首页|低压配电系统断路器可靠性聚类分析方法

低压配电系统断路器可靠性聚类分析方法

扫码查看
针对用户停电的时间不具有规律性,不同类型的断路器具有不同的性能特点和失效机理,为准确分析低压配电系统的可靠性,提出低压配电系统断路器可靠性聚类分析方法.收集低压配电系统在长时间运行过程中的性能退化数据,结合低压配电系统分区模型,划分性能退化数据样本空间;引入聚类算法,建立粗聚类标签矩阵;在图正则约束作用下,有序化、信息化处理性能退化数据样本,设计无标签性能退化数据聚类;在干扰因子图正则约束下,结合小路矩阵,获取低压配电系统断路器可靠性分析指标,并进行等效处理,判断低压配电系统断路器可靠性.测试结果表明,将锚点数设成 35 个,当正则项系数为 101.5 时,聚类结果的归一化互信息、聚类纯度趋于稳定,符合实际,断路器可靠性聚类分析方法具有可应用性.
Cluster Analysis Method for Reliability of Circuit Breakers in Low-voltage Distribution Systems
The time of user power outage is not regular,and different types of circuit breakers have different performance characteristics and failure mechanisms.To accurately analyze the reliability of low-voltage distribution systems,a reliabili-ty clustering analysis method for low-voltage distribution system circuit breakers is proposed.Collect performance degra-dation data of low-voltage distribution systems during long-term operation,combine the low-voltage distribution system partition model,divide the sample space of performance degradation data,introduce clustering algorithm,establish a coarse clustering label matrix,process performance degradation data samples in an orderly and informational manner based on graph regularization constraints,design unlabeled performance degradation data clustering,obtain reliability a-nalysis indicators of low-voltage distribution system circuit breakers under interference factor graph regularization con-straints,and perform equivalent processing to determine the reliability of low-voltage distribution system circuit breakers.The experimental test results show that whenthe anchor numberis set to 35 and the regularization coefficientis 101.5,the normalized mutual information and clustering purity of the clustering results tend to stabilize.The reliability clustering a-nalysis method for circuit breakers is practical and applicable.

partition modellow voltage distribution systemreliability analysispath matrixdata clustering

杨子江、程琦

展开 >

国网浙江省电力有限公司金华供电公司,浙江 金华 321000

国网浙江省电力有限公司武义县供电公司,浙江 武义 321200

分区模型 低压配电系统 可靠性分析 小路矩阵 数据聚类

2024

电工技术
重庆西南信息有限公司(原科技部西南信息中心)

电工技术

影响因子:0.177
ISSN:1002-1388
年,卷(期):2024.(23)