Forecast-Driven Stochastic Scheduling of a Virtual Power Plant in Energy and Reserve Markets
Jean-Franeois Toubeau,Thuy-Hai Nguyen,Hooman Khaloie,Yi Wang,Franeois Vallee
DOI: https://doi.org/10.1109/jsyst.2021.3114445
IF: 4.802
2021-01-01
IEEE Systems Journal
Abstract:Virtual power plants (VPPs) offer a cost-effective solution to incentivize coordination between different resources participating in joint energy and reserve markets. However, emerging technologies such as storage and demand response cannot deliver flexibility over long periods due to inherent energy limitations. In this article, we, therefore, inform the day-ahead scheduling of VPPs with forecast scenarios of the balancing stage, which are complemented with information on nonshiftable load, renewable generation, and electricity prices. These multivariate scenarios (incorporating both time and cross-variable dependencies) are obtained using a machine learning framework in which probabilistic forecasts are converted into time trajectories using a copula-based sampling. The model is enriched with a detailed representation of the intraday decision stage wherein all flexible resources are dynamically allocated over the daily horizon. To ensure the reliability of the solution, we take into full consideration the revenues and unit-specific costs related to the activation of reserves. Outcomes show that improving the representation of the intraday dispatch stage while relying on representative balancing uncertainties are two complementary components for increasing the quality of the VPP day-ahead strategy, which ultimately fosters its economic value.
computer science, information systems,telecommunications,engineering, electrical & electronic,operations research & management science