Speed sensorless stator flux-oriented control of induction machine in the field weakening region

In a conventional speed sensorless stator flux-oriented (SFO) induction machine drive system, when the estimated speed is transformed into the sampled-data model using the first-forward difference approximation, a modeling error occurs in the sampled data model. As the result, an error in the rotor speed estimation is produced. The error included in sampled data model of the estimated speed is removed by the use of a digital low pass filter (LPF). However, the delay of the estimated speed occurs in transients due to the use of the LPF. Consequently, current control loss occurs at the transition to field weakening region by the delay of the estimated speed. This paper investigates the problem of a conventional speed sensorless SFO system produced by the delay of the estimated speed in the field weakening region. In addition, this paper proposes a new method to estimate exactly rotor speed by using a Kalman filter. The proposed method is verified by simulation and experiment.