Torque ripple minimization in direct torque control of induction motors

As kind of new fuzzy to large torque ripple in direct torque control (DTC) of induction motors at low speed, a neural networks (FNN) approach was proposed based on the merits that fuzzy control absorbs man's empirical thinking and neural networks have self-organization and self-study ability. The new approach achieved inverter switch's duty ratio control and made torque ripple minimum. The problem that BP algorithm easily gets into local minimum was solved by using least square method for the training of fuzzy neural networks. By comparing traditional DTC approach with FNN duty ratio control scheme, the effectiveness of FNN duty ratio control scheme was verified.