Scholarlink

<p>To enhance the privacy and security of the coupled network in the power system, this study provides a privacy preserving method based on the K-means clustering algorithm. The K-means algorithm is utilized to perform clustering analysis on the data from the power system’s coupled network, and further optimized by integrating the density peak method, resulting in the proposed DPK-means clustering algorithm. To overcome the limitation of local optima, appropriate initial centroids and the number of clusters are selected. Moreover, the BDPK-means clustering algorithm is introduced by combining the farthest distance technique with within-cluster sum of squared errors (SSE) to identify initial centroids located in densely populated areas. Additionally, the differential privacy algorithm with both sequential composition and parallel composition properties is employed. Laplace-distributed noise is added during the iterative centroid update process of the BDPK-means clustering algorithm Experimental results demonstrate that the proposed method achieves effective security and privacy preservation for the coupled network of the power system. The method exhibits superior clustering performance, high accuracy, and robust resistance against attacks, thereby successfully ensuring privacy protection for the coupled network in the power system.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Security privacy protection method of power system coupled network based on k-means clustering algorithm

  • 期刊论文

  • 发表日期：

    2026年7月14日

查看全文
  • Xiangbei Che,
  • Siliang Suo,
  • Man Li,
  • Junqiang Gong,
  • Shiqin Zeng

摘要

To enhance the privacy and security of the coupled network in the power system, this study provides a privacy preserving method based on the K-means clustering algorithm. The K-means algorithm is utilized to perform clustering analysis on the data from the power system’s coupled network, and further optimized by integrating the density peak method, resulting in the proposed DPK-means clustering algorithm. To overcome the limitation of local optima, appropriate initial centroids and the number of clusters are selected. Moreover, the BDPK-means clustering algorithm is introduced by combining the farthest distance technique with within-cluster sum of squared errors (SSE) to identify initial centroids located in densely populated areas. Additionally, the differential privacy algorithm with both sequential composition and parallel composition properties is employed. Laplace-distributed noise is added during the iterative centroid update process of the BDPK-means clustering algorithm Experimental results demonstrate that the proposed method achieves effective security and privacy preservation for the coupled network of the power system. The method exhibits superior clustering performance, high accuracy, and robust resistance against attacks, thereby successfully ensuring privacy protection for the coupled network in the power system.