Minimization of Torque Ripple in Direct Torque Controlled Switched Reluctance Drive Using Neural Network

This paper proposes a new control approach for minimizing the torque ripple in four phase switched reluctance motor using Neural Network based direct torque control technique. The Direct torque Control technique is familiar for its comparable performance and simple control structure. The optimal selection of the voltage space vector is the major drawback in the classical DTC method. To overcome the drawback Neural Network controller is proposed along with flux and sector to select the appropriate voltage space vectors from the voltage lookup table. This approach improves the torque and speed response in terms of reduced torque ripple and faster settling time over a wide speed range both in static and dynamic condition. The drive is simulated in MATLAB/SIMULINK and the results are observed and tabulated. The results reveals that the torque ripple is in range of 3.5% to 2.2% for wider range of speed and the load torque which is comparable better compared to Classical direct torque control technique.