CMAC neural network application for induction motor drives

This article presents a newly developed speed-sensorless induction motor drive using cerebellar-model-articulation-controller (CMAC). The gradient-type learning algorithm is used to train the CMAC neural network online in order to provide a real-time adaptive identification of the motor speed. The CMAC is then viewed as a speed estimator that produces the estimated speed to the speed control loop that accomplishes the speed-sensorless vector control drive. From the experimental results, the proposed CMAC speed estimator forces the estimated speed to follow the actual motor speed precisely. The article describes both the theoretical analysis as well as the simulation results to verify the effectiveness of the proposed method.