Improving the accuracy of the rotor resistance estimate for vector-controlled induction machines

The estimation of rotor resistance in a vector-controlled induction machine is necessary to achieve high performance torque control. The extended Kaiman filter (EKF) or the extended Luenberger observer (ELO) have been used to estimate this machine parameter. Three techniques are presented for use with the EKF and ELO which improve the accuracy of the rotor resistance estimate, either in both estimators, or in the EKF alone. These techniques are: the use of the synchronous two-axis (dc-qc) frame model of the induction machine with the EKF, the inclusion of the core loss resistance to precalculate the phase currents used by the estimators, and the injection of a high frequency sine wave on the flux current reference command. These improvements are achieved without increasing the complexity of the estimation algorithms. The consequent improvements in the rotor resistance estimation are illustrated through simulation and practical implementation of a vector-controlled induction machine. A high performance digital signal processor (DSP) is used in the practical implementation. © IEE, 1997.