DSPACE Implementation of a Neural SVPWM Technique for a Two Level Voltage Source Inverter

Abstract

This article presents the implementation of an improved space vector pulse width modulation (SVPWM) technique based on neural network for a real two level voltage source inverter (VSI) realized in our Lab. The major goal of using this new technique is the amelioration of the voltage quality in the output of the VSI by decreasing the effect of the harmonics. The used technique has been simulated by MATLAB/Simulink and then implemented using a DSPACE card on a real two level VSI. The advantages of the used technique are shown by simulation and experiment results.