Sharing Distributed Energy Resources in Consumer Communities: The Design of Optimal Combinatorial Auctions

The high penetration of renewable energy (RE) resources is essential in the move towards sustainable development. The integration of distributed energy resources (DERs), which include RE and energy storage systems (ESS) can have many environmental and economic benefits, in addition to their technical benefits. However, the high investment and space requirements, coupled with the low utilisation of these resources when individually operated, have motivated the design of market mechanisms for sharing medium-scale (10-100 kW) DERs. This can also open the opportunity for sharing economy and economies of scale. This research proposes novel game-theoretic, optimisation, and data science approaches to design local energy markets (LEMs) for sharing medium-scale DERs in consumer communities. In the first part of this thesis, the literature on LEM design is reviewed. These advances are classified based on four design criteria, and an evaluation of the strengths and weaknesses of these design criteria is provided. Flaws in the strategic modelling of consumers are highlighted by studying the utility function formulations used in the reviewed literature. A thorough critique of the proposed combinatorial market designs is also presented. The review of these approaches has revealed several concerns about their practical implementation. These include privacy concerns, sub-optimal market outcome, and failing to capture the complimentary nature of energy resources. In the second part of this thesis, a market framework for DER sharing between multiple DER providers and a community of consumers is developed. The proposed market is based on a novel combinatorial double auction (CDA) that allocates the limited DER resources efficiently without compromising the privacy of its participants. It also supports the environmentally sustainable behaviour of consumers. Nonetheless, developing optimal pricing strategies for DER providers remains a challenge. This is addressed in the third part of this thesis, where a novel two-stage LEM mechanism for sharing DERs is proposed, combining a combinatorial clock auction (CCA) methodology with an artificial neural network (ANN) implementation. In this approach, an iterative ascending price auction for each of the DER services is presented. Then, an ANN design that uses the price-demand bidding data to learn the strategic behaviour of consumers is proposed. Optimal pricing strategies for revenue-maximising DER providers are also investigated. Simulation results show that the proposed DER sharing mechanisms can greatly reduce energy emissions while also enhancing the revenues of DER providers, when compared with existing approaches from the literature.