Neuro-fuzzy-based torque ripple reduction and performance improvement of VSI fed induction motor drive

An indirect vector control is used in various industrial induction motor drives for better dynamic performance. But the performance of closed-loop controlled induction motor drive is largely influenced by the type of current controllers used. Usually, proportional-integral (PI)-based current controllers are used due its simplicity. But these PI controllers suffer from tuning problem. To overcome the problem, this paper presents adaptive neuro-fuzzy (ANFIS)-based current controllers for indirect vector controlled VSI fed induction motor drive in order to minimise the torque ripple. The proposed ANFIS controller significantly reduces the torque ripple compared to that of conventional PI controllers without using any filter. In this, the current command values and actual values of current are fed to the ANFIS-based controllers and the controller is trained based on the obtained values. The performance of the indirect vector controlled induction motor drive has been simulated at different operating conditions using the ANFIS controller and the obtained results are compared with PI controller. Additionally in the implemented SVM algorithm, the switching times proportional to the instantaneous values of the reference phase voltages. It eliminates the need the calculation of sector and angle information unlike conventional SVM method.