EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Probabilistic Forecasting of PV Adjustable Capacity and Two‐Timescale Volt‐Var Control in Active Distribution Networks

Wei Li, Yichuan Shi, Yiming Qian, Kai Ding, Jianyu Luo, Jianxing Fu, Ying Wang, Huaxi Yu

IET Generation, Transmission & Distribution · 2025-01 · 19 (1)

DOI: 10.1049/gtd2.70185 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IET Generation, Transmission & Distribution, online erschienen am 06.11.2025. Verfasst von Wei Li und 7 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The high penetration of renewable energy and increasing load volatility in active distribution networks have led to rapid voltage fluctuations, posing significant challenges to conventional voltage‐var control (VVC) strategies. Photovoltaic (PV) inverters offer strong potential for voltage regulation by utilizing their adjustable capacity. However, existing studies largely overlook the need for forecasting this capacity. Additionally, effective VVC based on PV adjustable capacity (PVAC) remains difficult to achieve under uncertainty and communication constraints. Therefore, this paper proposes a multi‐timescale VVC framework that integrates PVAC forecasting with voltage regulation based on deep reinforcement learning. First, the PVAC available for VVC is formally defined, and an enhanced Transformer‐based model is developed to forecast its probabilistic intervals. Second, a hybrid‐action deep reinforcement learning (DRL) algorithm is proposed to coordinate continuous inverter VAR support and discrete tap/capacitor switching. To improve robustness under limited communication, the method incorporates a partially observable Markov decision process (POMDP) formulation and recurrent policy networks. Simulation results demonstrate that the proposed approach provides accurate PVAC interval predictions, offering enhanced robustness compared to point forecasting. Extensive VVC experiments on the IEEE 123‐bus feeder confirm improved voltage regulation, reduced power losses, and fewer discrete control actions, even under communication impairments.

Zitieren

APA

Li, W., Shi, Y., Qian, Y., Ding, K., Luo, J., Fu, J., Wang, Y., & Yu, H. (2025). Probabilistic Forecasting of PV Adjustable Capacity and Two‐Timescale Volt‐Var Control in Active Distribution Networks. IET Generation, Transmission & Distribution, 19(1). https://doi.org/10.1049/gtd2.70185

Vancouver

Li W, Shi Y, Qian Y, Ding K, Luo J, Fu J, et al. Probabilistic Forecasting of PV Adjustable Capacity and Two‐Timescale Volt‐Var Control in Active Distribution Networks. IET Generation, Transmission & Distribution. 2025;19(1). doi:10.1049/gtd2.70185

BibTeX
@article{Li2025,
  author = {Li, Wei and Shi, Yichuan and Qian, Yiming and Ding, Kai and Luo, Jianyu and Fu, Jianxing and Wang, Ying and Yu, Huaxi},
  title = {Probabilistic Forecasting of PV Adjustable Capacity and Two‐Timescale Volt‐Var Control in Active Distribution Networks},
  journal = {IET Generation, Transmission & Distribution},
  year = {2025},
  volume = {19},
  number = {1},
  doi = {10.1049/gtd2.70185}
}
Kostenlos starten →