Game-theoretic methods for small-scale demand-side management in smart grid

Abstract

Increasing electricity demand with inadequate power generating resources has become a key challenge for power system operators. Adding extra power generating sources to meet peak electricity demand can reduce the sustainability of the electricity grid and increase energy generation cost to power system operators. Demand-side management is essential to smart grid because it manages peak electricity demand without requiring major upgrades to existing grid infrastructure. Successful demand-side management strategies include employing distributed energy resources, such as renewable energy generation, near end-users to satisfy users' peak energy demand.

The primary goal of this dissertation is to devise and study novel decentralised energy trading systems for small-scale demand-side management incorporating a community energy storage (CES) device and customer-owned photovoltaic (PV) energy generation. First, we introduce an energy trading system between a CES device and residential PV energy users in a neighbourhood area network. In the energy trading system, users decisions to minimise their individual energy costs are studied by implementing a non-cooperative dynamic game. Then, we develop a hierarchical energy trading system between the CES device and the users, implementing a non-cooperative Stackelberg game, to enable both the CES operator and the residents to maximise their individual payoffs. Our results demonstrate that the proposed systems offer significant electricity cost savings for users, increased CES operator revenue, and peak-to-average load ratio reductions. Moreover, the systems are shown to be robust to imperfect information with significant forecast errors in demand and PV energy.

Next, the impacts of non-ideal participating actions of users, that are not completely rational, on the hierarchical energy trading are investigated using a prospect-theoretic approach. The results demonstrate that the benefits of the energy trading system are robust to users' actions when they significantly deviate from complete rationality. Finally, we use prospect theory to investigate the impact of non-ideal participating actions of multiple electric vehicle (EV) aggregators on a non-cooperative game-theoretic EV charging system. Extensive numerical analysis shows that the performance of the EV charging system, in terms of energy cost reductions and peak-to-average ratio reductions, is resilient to aggregators' non-ideal participating actions.