DSVM based Predictive Torque Control of Induction motor for EV

Abstract

This paper presents discrete space vector modulation (DSVM) based predictive torque control (PTC) of induction motor for Electric Vehicle (EV). In the conventional PTC, only seven voltage vectors are utilized for determination of flux and torque. In order to utilize the converter control region effectively without decreasing the sampling time of the controller and to decrease the ripple in the torque, flux and current, DSVM based PTC is applied for induction motor. According to the DSVM, each sampling time is divided into 3 equal parts and by applying standard voltage vectors totally 37 virtual voltage vectors are generated. These voltage vectors are used for flux and torque prediction. To decrease the computational burden of DSVM-PTC, number of voltage vectors used for the prediction is reduced by direct torque control (DTC) based sector selection method. The starting current of the motor is limited by a nonlinear function added in the cost function. The speed of the motor is estimated by rotor flux based speed estimation technique. Torque ripple, motor current THD are analysed and compared with conventional techniques in different modes of operation of the motor. The DSVM – PTC algorithm is evaluated in simulation and experimental platform, which gives excellent response even in low sampling frequency compared to conventional methods.