From hydropower to wind power, renewable energy is converted to electricity, putting the electric grid at the crossroads of the green energy generation and consumption. The grid must also adapt and scale up to an increased capacity and to the stochastic nature of the wind and solar production. Consumption is also on the rise as the world turns more electric and adopts new power-hungry lifestyles, from driving electric vehicles to consuming data centers-based content. These evolutions impact the grid in many ways. The transfered energy not only grows in volume but also in location, as there is not anymore a clear geographical separation between remote large generation points and consumption points concentrated in the power distribution networks of cities and factories. With the introduction of photovoltaic panels that can be mounted on the roofs of buildings or the use of battery storage systems, the production and storage can now be done at the distribution network level with consumers becoming producers as well. Secondly, the DC nature of some of the loads and sources (e.g., electric vehicle, photovoltaic power sourcesâ Š) shakens the AC-only grid paradigm towards hybrid AC-DC and DC-only grids. Finally, this evolution has also led to the massive adoption of power electronics converters in the grid, from its utilisation in energy generation resources (e.g., inverters of photovoltaic power sources) to distribution networks (e.g., active-front end converters interfacing AC and DC nodes) and eventually, to consumers (e.g., electric vehicle chargers). In this context, converters must be operated reliably and be integrated to networks easily. These grid requirements enable a large variety of opportunities for power electronics researchers and in this thesis, two main directions have been investigated, with the support of a full-scale low-voltage AC-DC microgrid. Amongst all the converters concepts, the DC transformer as the DC counterpart of the AC transformer is one that presents a strong potential for a wide adoption in power distribution networks, with its galvanic isolation and its simple natural power flow-based operation. As a first research direction, in this thesis, its integration in AC-DC power distribution networks and microgrids has been investigated. By enabling novel power distribution network controls, DC transformers affect largely the power flow dynamics and a framework for estimating the grid dynamics is proposed. This thesis also presents grid operation techniques such as grid reconfiguration and optimal power flow-based central control, where grid-connected converters play key roles. The interaction between the impedances of the AC grid and of the grid-connected converters can greatly impact the overall grid stability and a second research direction focuses on proposing methods to estimate the grid impedance accurately during grid operation. A new impedance estimation methodology based on discrete Fourier transforms is proposed. The use of existing grid-connected converters for perturbation injection in the grid is then investigated. A perturbation is employed to estimate the grid impedance and must be of desirable magnitude to allow correct estimation. Several techniques are proposed here to circumvente some of the many limitations of such converters and enable a perturbation injection function in grid-connected converters.
These contributions could strengthen the adoption of power electronics in future power grids.