Hierarchical Control Strategy for Load Regulation Based on Stackelberg Game Theory Considering Randomness

Abstract

Demand response has been recognized as a valuable functionality of power systems for mitigating power imbalances. This paper proposes a hierarchical control strategy among the distribution system operator (DSO), load aggregators (LAs), and thermostatically controlled loads (TCLs); the strategy includes a scheduling layer and an executive layer to provide load regulation. In the scheduling layer, the DSO (leader) offers compensation price (CP) strategies, and the LAs (followers) respond to CP strategies with available regulation power (ARP) strategies. Profits of the DSO and LAs are modeled according to their behaviors during the load regulation process. Stackelberg game is adopted to capture interactions among the players and leader and to obtain the optimal strategy for each participant to achieve utility. Moreover, considering inevitable random factors in practice, e.g., renewable generation and behavior of users, two different stochastic models based on sample average approximation (SAA) and parameter modification are formulated with improved scheduling accuracy. In the executive layer, distributed TCLs are triggered based on strategies determined in the scheduling layer. A self-triggering method that does not violate user privacy is presented, where TCLs receive external signals from the LA and independently determine whether to alter their operation statuses. Numerical simulations are performed on the modified IEEE-24 bus system to verify effectiveness of the proposed strategy.