Artificial neural networks-based decoupling approach in the vector control block of the single-phase induction machine

Abstract

The vector control of single phase induction machine using conventional decoupling approaches has a remarkable decrease of rotor flux when the machine is powered by a real voltage source inverter. To solve this problem, we propose as an efficient solution, a decoupling approach based on artificial neural networks in the vector control block of single phase induction machine. The application of this approach to the single-phase machine increases its dynamic performance and constitutes a contribution to the study of thismachine. Indeed, this type ofmachine has not yet taken its whole share from various works present until now, compared to the threephase induction machine. For three different decoupling approaches with two types of supply: perfect voltage source and real voltage source inverter; a comparative study through numerical simulations is presented. The simulation results show the feasibility and good performance obtained by the proposed approach.