EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Forecast-Driven Aggregation and Real-Time Coordination of Demand-Side Flexibility for Balancing Market Participation

Arqum Shahid, Argo Rosin, Roya AhmadiAhangar, Tarmo Korotko, Noman Shabbir

IEEE Access · 2026 · 14 · 107225-107241

DOI: 10.1109/access.2026.3711805 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Arqum Shahid und 4 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Project “Increasing the Knowledge Intensity of Ida-Viru Entrepreneurship, European Commission, European Union and Estonian Research Council.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Residential aggregators remain underrepresented in fast-acting balancing services, such as Frequency Containment Reserve for Normal operation (FCR-N), due to behavior-dependent flexibility, challenges in accurate forecasting with limited target-domain data, heterogeneous portfolio coordination, and stringent dynamic requirements for real-time delivery. This paper presents an integrated forecast-driven aggregation and real-time coordination framework with closed-loop RTDS validation, facilitating residential aggregator participation in the FCR-N balancing market. A transfer learning-based multi-input multi-feature neural network (MIMF-NN) is developed to generalize from representative households, enabling community-level flexibility forecasting. This flexibility is further optimized through mixed-integer linear programming (MILP) to identify market-compliant participation windows, accounting for device dynamics by mapping dynamically deliverable capacity. To effectively allocate commitments among individual households, a risk-aware hierarchical disaggregation strategy is employed, utilizing availability-weighted driver–energy quantiles and CVaR-lite caps. Upon activation within the 15-minute FCR-N interval, real-time forecast–delivery mismatches are addressed through MILP-based coordination, leveraging updated resource-availability signals. The framework is validated using a real-time digital simulator (RTDS) with the IEEE 9-bus testbed system, demonstrating compliance with FCR-N dynamic performance requirements. Numerical results indicate that MIMF-NN achieves an NRMSE of 0.092 and an R2 of 0.925 under high comfort disturbance. During RTDS-based FCR-N activation, the aggregator delivered 2.10 kWh against a requirement of 2.03 kWh, resulting in a mean coverage ratio of 1.033 and a mean absolute tracking error of 1.74 kW, thereby demonstrating the practical viability of the proposed framework for reliable balancing market participation.

Förderung

  • Project “Increasing the Knowledge Intensity of Ida-Viru Entrepreneurship
  • European Commission
  • European Union and Estonian Research Council

Zitieren

APA

Shahid, A., Rosin, A., AhmadiAhangar, R., Korotko, T., & Shabbir, N. (2026). Forecast-Driven Aggregation and Real-Time Coordination of Demand-Side Flexibility for Balancing Market Participation. IEEE Access, 14, 107225-107241. https://doi.org/10.1109/access.2026.3711805

Vancouver

Shahid A, Rosin A, AhmadiAhangar R, Korotko T, Shabbir N. Forecast-Driven Aggregation and Real-Time Coordination of Demand-Side Flexibility for Balancing Market Participation. IEEE Access. 2026;14:107225-107241. doi:10.1109/access.2026.3711805

BibTeX
@article{Shahid2026,
  author = {Shahid, Arqum and Rosin, Argo and AhmadiAhangar, Roya and Korotko, Tarmo and Shabbir, Noman},
  title = {Forecast-Driven Aggregation and Real-Time Coordination of Demand-Side Flexibility for Balancing Market Participation},
  journal = {IEEE Access},
  year = {2026},
  volume = {14},
  pages = {107225--107241},
  doi = {10.1109/access.2026.3711805}
}
Kostenlos starten →