Estimation using TLS-ESPRIT and Least Mean Squares Sign-Data (LMSSD) Adaptive Filtering

Abstract

In this paper, we propose a new low-frequency power system oscillating modes estimation method i.e. Total Least Square-Estimation of Signal Parameters using rotational invariance technique (TLS-ESPRIT) based on Least Mean Squares Sign-Data (LMSSD) Adaptive filtering. LMSSD adaptive filtering has considered the effect of Additive White Gaussian Noise (AWGN) produces due to filters used for preprocessing of a signal from Phasor Measurement Unit (PMU) and efficiently reduced its effect without any phase shift. The comparison of the LMSSD adaptive filtering method has been carried out with the ESPRIT, the LS-ESPRIT and the TLS-ESPRIT method on a test signal at different Signal-to-Noise Ratio (SNR). Robustness of the LMSSD adaptive filtering is demonstrated in the presence of AWGN through 50000 Monte-Carlo simulations. The LMSSD adaptive filtering Estimation technique has been applied to Kundur’s two area power system using MATLAB/SIMULINK. From estimated results obtained, it is observed that the LMSSD adaptive filtering performs much better than the standard TLS-ESPRIT regarding the standard deviations and mean in the modes estimation.