No abstract

The majority of speed estimation schemes for sensorless vector control of induction machines use a mathematical model of the machine in the estimation process. These schemes are therefore inherently sensitive to parameter variation eOE ects in the machine. The variety of speed estimation methods makes any attempt to develop a universal approach to compensation of parameter variation eOE ects impossible. This paper concentrates on one of the most frequently used schemes, in which speed is estimated using model reference adaptive control approach (MRAC) on the basis of two estimated values of the rotor ux space vector. The estimator is analyzed in conjunction with indirect feed-forward rotor ux-oriented induction machine. An attempt is made to improve the accuracy of the speed estimation by appropriate modi cation of the speed estimator and the indirect vector controller structures using modi ed induction motor models that account for one or more of the phenomena that are neglected in development of the basic constant parameter scheme. In particular, compensation of main ux saturation, compensation of iron loss, and simultaneous compensation of both the iron loss and main ux saturation are elaborated. Novel structures of the speed estimator are developed for each of the three cases and are applied in conjunction with the appropriate modi ed form of the indirect rotor ux-oriented controller. Excellent compensation capability is demonstrated in all the cases by performing extensive simulation studies.