期刊
ELECTRIC POWER SYSTEMS RESEARCH
卷 220, 期 -, 页码 -出版社
ELSEVIER SCIENCE SA
DOI: 10.1016/j.epsr.2023.109366
关键词
Power quality; Comprehensive assessment; Quantum clustering; Pattern recognition
This paper proposes an analysis method based on Quantum clustering to address the imbalanced structure of power quality monitoring data. The method effectively distills diverse abnormal patterns from imbalanced data and allows for easy detection of regions with issues exceeding defined limits. By quantifying cluster and region severity indices, the method provides a reliable assessment of power quality performance.
Steady-state power quality (PQ) indices seldom exceed the limits in power systems. Therefore, steady-state PQ monitoring data show a distinct imbalanced structure (i.e. unequal distribution). To ensure the efficient assessment of PQ performance, new and more appropriate tools for analyzing these data are needed. In this paper, an analysis method based on Quantum clustering is proposed, aiming to address the imbalanced structure of monitoring data. Firstly, the imbalanced structure of PQ monitoring data is analyzed. Secondly, Quantum clustering is performed on the PQ monitoring data and several disparate patterns of clusters are recognized. Thirdly, cluster severity index of cluster and the region severity index of region are defined to quantify the PQ performance of a cluster and a region, respectively. The cluster severity index is defined according to the sum of the proportions between cluster centers and limit for the respective PQ index. The region severity index is defined according to the proportions of PQ monitoring data of each region belonging to the different clusters with the corresponding cluster severity index. By sorting the region severity index in descending order, regions with poorer PQ performance become a higher-ranking place, which enables an easy comparison among regions. Finally, the proposed method is applied to the PQ monitoring data of 11 regions in a large city in China. The method can effectively distill diverse abnormal patterns from substantially imbalanced monitoring data, which is beneficial for easy and automatic detection of regions having issues with PQ indices exceeding defined limits.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据