Novel Methodologies for the Analysis and Management of Low Voltage Active Distribution Networks

The continuous proliferation of renewable Distributed Energy Resources (DERs) in Medium Voltage (MV) and Low Voltage (LV) Distribution Networks (DNs) is playing a prominent role in reducing the carbon footprint of fossil-fuel-based electricity generation plants. However, the rapid integration of these resources has also changed the passive nature of these networks, and consequently, special attention is required to ensure their optimal, secure and reliable operation. Moreover, the installed DERs cannot be simply operated on the basis of a fit-and-forget policy due to the emergence of several technical problems caused by their interconnection. Issues such as voltage unbalance, reverse power flow, unintentional islanded operation etc. are frequently occurring in these networks which, in turn, require novel control and management schemes to ensure the continuous supply of electricity to end-users. In this context, several countries and network operators have updated their grid codes which put additional constraints on renewable DERs to participate in the provision of ancillary services in the context of maintaining network stability. In view of this, multi-phase analysis of MV and, especially, LV networks gains unprecedented attention due to the non-capability of single-phase equivalent representation-based network analysis approach in depicting the true picture of system’s operating conditions and state variables. Such a detailed network analysis, in turn, sets the path for innovative solutions to effectively manage these networks. Furthermore, concerning LV DNs, neutral conductor must be treated like phases conductors due to the significant presence of current in it under highly unbalanced network loading scenario. Resultantly, the application of the Kron reduction methodology to these networks must be avoided because of its reliance upon the unrealistic assumption of an equipotential neutral conductor. Since LV active DNs will become a central pillar of a decentralized electrical grid in near-future, it is, therefore, incumbent to analyse and manage them by taking into account each aspect of their structure. With a special emphasis and focus on LV active DNs, this thesis presents novel analysis and management techniques for such networks by taking into account the explicit representation of both phases and neutral conductors in the context of losses management, which includes both losses minimization and losses allocation concepts in its scope. Regarding the latter notion, a novel multi-phase losses allocation procedure is proposed which avoids allocating neutral losses to phases conductors and, consequently, provides explicit information about phase- and neutral-losses allocation factors. The proposed approach successfully takes into account the impact of unbalanced loading through self- and mutualvariable losses coefficients while allocating losses to end-users. Based on the information provided by positive neutral losses allocation factors, a novel unbalance reduction strategy is further developed which allows the commutation of loads from heavily loaded phases of a critical node to its lightly loaded phases and, therefore, improves network balancing significantly. In the framework of losses reduction, a multi-phase Optimal Power Flow (OPF) model based on Semi-Definite Programming (SDP) technique is formulated which successfully incorporates neutral conductors as well as wye- and delta-connected loads in its formulation. Moreover, a novel complex voltage variable-based approach is proposed for the incorporation of a complete ZIP load in the proposed relaxation. However, due to excessively high computational requirements of this relaxation in the case of medium- and large-size DNs, a cheap SDP-based OPF model is additionally proposed and three novel propositions are developed in this context for the modelling of constant current loads. The proposed cheap relaxation can be practically realized due to its low computational requirements and provision of exact solution under a range of ZIP load parameters. However, due to the dominance of multi-phase SDP-based approach over cheap SDP-based relaxation, the latter relaxation is further tightened by imposing additional constraints in the form of convex envelopes. As a result, a tight solution is obtained without making any compromise on the computational benefits of cheap SDP-based OPF relaxation. Finally, all the proposed strategies are simulated on several test cases to demonstrate their positive potential in the context of LV active DNs management.