Application of a neural network based space vector PWM algorithm to multi-level inverters

Abstract

In this paper, application of an artificial neural network (ANN) based space vector pulse width modulation technique (SVPWM) to a three-level inverter has been realized. However, the proposed modulation technique can be also applied to multi-level inverters having more than three levels. Two ANN structures are applied in the modulation algorithm. One of them is used for determining the sector in the space vector hexagon, and the other one is used for calculating the region in the sector. Modeling and simulation of the proposed control algorithm and the three-level inverter feeding either an R-L load or induction motor load have been performed. Simulation results are presented illustrating good performance of the drive system.