|
This dissertation presensts a new sensorless speed estimation
method for induction motor, which replaces the feedback signal of
conventional speed sensors. Using the proposed method, closed-loop
speed control is accomplished without any speed measurement.
The proposed method is based on an adaptive flux observer, in
which a second-order Kalman filter is employed to modify the
esitmated rotor flux using the actually measured stator currents. The
modified rotor flux estimate is substituted into the speed estimation
equation derived in the adaptive observer to improve the speed
estimation results. In comparison with the reduced-order Kalman
filter method, the proposed method reduces the computation
complexity. Moreover, the proposed method improves the accuracy
of speed estimation compared with the conventional adaptive
observer, while maintaining the computational load moderate.
In the experimental implementation, vector control and direct
torque control are investigated to verify the effectiveness of the
proposed speed and flux observer. With the proposed method, speed
sensorless control of induction motor is achieved. Experimental
results shwo that the proposed speed estimation scheme improves the
accuracy of speed estimation compared with the conventional
adaptive observer. The proposed method is proven to be an effective
speed sensorless control strategy of induction motor with moderate
computational load.
|