Neuro-fuzzy control of an input output linearization of a permanent magnet synchronous machine fed by a three levels inverter

In this paper, we present the linearizing control technique controlled by a neuro-fuzzy regulator applied to the permanent magnet synchronous machine (PMSM). It permits decoupling and linearizing the system without taking into accounts the flux orientation. The nonlinear control (NLC) applied to the PMSM decompose the system into two mono variable, linear and independent subsystems. The neuro-fuzzy control permits to the speed and the I_d current control is carried out by neuro-fuzzy regulators (ANFIS). The analysis of the results obtained by this type of nonlinear regulator shows the robustness characteristic with respect to the load perturbations and the parametric variations. A qualitative analysis of the evolution of the principal variables describing the behaviour of the global system (PMSM-Inverter (PWM)-Control) is developed by several tests of digital simulation in last stage.