This Ph.D. thesis focuses on a novel combination of a frequency converter and an electric machine specially dedicated to traction drive and battery recharging modes of an electric vehicle (EV). This power architecture is composed of a six legs voltage inverter connected to a three-phase open-end winding machine. Chapter II details a quantitative comparison between two industrial power architectures and concludes that the SOFRACI powertrain is a competitive solution in terms of power converter efficiency, drive mechanical performances, and required silicon area.This architecture offers the attractive possibility of combining three important functions: traction and braking, battery charging and connecting the energy storage to a smart grid. In addition, this topology offers several advantages such as various motor feeding possibilities and a high degree of reconfiguration in degraded operating mode. The third and fourth chapters of this thesis concern the optimization of control strategies with regard to two types of faults: firstly the inherent imperfections in the converter itself (non modeled non-linearity and ineffective synchronization of control values) and secondly accidental failures. In the first case, an analysis of the pulse width modulation (PWM) methods enables the creation of a PWM strategy with a very low sensitivity to PWM uncertainties and the non-linear behavior of the power converter.In the second case, in the event of a faulty semiconductor device, it is shown that a hardware reconfiguration is required to enable an emergency traction mode. The sustainability of the traction mode is then examined with respect to the control strategy. This analysis leads to an innovative control structure based on basic and easy to implement solutions. Finally, the degraded mode operation principles have been extended to normal mode operation for the purpose of enhancing the cycle efficiency.
