The increasing focus and global support for the transition to electric vehicle traction systems drive rapid advancements in electric vehicle technologies. One essential component of electric vehicles is the main traction converter, which controls the power flow and drives the electric motor. Additionally, a dedicated arrangement is required for managing the batteries state of charge and the charging process. The modular multilevel converter with submodules that comprises half-bridges and batteries is currently being investigated for its suitability in traction systems. This converter offers many advantages in terms of integration, modularity, and redundancy. It combines multiple functions of various electrical components within the vehicle, and it can partially support vehicle operation even during undesired battery conditions.

In this research, a real-time emulation approach of the modular multilevel converter is developed using MicroLabBox/dSPACE to investigate the operation of the converter. A physical prototype of the converter is also constructed in the laboratory and compared to the real-time model to validate its accuracy and effectiveness. This model is implemented by utilising a single look-up table in each arm of the converter. The dynamic performance of the converter in the emulated system resembles the dynamic operation of the converter prototype, confirming its fidelity and suitability for further investigations. The developed model is then integrated into a simplified real-time model of the vehicle to demonstrate the converter’s potential in driving the motor, balancing battery state of charge, and regenerative braking.

Furthermore, the research investigates a novel approach for balancing the average state of charge of the converter phases and demonstrates the override capability of the converter through the integration of an override subroutine in the central control and balancing system. The balancing arrangement is independent on the system parameters and reduces the number of controllers required in the converter’s conventional balancing arrangement. This approach balances the average state of charge of the converter legs during different vehicle driving conditions. This investigation also shows that a balanced converter operation can be achieved while overriding particular submodules due to undesired conditions.

A practical approach to minimise the sensing devices in the converter is developed in this thesis. Instead of using hundreds of sensors, only six sensors are utilised to estimate the voltage of the individual submodules. This approach reduces the cost and minimises the potential points of failure in the system. Also, this approach reduces the required analogue inputs in the controller system, particularly for systems with a large number of submodules.

Finally, this research investigates the utilisation of the converter as an onboard charger without altering the converter connection or adding extra components. This charging arrangement enables faster charging of batteries with a lower state of charge and ensures equalised charging of all batteries in the converter. The charging arrangement is for a single-phase source and can be used for three-phase source charging if needed.

In summary, this thesis studies the utilisation of the modular multilevel converter as a central traction converter in electric vehicles. The investigation provides an optimal and reliable converter integration to achieve multiple functions in the electric vehicle system. The optimisation includes a balanced converter operation under various driving conditions and implementing a submodule override algorithm. The reduced number of sensors enhances the overall reliability of the converter while reducing the cost. The research is conducted using a validated real-time emulated model of the converter in conjunction with a laboratory prototype converter. This work can also be extended to investigate many potential electric vehicle drive system optimisations.