Contribution to the Improvement of Control Strategies for Electric Drive Systems Dedicated to Electric Vehicles

Abstract

This thesis presents the development, implementation, and experimental validation of advanced sensorless control strategies for induction motor–based electric vehicle (Ev) drive systems. To address the growing demand for efficient, reliable, and costeffective electric propulsion, the proposed approaches eliminate mechanical speed sensors, thereby reducing system cost and complexity while enhancing robustness in harsh operating environments. The first contribution of this work is a lowcost sensorless scalar control strategy based on a SecondOrder Generalized Integrator FrequencyLocked Loop (SogiFll), in which rotor speed and stator frequency are estimated in real time using only a single stator current sensor. This configuration significantly reduces hardware requirements and computational burden, making it particularly suitable for highspeed operating regions and applications where simplicity and energy efficiency are prioritized. Experimental results demonstrate stable operation and satisfactory speed regulation at medium and high speeds, although performance degrades at reduced speeds due to inherent limitations associated with fixedslip operation. To overcome these limitations and enhance the dynamic performance of the drive system, an improved sensorless Direct Torque Control (SDtc) strategy is subsequently developed using a MultipleEnhanced SogiFll (MesogiFll) estimator. In this scheme, two stator current sensors are employed to enable more accurate estimation of torque and stator flux. The MesogiFllimproves harmonic and DCoffset rejection as well as frequency estimation accuracy, resulting in faster torque response and superior flux regulation under dynamic operating conditions. While the proposed sensorless Dtc approach significantly enhances overall drive performance, its effectiveness remains constrained in very lowspeed regions, where estimation sensitivity increases. Comprehensive simulation studies are conducted using Matlab/Simulink, followed by realtime experimental validation on a dSPACE MicroLabBox control platform using ControlDesk. The experimental test bench, comprising an induction motor, a power electronic inverter, and realtime control hardware, enables thorough evaluation of the proposed strategies under realistic operating scenarios. Overall, the results confirm that the sensorless scalar control offers a simple and economical solution for highspeed applications, whereas the sensorless Dtc scheme provides superior torque and flux control for demanding dynamic conditions, highlighting the complementary nature of the proposed approaches for electric vehicle drive systems. KeywordsInduction Motor (Im), Direct Torque Control (Dtc), Speed Sensorless Control (Ssc),