Selective On-line Parameter Tuning of Adaptive PSS by Trinomial Tree Method in Wide-Area Power System

Abstract

In this study, an adaptive PSS using data measured at its generator bus is developed. The low-order linear models are utilized for the eigenvalues analysis of the large-scale power system and for the PSS on-line tuning. A generator with the PSS is modeled as the 5th order model, and the models are identified by the least squares method using measured variables at each generator. Control parameters of the phase compensator in the PSS are updated online by using the trinomial tree method based on the eigenvalue analysis. The 4th order model is also developed to obtain the dominant oscillation mode of the whole system. The PSS parameters are updated only when the dominant eigenvalues detected at the 4th and 5th order models are converged for a given period, and the identification is considered to be accurate enough to update the parameters. Dynamic digital simulation is carried out for IEEJ WEST 10-machine system model to show the effectiveness of the proposed tuning method.