Identification of synchronous machine linear parameters from standstill step voltage input data

This paper presents the results of a time-domain identification procedure to estimate the linear parameters of a 15 kVA salient-pole synchronous machine at standstill. A step voltage input test is performed, and the parameters of the time constant models and equivalent circuit models are estimated. The maximum likelihood algorithm is used for the estimation, and the best-fit model is selected from a set of increasing order models. The initialization values for the parameters to be estimated are determined from the operational inductances derived directly from the measured time domain data. The simulated equivalent circuit model response is validated against the measured standstill time domain and frequency domain data. In addition, simulation of the model response to an on-line small disturbance test is compared to the measured dynamic response.>