Co-Optimizing Virtual Power Plant Services Under Uncertainty: A Robust Scheduling and Receding Horizon Dispatch Approach

• Published in: IEEE transactions on power systems Vol. 36; no. 5; pp. 3960 - 3972
• Main Authors: Naughton, James, Wang, Han, Cantoni, Michael, Mancarella, Pierluigi
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.09.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Co-optimization; Distributed energy resources; Distributed generation; Distributed power generation; Energy sources; Energy storage; Frequency response; Load flow; Optimal power flow; Optimization; Power flow; Power plants; Pricing; Reactive power; Receding horizon optimization; Resource scheduling; Robustness; Scenario-based robust optimization; Uncertainty; Virtual power plant; Virtual power plants
• DOI: 10.1109/TPWRS.2021.3062582
• ISSN: 0885-8950

Market and network integration of distributed energy resources can be facilitated by their coordination within a virtual power plant (VPP). However, VPP operation subject to network limits and different market and physical uncertainties is a challenging task. This paper introduces a framework that co-optimizes the VPP provision of multiple market (e.g., energy, reserve), system (e.g., fast frequency response, inertia, upstream reactive power), and local network (e.g., voltage support) services with the aim of maximizing its revenue. To ensure problem tractability, while accommodating the uncertain nature of market prices, local demand, and renewable output and while operating within local network constraints, the framework is broken down into three sequentially coordinated optimization problems. Specifically, a scenario-based robust optimization for day-ahead resource scheduling, with linearized power flows, and two receding horizon optimizations for close-to-real-time dispatch, with a more accurate second-order cone relaxation of the power flows. The results from a real Australian case study demonstrate how the framework enables effective deployment of VPP flexibility to maximize its multi-service value stack, within an uncertain operating environment, and within technical limits.