Decoupling control of induction motors using artificial neural networks

In order to make a induction motor has excellent control performance of torque and speed as DC motors, a novel approach of the decoupling control of the indution motor based on artificial neural networks is presented. A decoupling controller built by a real time recurrent neural network has excellent dynamic performance, because neural network has the capability of representing and processing temporal information, which is applicable for nonlinear dynamic systems. In order to reduce the training time of neural network decoupling controller, an improved adaptive training algorithm is presented. The learning rate is adjusted incessantly during the training of the network. Thus, the training of the network is accelerated. Simulation result shows that the neural network decoupling control has a good dynamic response characteristic.