Implementation of neural network based Space Vector Pulse Width Modulation inverter- induction motor drive system

Abstract

This paper presents a voltage source inverter which is controlled with Space Vector Pulse Width Modulation (SPWM) designed using three layers feed forward back propagation based artificial neural network (ANN). The input output samples were obtained using simulations in Matlab Simulink; which were used for extensive training of network. With calculated weights and biases, the neural network was implemented using PIC microcontroller. The designed ANN model and speed controller receives command voltage and reference speed as the inputs and generates pulse width modulated waves for three phase inverter-bridge. The designed system controls the frequency and stator voltage keeping V/f ratio constant. Results, harmonic analysis with designed ANN structure and cost comparison with other techniques is presented. Low cost due to ANN and micro-controller based implementation, flexibility of on-line training for incorporating different conditions are the main advantages of presented scheme.