The evolution of high performance vector control has transformed the transient behaviour available from induction motor drives. The variation of the rotor resistance in such drives remains a problem of significant industrial interest. Parameter estimators are needed which operate in both the transient and steady-state regions of the drive. Such a parameter estimator is described, which is a novel variation on the extended Kalman filter (EKF). Execution of the EKF demands a high computational performance. The algorithm presented in the paper makes use of a model order reduction process that cuts the computational requirements to approximately one third of that demanded by the EKF. The theoretical development of the algorithm is followed by a simulation study which is used to illustrate the possible range of behaviour including the introduction of noise and modelling errors. Finally, an experimental examination of performance is presented, which shows the high standard obtained when the new estimator is applied to a practical inverter machine drive.
[1]
Rupprecht Gabriel,et al.
Field-Oriented Control of a Standard AC Motor Using Microprocessors
,
1980,
IEEE Transactions on Industry Applications.
[2]
Silvio Stasi,et al.
A new EKF-based algorithm for flux estimation in induction machines
,
1993,
IEEE Trans. Ind. Electron..
[3]
Scott Wade.
Parameter identification for vector controlled induction machines
,
1994
.
[4]
K. J. Hunt.
A survey of recursive identification algorithms
,
1986
.
[5]
P. Eykhoff.
System Identification Parameter and State Estimation
,
1974
.
[6]
G. W. Irwin,et al.
The Luenberger canonical form in the state/parameter estimation of linear systems
,
1976
.
[7]
A. Johnson.
Process Dynamics Estimation and Control
,
1985
.
[8]
A. F. Puchstein,et al.
Alternating-current machines
,
1942
.
[9]
A. Jazwinski.
Stochastic Processes and Filtering Theory
,
1970
.