A direct torque intelligent control strategy for induction motor based on matrix converter

Abstract

Direct torque control (DTC) technology as a high-performance control method uses stator flux to determine the direction of magnetic field and controls converter switch state through discrete two-point regulator to obtain high torque dynamic performance. But none of switch states can generate exact voltage vector to produce desired torque and flux changes in most of switch instances. This causes a high torque ripple. To solve this problem, a novel DTC strategy using fuzzy neural network (FNN) instead of hysteresis controller is proposed. Flux error, torque error and flux angle are taken as FNN variables. In order to solve nonlinear problem of parameter variation in DTC, a fuzzy adaptive speed PI regulator of analysing optimizing PI control speed response and combining experts experience is proposed. The simulation results show that control system obtains rapid speed response, stronger robustness, higher precision of speed control and small torque ripple.