Thesis Topics

The research activity aims to investigate the possibility of using multilevel converters, such as the Cascaded H-Bridge (CHB), the Modular Multilevel Converter (MMC) or reconfigurable multilevel structures, to replace the conventional two-level inverter in BESS grid interface. The main reason is related to the possibility of subdividing the battery pack cells into smaller groups with reduced supply voltage directly connected to the individual cells of the multilevel converter. Ideally, each module is connected to a single battery cell. The advantages of using this application solution include the possibility of using low-power semiconductor devices, reducing costs and losses, eliminating the need for a BMS whose functions can be directly managed by the converter’s individual cells, improving the quality of voltage and current waveforms, resulting in increased efficiency, and reducing the volumes and costs of filters and passive elements of the converter. However, the complications from a topological, constructional, and control point of view pose challenges for the implementation of this technology in the automotive field, representing highly current and innovative research topics. Thesis activities in this field can include:

• Design and construction aspects related to the optimization of topologies, devices, and PCBs.
• Hardware implementation of experimental prototypes and testing.
• Development of innovative modulation techniques to improve system performance, power quality, and efficiency.
• Development of control techniques for battery-side management (balanced management of battery cell charging and discharging).
• Study and comparison of different topological solutions with respect to current standards of efficiency, performance, volumes, weights, and costs.

DC microgrids are more and more becoming an interesting solution adopted in modern power grids. A DC microgrid is a localized power system that distributes and manages direct current (DC) electricity among interconnected sources (like solar panels, fuel cells, and batteries) and DC loads (such as LED lighting, electronics, and EV chargers). Unlike traditional AC microgrids, DC microgrids eliminate multiple AC/DC conversions, improving efficiency, reliability, and compatibility with renewable energy and storage systems. They are increasingly used in remote areas, data centers, and smart buildings, offering flexible, resilient, and energy-efficient power management solutions.

This activity aims to develop innovative architectures for the integration of storage systems with RES within DC microgrids, exploring multi-stage and/or multi-port conversion topologies with a minimal number of components, high efficiency, low cost, and optimized control systems for power flow management, in order to maximize energy extraction from RES and storage efficiency.

Thesis topics in this area may include:

• Study, design, and development of multi-stage and/or multi-port bidirectional conversion architectures for the integration of storage systems, RES, in a DC microgrids including a port for eventual AC connection, aimed at maximizing power extraction from RES, conversion and storage efficiency, reliability, and power quality.

• Design and construction aspects of converters focused on topology optimization regarding efficiency, cost, and power density.

• Development and implementation of control systems for power flow management.

• Study and development of innovative modulation and control techniques for the interface between batteries and RES.

Converters in power-electronics-dense grids

With the advent of smart grids, the increase in distributed generation, and the spread of loads requiring regulated power supplies, more and more power electronic converters are being connected to the electrical distribution system. This increases the injection of harmonics into the grid with power quality reduction and generate unwanted interactions among different converters. This issue is particularly critical in weak or isolated networks, such as microgrids or those on ships and airplanes. In these contexts, converters—often from different manufacturers—interact in an uncoordinated manner. Although designed for standalone operation, once integrated into the same network, their performance degrades, often causing instability.

This research activity aims to investigate innovative solutions for integrating converters and systems into the grid through advanced power flow control techniques. These include the estimation of network impedance and parameters as seen from the converter and optimizing control actions based on the grid’s status and predefined cost functions.

Thesis topics in this area may include:

• Study of the effects of converters on grid stability and system modeling.

• Development and implementation of techniques for estimating the network impedance seen at the terminals of connected converters.

• Development and implementation of optimized control techniques based on the type and state of the network in which the converter is integrated.

Electric drives are at the heart of electric mobility, industrial automation, aerospace and renewable generation. Their performance depends on how the power converter and the electric machine are controlled together: dynamic response, efficiency, torque quality, robustness to parameter variations and fault tolerance. Research in this area covers the modelling and control of drives with permanent magnet synchronous machines, induction machines and multiphase machines, from classical field-oriented control to advanced strategies such as Model Predictive Control (MPC), sensorless control and fault-tolerant control, implemented on microcontrollers, DSPs and FPGAs. PELAB has an experimental motor-drive setup in the laboratory, so the techniques developed in the thesis can be tested and validated on real hardware. Thesis topics in this area may include, for example:

• Design and implementation of high-performance control strategies (field-oriented control, direct torque control, MPC) for electric drives.
• Sensorless control and online estimation of machine parameters.
• Fault-tolerant control of multiphase machines and converters.
• Digital implementation on DSP and FPGA platforms and experimental validation on the lab setup.

Artificial intelligence is changing how power converters are modelled, monitored and controlled. This research combines data-driven methods with the physics of power electronics to build tools that are fast, accurate and suitable for real-time and embedded applications. Main research lines:

• Grid impedance estimation: physics-informed neural networks (PINN) and LSTM architectures that estimate the grid impedance seen by the converter from operating transients, to adapt its control and keep the grid stable.
• Parameter estimation and digital twins: sequential learning and physics-informed models that identify converter parameters online and create digital twins for predictive maintenance and ageing assessment.
• AI-based and model-free control: control strategies that learn from data or adapt online, reducing the dependence on an accurate model of the system.
• Fault diagnosis and condition monitoring: LSTM, CNN and Convolutional-Transformer networks for fast fault detection and localization in converters with many switches.

Thesis activities can range from the development and training of the AI models to their implementation on embedded platforms and experimental validation in the lab.

AI-driven hyperscale data centers are a new class of electrical load. A single site can draw hundreds of megawatts, its consumption can change quickly following the AI workloads, and the whole load is connected to the grid through power electronics. How these facilities are supplied from the grid has become a key question for both data center operators and grid operators. This research studies the grid-side power supply architectures of data centers: medium-voltage interfaces, active front-end and PFC rectifiers, UPS systems, solid-state transformers and the emerging 800 V DC distribution, together with battery storage for support and peak shaving. Thesis topics in this area may include:

• Modelling and design of grid-side conversion stages (active front-end, PFC, solid-state transformers) for data center power supply.
• Fault ride-through (LVRT/HVRT) and compliance with grid-code requirements for large loads.
• Stability analysis of grids with large converter-interfaced loads and grid-forming / grid-supporting data center interfaces.
• Impact of dynamic AI workloads on the grid and mitigation with storage systems.
• Validation through EMT simulation and Hardware-in-the-Loop (HIL) testing.

Isolated DC/DC converters for data centers, transportation and energy Isolated DC-DC converters, such as Dual Active Bridge (DAB) converters and LLC resonant converters, are fundamental components across diverse sectors due to their high efficiency, compact size, and galvanic isolation, which ensures safe and reliable power transfer between voltage domains. They enable critical functions in: • Data center power supplies: used to obtain a low-voltage (LV) DC output (e.g., 12 V or 48 V) from a typical 400 V DC bus, or from the emerging 800 V DC power supply architectures. • Medium-voltage DC (MVDC) grids: for integrating renewable energy sources, interconnecting different voltage levels, and providing galvanic isolation to enhance system reliability. • Electric and hybrid vehicles (EVs and HEVs): to derive a low-voltage DC (LVDC) bus from the high-voltage (HV) battery pack. • Transportation systems beyond EVs and HEVs, including railways and more-electric aircraft, where their high power density and efficiency contribute to reduced size and weight. Thesis topics in this area include: • Design and simulation of DAB converters using conventional or novel control schemes. • Design and simulation of LLC converters in open-loop operation or with voltage control. • Frequency-domain and time-domain modeling techniques for DAB and LLC converters. • Efficiency evaluation through analytical and simulation-based methods.

Nuclear fusion represents a cutting edge technology under development and a central topic for many academic institutions and companies around the world. Europe currently leads the most ambitious fusion experiment, ITER, located in Cadarache, France. Italy contributes to this effort through the Divertor Tokamak Test facility in Frascati. In a fusion reactor, plasma at extraordinary high temperatures must be confined in vacuum using magnetic fields generated by currents controlled by power electronics devices, making this branch of engineering fundamental for achieving stable plasma confinement. A thesis in this area may deal with the design and optimization of a power electronics converter and the control strategies required to generate the currents used for plasma confinement. Alternatively, it may address the characterization of electrical systems dedicated to plasma heating, such as neutral or negative beam injection units. The main challenges set by such a thesis are not only the control of the converters but also the management of very high currents which can reach up to tens of kA. A work project of this kind would imply: – studying the fundamentals of plasma fusion and the role of coils and actuators inside the reactor – modelling the most suitable converter architecture to meet plasma-current requirements – design of the most suitable control strategy for the converters – analysis of the semiconductors devices needed to implement the system