Skip to Main Content
This paper presents a new speed-sensorless vector control drive system for induction motor. In order to produce low harmonics in output voltage and current, reduce the torque fluctuation, and avoid the high voltage jump in switching time, the system utilizes three-level inverter to supply power for the induction motor and a SVPWM scheme with neutral point voltage balance strategy is applied for the three-level inverter. The speed estimator in this system is constructed based on artificial neural network (ANN) theory. Compared with the conventional model reference adaptive system (MRAS) method realized dependent on the PI controller, ANN speed estimator uses the intelligent learning algorithm to complete the speed estimation so that a real-time identification of the motor speed is acquired precisely with excellent dynamic response in various conditions. This paper describes both the theoretical analysis as well as the simulation results to verify the effectiveness of this drive system.