A Neuro-Fuzzy Direct Torque Control Using Bus-Clamped Space Vector Modulation

This paper presents proportional integral direct torque control (PIDTC) and neuro-fuzzy-based DTC (NFDTC) using bus-clamped space vector modulation for two-level inverter. The PIDTC provides more ripples in flux, speed, and torque during speed/load changing conditions, and also gives more total harmonic distortion (THD) in current of the induction motor (IM). In order to improve the dynamic and the ripple performance of IM drive, the NFDTC is used in which the PI-controllers (speed and torque) are replaced with neuro-fuzzy controls using hybrid learning algorithm. The hybrid algorithm divides into two parts such as back propagation and least square estimation methods to improve the controller performance. Also, the duty ratios are controlled as independent of sampling time to improve the gating pulse of the inverter. A prototype IM drive is developed with a DSPACE DS-1104 controller to validate their simulated performance. The experimental results show that the NFDTC has less current THD than PIDTC leading to significant improvement in torque and flux ripples with fast dynamic response of IM drive.