Optimal dispatch strategy for the agile virtual power plant

Mette Højgaard Petersen, Jan Dimon Bendtsen, Jakob Stoustrup

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

13 Citations (Scopus)
1193 Downloads (Pure)


The introduction of large ratios of renewable energy into the existing power system is complicated by the inherent variability of production technologies, which harvest energy from wind, sun and waves. Fluctuations of renewable power production can be predicted to some extent, but the assumption of perfect prediction is unrealistic. This paper therefore introduces the Agile Virtual Power Plant. The Agile Virtual Power Plant assumes that the base load production planning based on best available knowledge is already given, so imbalances cannot be predicted. Consequently the Agile Virtual Power Plant attempts to preserve maneuverability (stay agile) rather than optimize performance according to predictions. In this paper the imbalance compensation problem for an Agile Virtual Power Plant is formulated. It is proved formally, that when local units are power and energy constrained integrators a dispatch strategy exists, which is optimal regardless of future load/imbalances. The optimal dispatch is obtained at each sample by solving a quadratic program. Finally a simulation example illustrates the optimal dispatch strategy and compares the performance with a (non-optimal) MPC-strategy.
Original languageEnglish
Title of host publication2012 American Control Conference
Number of pages7
PublisherAmerican Automatic Control Council
Publication date2012
ISBN (Print)978-1-4577-1095-7
ISBN (Electronic)978-1-4673-2102-0
Publication statusPublished - 2012
Event2012 American Control Conference - Montreal, Canada
Duration: 27 Jun 201229 Jun 2012


Conference2012 American Control Conference
SeriesAmerican Control Conference


Dive into the research topics of 'Optimal dispatch strategy for the agile virtual power plant'. Together they form a unique fingerprint.

Cite this