Research of speed observer based on Neural Network optimized by fast modified ACO in DTCs

To enhance its global optimization speed, the basic ant colony optimization. (ACO) is modified, then it is used to optimize the neural networks (NN), and the optimized NN is applied to the direct torque control (DTC) system, so that the rotate speed can be observed. The DTC with speed sensorless is implemented at last. The research of simulation shows that, the modified ACO has eminent global optimization performance and fast convergence rate, the rotate speed of the system is able to be observed by the DTC system with the NN optimized by this method exactly, thereby, the DTC with speed sensorless can be implement.