Nonlinear Estimation of Synchronous Machine Parameters Using Operating Data

Abstract

This paper presents a nonlinear parameter estimator for synchronous machines based on the unscented Kalman filter. The proposed methodology uses voltages and current signals recorded from the stator and the field winding to update the parameters of the classical model of the synchronous machine for stability studies. The methodology can be applied without interrupting the normal operation of the generator. Park's Transformation is included in the estimation process to relate the stator measurements (in abc components) to the nonlinear voltage equations in the qd0 reference frame. The proposed robust methodology has been validated using real and simulated data to estimate the model parameters of a 483-MVA round rotor machine.