Assessment of the accuracy of synchronous generator model parameter estimation based on noisy dynamic waveforms

In the paper, there are presented the results of parameter estimation of the mathematical model of a synchronous generator operating in a single-machine power system. Noisy dynamic waveforms caused by introducing a disturbance in the form of a step change in the reference voltage or a pseudorandom signal to the generator voltage regulator system were the basis of estimation. It was assumed that the generator load was close to the rated one. Finite and infinite impulse response zero-phase digital filters were used for filtering the waveforms. The least squares method was used for parameter estimation, and the gradient method was applied to minimization of the mean square error.