With the increased penetration of intermittent renewable energy sources in the power grid, power system operators need to maintain more ancillary services (AS) to ensure system reliability. Recently, demand-side management (DSM) has been proposed as a novel approach for AS provision in future smart grids. The idea is to change the power consumption behavior at the demand-side to achieve certain objectives like active power balancing and reactive power support. In this thesis, we study and tackle three issues related to the existing challenges in the implementation of DSM for AS provision.

Firstly, we focus on the issue of operating reserves provision and show how to use load aggregators to enable residential users to provide operating reserves. A hierarchical structure is int...[ Read more ]

With the increased penetration of intermittent renewable energy sources in the power grid, power system operators need to maintain more ancillary services (AS) to ensure system reliability. Recently, demand-side management (DSM) has been proposed as a novel approach for AS provision in future smart grids. The idea is to change the power consumption behavior at the demand-side to achieve certain objectives like active power balancing and reactive power support. In this thesis, we study and tackle three issues related to the existing challenges in the implementation of DSM for AS provision.

Firstly, we focus on the issue of operating reserves provision and show how to use load aggregators to enable residential users to provide operating reserves. A hierarchical structure is introduced. In the lower level, we propose an incentive compatible and social optimal mechanism for user aggregation. In the upper level, the reserve market is modeled using the conjectured supply function approach. We prove the existence and uniqueness of the market equilibrium, and analyze its efficiency. A distributed algorithm is also developed that converges to the equilibrium. We find that by applying our aggregation mechanism, both power system operators and residential users can benefit financially.

We also investigate the control problem of charging electric vehicles (EV) under uncertainty. Both synchronous and asynchronous distributed algorithms are proposed to solve this stochastic optimization problem using the historical information only. The computation burden is distributed to the EVs, thus ensuring the scalability of the algorithms. We found the convergence rate of the synchronous algorithm and obtain a sparse charging solution so that the EV battery degradation caused by frequent charging can be mitigated. Simulation results validate our proposed schemes, and show how the algorithm shaves the load peak.

Finally, we consider the issue of supporting reactive power from demand-side users with photovoltaic (PV) inverters in a distribution system. A Stackelberg game is proposed to model this problem. As the game leader, the system operator solves an optimal power flow problem to determine the reactive power prices with the knowledge of the reaction of users. As game followers, users apply an online algorithm to control the battery and PV inverter to maximize the time average revenue based on techniques in Lyapunov optimization. This algorithm has low complexity and does not need prior knowledge of the uncertain parameters. We prove that it can achieve near-optimal performance when the battery size is large. Simulation results show the increased user revenue, reduced power loss and improved voltage profile in the distribution network.