Data-Driven Dispatchable Regions with Potentially Active Boundaries for Renewable Power Generation: Concept and Construction
The dispatchable region of volatile renewable power generation (RPG) quantifies how much uncertainty the power system can handle at a given operating point. State-of-the-art dispatchable region (DR) research has studied how system operational constraints influence the DR but has seldom considered the effect of the uncertainty features of RPG outputs. The traditional DR is generally described by a large number of boundaries, and it is computationally intensive to construct. To bridge these gaps, a novel type of DR is defined, which is enclosed by potentially active boundaries (PABs) that consider the operational constraints and uncertainty features of RPG outputs. The proposed DR is easier to construct because the PABs are only a small part of the traditional DR boundaries. The procedure for constructing the proposed DR is described in terms of the progressive search for PABs, which is formulated as a mixed-integer linear program by incorporating the discrete observed data points of RPG outputs as an approximate distribution. A parallel solution paradigm is also developed to expedite the construction procedure when using a large observed dataset. Simulation tests on the IEEE 30-bus and 118-bus systems verify the effectiveness and scalability of the proposed DR and the efficiency of the proposed algorithm.