Dynamic Parameter Estimation Based on Rank-Reduced Prony Analysis

Abstract

This paper presents a new least squares estimation (LSE)-based dynamic parameter estimation technique using phasor measurement unit (PMU) data. Generator parameters such as inertia constant, damping coefficients, and regulation speed constant are estimated from captured measurements during transient events. The key idea of this dynamic parameter estimation is based on unknown model structure and reduced-order model. This approach depends on measurement-based methods for ringdown signals. A rank-reduced Prony analysis is employed to accurately identify the system eigenvalues with reduced-order model. Then an optimization problem is formulated to obtain the system matrix and estimate the dynamic parameters. Sensitivity analysis is performed to the optimization problem to find the best parameter estimates.