Parameter validation for Kalman filter based dynamic state estimation of power plant dynamics

Abstract

A study of effect of parametric variability on dynamic state estimates of synchronous generator operating using terminal Phasor Measurement Unit has been conducted. Parametric variations have been modelled using Monte Carlo method and state deviation from actual ones has been presented using suitable metrics. In addition impact of individual parametric variations on all the states have been studied as well. Furthermore, two Kalman filter variants (Extended Kalman Filter with unknown inputs and Unscented Kalman Filter) has been considered to ascertain whether choice of Kalman filter affects state estimates when subjected to parametric variability. Initial results have been performed on a Single Machine Infinite Bus (SMIB) system and consistency of the results has been validated on an interconnected network using the benchmark IEEE 39 bus system.