Dynamic Cluster Partitioning of Distribution Networks With High Photovoltaic Penetration Based on K‐Means Integrated With Isolation Forest, KNN, and Whale Optimization
International Transactions on Electrical Energy Systems · 2026-01 · 2026 (1)
DOI: 10.1155/etep/7413622 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Transactions on Electrical Energy Systems, online erschienen am 14.11.2026. Verfasst von Congshan Li und 4 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Henan Provincial Science and Technology Research Project, Henan University.
Abstract (Originalsprache)
Cluster partitioning of distribution networks with distributed photovoltaic (PV) generation is a prerequisite for effective coordinated control, but the conventional K‐means algorithm is highly sensitive to initial cluster centers and easily disturbed by abnormal data, often converging to local optima. To address these limitations, this paper proposes an improved K‐means algorithm integrating isolation forest (IF), the K‐nearest neighbor (KNN) algorithm, and the whale optimization algorithm (WOA) and applies it to the dynamic cluster partitioning of high‐PV‐penetration distribution networks. The proposed method first removes outliers through a cascade model in which IF performs rapid global screening and KNN conducts fine‐grained local verification. WOA is then applied to optimize the initial cluster centers by minimizing the within‐cluster sum of squares, overcoming the local‐optimum problem caused by random initialization. On this basis, a partitioning framework is established, combining a clustering indicator system (active/reactive regulation capability and sensitivity‐based electrical distance) with an adaptive weighting mechanism driven by clustering‐quality‐oriented optimization. The main contributions of this paper are as follows. First, a two‐stage mechanism coupling cascaded anomaly detection with population‐based initial‐center optimization, which enhances the robustness and accuracy of K‐means. Second, its adaptation to PV‐integrated distribution networks through a dedicated indicator system and adaptive weighting scheme. Simulations on an IEEE 33‐node network show that, compared with genetic algorithm K‐means (GA‐K‐means), particle swarm optimization K‐means (PSO‐K‐means), conventional K‐means, and a fixed partitioning scheme, the proposed method improves voltage regulation capability by about 9.1% and load fluctuation suppression capability by about 65%, strengthening the operational stability of distribution networks under high PV penetration.
Förderung
- Henan Provincial Science and Technology Research Project
- Henan University
Zitieren
APA
Li, C., Zhang, J., He, P., Ji, Y., & Zhao, C. (2026). Dynamic Cluster Partitioning of Distribution Networks With High Photovoltaic Penetration Based on K‐Means Integrated With Isolation Forest, KNN, and Whale Optimization. International Transactions on Electrical Energy Systems, 2026(1). https://doi.org/10.1155/etep/7413622
Vancouver
Li C, Zhang J, He P, Ji Y, Zhao C. Dynamic Cluster Partitioning of Distribution Networks With High Photovoltaic Penetration Based on K‐Means Integrated With Isolation Forest, KNN, and Whale Optimization. International Transactions on Electrical Energy Systems. 2026;2026(1). doi:10.1155/etep/7413622
BibTeX
@article{Li2026,
author = {Li, Congshan and Zhang, Jialu and He, Ping and Ji, Yuqi and Zhao, Chen},
title = {Dynamic Cluster Partitioning of Distribution Networks With High Photovoltaic Penetration Based on K‐Means Integrated With Isolation Forest, KNN, and Whale Optimization},
journal = {International Transactions on Electrical Energy Systems},
year = {2026},
volume = {2026},
number = {1},
doi = {10.1155/etep/7413622}
}
EUVIMED