Simulation and Experimental Study of Vector Controlled Induction Motor Drive Based on RBFNN-MRAC Controller

Although induction motors have many structural advantages, they are difficult to control due to complicated and non-linear structures. In recent years, the control performance of induction motors has been improved by using intelligent control methods such as fuzzy control and artificial neural network control. This article presents implementation of model reference adaptive control (MRAC) based on radial basis function neural network (RBFNN) for vector controlled induction motor drive at real time. The complex calculations used in vector control of induction motors can be easily done with digital signal processors, powerful microprocessors. In this study, a dsPIC30F6010A microcontroller based speed control system was designed for an induction motor with three-phase squirrel cage. The system consists of control board based on dsPIC30F6010A device, power module, three phase induction motor and DC generator that acts as a proper load. The rotor flux vector required for the coordinate transformations is estimated using indirect field-oriented control technique. The results show that the proposed RBFNN based MRAC scheme has better tracking performance than traditional PI controller for different references and loading conditions. The success and superiority of the proposed RBFNN based MRAC controller over the traditional PI controller was demonstrated by simulation and experimental studies.