The future renewable-based power system will have an increased need for balancing power. Prosumers, having both generation and consumption capabilities, are expected to provide balancing power to the grid, if their flexibility can be appropriately managed. Meanwhile, undesirable line congestion and voltage violations may arise in the distribution network, when flexible resources respond to external control or price signals on a large scale. Hence, the development of an effective framework to coordinate flexibility at the distribution system level is of utmost importance. Such a framework should allow for an optimal provision of prosumer balancing power services within the boundaries of local network security constraints. In this study, a balancing market participation framework is proposed, adopting the concept of network-constrained transactive energy, to facilitate the interactions between the transmission system operator and aggregators who manage prosumer energy profiles. This framework retains user privacy and complies with the current market setup, where flexible energy is traded on the spot and balancing markets; however, it is ensured that the resulting energy profiles do not cause problems in the distribution network.

Due to increasing shares of renewable energy sources, more frequency reserves are required to maintain power system stability. In this paper, we present a decentralized control scheme that allows a large aggregation of refrigerators to provide Primary Frequency Control (PFC) reserves to the grid based on local frequency measurements and without communication. The control is based on stochastic switching of refrigerators depending on the frequency deviation. We develop methods to account for typical lockout constraints of compressors and increased power consumption during the startup phase. In addition, we propose a procedure to dynamically reset the thermostat temperature limits in order to provide reliable PFC reserves, as well as a corrective temperature feedback loop to build robustness to biased frequency deviations. Furthermore, we introduce an additional randomization layer in the controller to account for thermostat resolution limitations, and finally, we modify the control design to account for refrigerator door openings. Extensive simulations with actual frequency signal data and with different aggregation sizes, load characteristics, and control parameters, demonstrate that the proposed controller outperforms a relevant state-of-the-art controller.