Artificial Neural Network in Field Oriented Control for Matrix Converter Drive

Matrix Converters are becoming popular in industrial applications due its better performance than the conventional VSI converters. The advantages of Matrix Converters are bidirectional current flow, adjustable displacement angle and sinusoidal input current and output voltage. Two control schemes used to control the speed of Induction motor. They are Field Oriented Control and Direct Torque Control systems. The Direct Torque Control system uses hysteresis controllers for flux and torque control because of which the ripples in the stator current and torque are more. Also the conventional Direct Torque Control system uses lookup table method to select the switching vectors which reduces the accuracy of the system. In the Field Oriented Control system the torque and flux components are separately controlled using PI controllers and then recombined to create the motor phase current. This gives better control than the conventional DTC system. The system uses Space Vector Modulation Circuit to choose the switching vectors for Matrix Converter. In this paper the Artificial Neural Network is introduced to replace the SVM in Field Oriented Control system. The complete Field Oriented Control system using ANN for in Matrix Converter Drive is simulated using MATLAB/SIMULINK. It is observed that in the ANN system the speed of the motor tracks the reference speed without any overshoot and speed control is achieved with zero steady state error.