The transportation industry is causing 14% of the worldwide greenhouse gas emission. Therefore, the United Nation concluded with the Paris climate agreement to reduce and limit their emissions to keep the global average temperature increase below 2 degrees Celsius. Based on the agreement the European union introduced new regulations to limit the average fleet emission of car companies. In this push towards lowering emission the companies introduced their first generations of electric cars. To further improve the efficiency of the electric drive train new multilevel topologies were introduced. One of the topologies is called modular multilevel converter (MMC) introduced by R. Marquardt and A. Llesnicar in 2003. This new topology main advantage are its modularity and the option to use low voltage component. However, no company industrialized the MMC for drive train application. The reason therefore are the extensive control architecture, hardware and software wise. This thesis investigated possible ways to reduce the wiring and complexity by introducing a low level communication bus together with distributing the control to remove the centralized architecture. Therefore, the different possible low level communication bus were analysed and compared. The different bus protocols were compared based on speed and time to send the same amount of information to control a MMC. It was concluded that the serial peripheral interface bus (SPI) shows the highest capability for the proposed single phase 5 level converter set-up. Therefore, an experimental system was built to validate the performance. The system itself was designed with the goal to use low cost and low voltage component. The validation of the experimental set-up showed the stable operation of the system with an open loop control. Thereby it proofs the possibility to integrate and control the MMC with a low level communication bus. The results can be used as a baseline to further research on integrating theMMC and make it viable for drive train applications.

The small and lightweight Lunar Zebro rover must survive in the harsh lunar environment for several Earth days following its moon landing. The mission of the rover is to map the radiation environment on the moon. The success of the entire mission depends on the Power Electronic System (PES), which supplies power to all subsystems and charges the batteries using solar panels. The current PES of the Lunar Zebro rover does not comply with all mission-specific requirements and does not perform satisfactorily when integrated into the rover. Therefore, the need for a reliable PES that conforms to all requirements arises for the Lunar Zebro rover. In this research, the design and implementation of an efficient, compact, and redundant PES for the Lunar Zebro rover is developed. First, the optimal Direct Current (DC) bus is designed to obtain a system with the highest efficiency. This is done by modelling the efficiency of the DC/DC converters for different bus voltages and estimating the overall losses in these converters during the deployment of the rover. Moreover, the effect of the bus voltage on the size of the passive components is investigated, and the bus voltage resulting in the most compact system is obtained. It is found that a 12 V bus results in the most efficient and compact system. No additional converter is required that regulates the 12 V output, and the inductance required for each converter is decreased compared to higher bus voltages. Besides the DC bus design, redundancy methods are compared to obtain the best tradeoff between redundancy and footprint added. The two-phase interleaved converter was found to have only an 8.59% increase in footprint compared to single-phase converters, while failure in a switch, diode, input capacitor, and output capacitor are accounted for in each converter. Finally, the mode of operation that results in the highest efficiency is obtained by designing each converter and modelling the corresponding losses for Continuous Conduction Mode (CCM) and Discontinuous Conduction Mode (DCM) operation. For both the single-phase and two-phase interleaved converters hold that operating in CCM results in a significant increase in efficiency compared to DCM operation. Moreover, the PES utilising two-phase interleaved converters is more efficient during rover operation than the single-phase counterpart. However, charging is less efficient than for the single-phase counterpart. Simulink and LTspice simulations have been carried out to verify the operation of each converter. Finally, experiments on a functional prototype are carried out to provide experimental validation of the design.

In the world of DC railway trains nowadays, asynchronous-traction machine with traction controllers, i.e. an inverter with variable output frequency, is implemented. This controller changes the incoming DC voltage into an AC voltage of a variable frequency and RMS amplitude. The disadvantages of switching inverters, compared to a pure sine wave source, are the harmonics they inject in both the output as in the input of the inverters connection. At the incoming power line, it is necessary to damp these harmonics, in order to avoid resonance issues or issues regarding other systems, for instance, train detection. For this purpose of harmonic and transient filtering, generated within as well as outside the train, an LC filter is utilised. However, this LC-filter has also influence on stability. Due to the resulting impedance to current variation in the inductance of the LC-filter, the power flow dynamics towards the train are decreased. When applying a certain amount of power, the voltage over the capacitance can become unstable very quickly. In this thesis, a graphical user interface simulation model is made to simulate these stability phenomena in Simulink in order to find the optimum size of the LC-filter in a train on the Dutch DC railway. The simulations are achieved with a model representing a generalised train. The model is simple to modify, and consists of the important parts for determining the stability of the system. Different simulations have been carried out, in order to examine the effect certain parameter variations have on the stability. Two systems have been considered, a constant power controlled system and a system where the motors had no controlling regime. An important factor is the value of the capacitance and inductance. A larger filter inductor results in an unstable system. Likewise, implementing a higher value capacitance will cause the system to be more stable. Simulations have shown that a motor controller with a simple constant power control regime is unstable with normal values for the inductance and capacitance. However a damping branch can solve this problem up to an extent. A system without any controlling regime has proven to be more stable with smaller capacitance. By utilising the model presented in this thesis, the stability of the system, consisting purely of a controlled constant power load or as a non-controlled load, can be investigated, and the impact of the LC-filter can be determined.