Parameter, speed, position estimations and torque ripple minimization in permanent magnet synchronous motor drives

Permanent magnet synchronous motor drives have been widely used in industrial areas. The applications of accurate current control, low-resolution position sensors and accurate position estimators are attractive for the motor dynamic performance and cost minimization. This research has focused on the development of a comprehensive parameter estimation strategy, a hybrid torque ripple minimization strategy, an anti-resistance-variation speed estimator and a position/speed estimator for high performance permanent magnet synchronous motor drives.; First, a comprehensive parameter estimation strategy is presented to eliminate the uncertainty in the PMSM model. The stator inductance is off-line modeled according to the stator currents. The torque constant and stator resistance are on-line estimated with a feed-forward neural network. A constant magnetizing current is injected into the stator winding to facilitate the resistance estimation. This strategy can provide a complete electrical model of PMSM in real time with acceptable accuracy and fast response in current and speed control.; Second, a hybrid torque ripple minimization controller for a PMSM drive is presented. The stator slot ripple is handled with off-line compensating currents, since it is nearly constant with respect to operating conditions. The rotor-flux linkage ripple is on-line minimized with the help of the torque constant neural estimator introduced earlier. The PMSM drive can generate fast compensating currents under deadbeat current control. Without the comprehensive parameter estimation method or with considerable mismatch of the motor parameters in the controller, the current from the stator winding can not follow the commands accurately.; Third, an anti-resistance-variation speed estimation strategy in PMSM is presented. The rotor speed is estimated with the adaptation of the stator resistance. The stator inductance is modeled off-line according to stator currents.; Finally, a back-emf based position/speed estimator is introduced. For simplicity, only the stationary back-emfs are estimated with a feed-forward neural network.; The on-line neural parameter estimator, hybrid torque ripple minimization controller and neural speed/position estimator have been simulated using a 0.75 hp and 1.2 hp PMSM motors. Simulation results of the torque ripple minimization controller provide smooth torque performance without off-line rotor flux linkage measurements. The speed/position estimators were able to extract dynamically accurate rotor speed and position information from the inverter currents and voltages.; Experimental implementations of the high performance permanent magnet synchronous motor drive were carried out to verify the performance. The motion controller was tested on the 1.2 hp PMSM with a Tiger-30 evaluation board that used the TMS320C30 DSP. The experimental results show that each newly developed control strategy improved the drive performance in terms of system modeling, torque ripple minimization and speed/position estimation.