Scalable control of islanded microgrids

In the recent years, the increasing penetration of renewable energy sources has motivated a growing interest for microgrids, energy networks composed of interconnected Distributed Generation Units (DGUs) and loads. Microgrids are self-sustained electric systems that can operate either connected to the main grid or detached from it. In this thesis, we focus on the latter case, thus dealing with the so-called Islanded microGrids (ImGs). We propose scalable control design methodologies for both AC and DC ImGs, allowing DGUs and loads to be connected in general topologies and enter/leave the network over time. In order to ensure safe and reliable operations, we mirror the flexibility of ImGs structures in their primary and secondary control layers. Notably, off-line control design hinges on Plug-and-Play (PnP) synthesis, meaning that the computation of individual regulators is complemented by local optimization-based tests for denying dangerous plug-in/out requests. The solutions presented in this work aim to address some of the key challenges arising in control of AC and DC ImGs, while overcoming the limitations of the existing approaches. More precisely, this thesis comprises the following main contributions: (i) the development of decentralized primary control schemes for load-connected networks (i.e. where local loads appear only at the output terminals of each DGU) ensuring voltage stability in DC ImGs, and voltage and frequency stability in AC ImGs. In contrast with the most commonly used control strategies available in the literature, our regulators guarantee offset-free tracking of reference signals. Moreover, the proposed primary local controllers can be designed or updated on-the-fly when DGUs are plugged in/out, and the closed-loop stability of the ImG is always preserved. (ii) Novel approximate network reduction methods for handling totally general interconnections of DGUs and loads in AC ImGs. We study and exploit Kron reduction in order to derive an equivalent load-connected model of the original ImG, and designing stabilizing voltage and frequency regulators, independently of the ImG topology. (iii) Distributed secondary control schemes, built on top of primary layers, for accurate reactive power sharing in AC ImGs, and current sharing and voltage balancing in DC ImGs. In the latter case, we prove that the desired coordinated behaviors are achieved in a stable fashion and we describe how to design secondary regulators in a PnP manner when DGUs are added/removed to/from the network. (iv) Theoretical results are validated through extensive simulations, and some of the proposed design algorithms have been successfully tested on real ImG platforms.