Distributed scheduling of smart buildings to smooth power fluctuations considering load rebound

Congying Wei, Qiuwei Wu*, Jian Xu, Yuanzhang Sun, Xiaolong Jin, Siyang Liao, Zhiyong Yuan

*Corresponding author for this work

Research output: Contribution to journalJournal articleResearchpeer-review

Abstract

With the similar regulation characteristics as the energy storage system, the thermostatically controlled loads in smart buildings show great potential in demand response. However, its dynamic regulation characteristics may also trigger the load rebound, which misleads the power grid operator to make non-optimal regulation commands. Therefore, this paper proposes a distributed look-ahead scheduling scheme for smart buildings to smooth power fluctuations in the distribution network, where the load rebound effect is focused on. Dynamic load baselines are adopted as a link between the load regulation and load rebound. It not only reflects the impact of the load regulation on the energy consumption plans of the consumers, but also helps the power grid operator to update the regulation plan in turn. Besides, the distributed solution approach is based on the column generation where the Dantzig-Wolfe decomposition and branch-and-bound methods are combined. It adapts to the mixed integer problems and the reuse of columns can speed up both the decomposition and bounding processes. The simulation results show that, the proposed scheduling scheme can make full use of the load rebound to improve the regulation efficiency, and the solution methodology can be applied in real-time even though the decomposition algorithm is called repeatedly.
Original languageEnglish
Article number115396
JournalApplied Energy
Volume276
Number of pages13
ISSN0306-2619
DOIs
Publication statusPublished - 2020

Keywords

  • Distribution network
  • Distributed optimization
  • Load rebound
  • Power fluctuation
  • Column generation

Fingerprint

Dive into the research topics of 'Distributed scheduling of smart buildings to smooth power fluctuations considering load rebound'. Together they form a unique fingerprint.

Cite this