Neural network based modeling of round rotor synchronous generator rotor body parameters from operating data

It is generally accepted that in order to account for the effect of eddy currents in the solid rotor-iron of a round-rotor synchronous machine, two or more fictitious rotor-circuits are to be used in each axis of the d- and q-axis equivalent circuit representations of the machine model. This paper presents a novel technique to estimate the parameters of these rotor-circuits (hereinafter referred to as rotor body parameters) from measurements collected online at several operating conditions. The effects of generator saturation, rotor position and loading are included in the estimation process. Tests conducted on a round-rotor synchronous generator reveal that certain rotor-body parameters are nonlinear functions of generator operating condition. A novel artificial neural network (ANN) based technique is used to map variables representative of generator operating condition to each parameter being modeled. The developed ANN models are validated with measurements not used in the modeling process