A new research method of direct torque control based on intelligent optimization and predictive control

The article adopts two different methods to solve the ripple problem of direct torque control in the low speed region according as different control frequency. They are variable structure control. In the subaltern low speed region, the partition of flux linkage region is added and the system optimizes the choice of switch vector. In exceeding low speed region, the system utilizes prediction to reduce the ripple of torque. Synchronously the genetic algorithm (GA) is used to optimize system parameters. GA solves the problem that predictive control makes the system increase dependence to parameters. The result of simulation testifies the method can effective reduce speed ripple amplitude. GA can accomplish parameter optimization at three generations. At this time, the torque ripple reduces 92.3%.