An ANN based identification and control approach for field-oriented induction motor

A high-performance speed control approach using artificial neural networks (ANNs) for the field-oriented induction motor is proposed. By using the proposed approach, speed of an induction motor can be controlled to follow an arbitrarily selected speed trajectory. Especially, an accurate track of the speed can still be obtained, when uncertainties of the motor and its load exist. The uncertainties include the unknown load on the motor and the variation of the motor rotor resistance due to temperature rise. To evaluate the performance of the proposed control system, the system has been simulated by using detailed models of the field-oriented induction motor, the current hysteresis controlled VSI (voltage source inverter) and the ANNs. The scenarios simulated on a typical induction motor are composed of unknown nonlinear load, different trajectories of speed and patterns of rotor resistance variation. Simulated results of the proposed system are compared to those of the traditional PI (proportional plus integral) controller. Preliminary results show that the proposed control system can achieve superior performances on speed trajectory tracking as well as adaptability to parameter variations of rotor resistance