MAXIMUM-LIKELIHOOD-ESTIMATION OF SYNCHRONOUS MACHINE PARAMETERS FROM FLUX DECAY DATA

A time-domain system identification procedure to estimate the paramete rs of a 5 kVA salient pole synchronous machine from standstill test me asurements is proposed. The test consists of a dc flux decay signal ap plied to the d-axis and q-axis of the machine. From the recorded respo nses to this signal, the admittance transfer function models and the S SFR equivalent circuit models are identified. The Maximum Likelihood a lgorithm is used to estimate the model parameter values, and the Akaik e Criterion is used to select the best-fit model. The performance of t he standstill models in the dynamic environment is studied through sim ulation of an on-line small-disturbance test. The results are compared with measured data.