Robust optimization approach with acceleration strategies to aggregate an active distribution system as a virtual power plant

Abstract

An effective way to make full use of the flexibility of distributed generations in an active distribution system (ADS) is to aggregate ADS as a virtual power plant (VPP) to participate in the day-ahead energy market. This paper proposes a robust optimization model to obtain parameters of the VPP that do not depend on the information of day-ahead energy markets, such as time-varying power bounds and ramp rates. The robust model considers the uncertainty of dispatch orders from the transmission system operator (TSO) and outputs of renewable distributed generations. Furthermore, the column-and-constraint generation algorithm is applied to solve the robust model. To accelerate its solution, the period decomposition and connection method is first proposed to change the max–min sub-problem into small-scale independent linear programming problems. Second, the directional per-unit method is proposed to establish the linear relationship between uncertain dispatch orders from the TSO and the parameters of the VPP, and this relationship is substituted into the master problem to reduce iterations. Finally, the proposed method is evaluated on a modified IEEE 33-bus system and a practical 185-bus distribution system. The results verify the correctness and effectiveness of the proposed method.