MAXIMUM-LIKELIHOOD-ESTIMATION OF SYNCHRONOUS MACHINE PARAMETERS FROM STANDSTILL TIME RESPONSE DATA

This paper presents a systematic approach for identification of a three-phase salient-pole synchronous machine rated at 5kVA from standstill time-domain data. The machine time constant models and the equivalent circuit models are identified and their parameters are estimated. The initialization of the estimated parameters is achieved by the Laplace transformation of the recorded standstill time-response data and the derivation of the well-known operational inductances. The estimation is performed using the Maximum Likelihood algorithm. Based on the best estimated equivalent circuit models, simulation studies using the measured on-line dynamic responses are performed to validate the identified machine models.