Author/Authors :
benbuhenni, habib ecole nationale polytechnique d’oran maurice audin - laboratoire d’automatique et d’analyse des systèmes (laas), departement de génie électrique, Oran, Algeria , zinelaabidine, boudjema hassiba benbouali university - laboratoire génie électrique et energies renouvelables (lgeer), electrical engineering department, Chlef, Algeria , belaidi, abdelkader ecole nationale polytechnique d’oran maurice audin - laboratoire d’automatique et d’analyse des systèmes (laas), departement de génie électrique, Oran, Algeria
Abstract :
This article presents an improved direct vector command (DVC) based on three-level space vector modulation (SVM). The essentiel disadvantage that is usually associated with DVC is the high harmonic current and power ripples. To overcome this disadvantages a SVM technique is proposed based on neural networks technique (NSVM). The neural proposed regulator is shown to be able to reducing the stator reactive and stator active powers ripples and to improve performance DVC command. The validity of the proposed schemes command is confirmed by the simulation results.
Keywords :
Direct vector command , space vector modulation , neural space vector modulation , artificial neural networks
