Asynchronous Distributed Scheduling of Active Distribution Network and Thermostatically Controlled Loads With MPC-Based Aggregation

The efficient aggregation, reliable scheduling and privacy protection can improve dispatching capacities of thermostatically controlled loads (TCLs). This paper proposes a statistically feasible chance-constrained programming (CCP) model for the active distribution network (ADN) with multiple TCL aggregators in a distributed manner. The distribution system operator (DSO) minimizes the operation cost of ADN under net load uncertainties. We construct the data-driven uncertainty set from samples based on an optimal shape approximation method and reformulate joint CCP into the tractable robust optimization. For the TCL dispatching, a model predictive control (MPC) based aggregation method is proposed to build TCL aggregators and optimize the operation of TCLs under ambient uncertainties. Furthermore, an asynchronous distributed algorithm is designed to solve the coordination problem between the ADN and TCLs as well as preserving participants' privacy. Case studies demonstrate the effectiveness and reliability of the proposed model and algorithm for improving the economy of system participants under various uncertainties compared with other methods.