电能质量监测数据中间距离法聚类分析

Middle Distance Clustering of Power Quality Monitor Data

电能质量监测点不断增多,导致电能质量监测数据规模的爆炸性增长。有效地分析电能质量监测数据,并从中获取有用的信息是一个亟待解决的问题。基于充分吸收大数据的思想方法,采用中间距离法对某区域电能质量监测中心的监测数据开展聚类分析。首先选定各种不同的电能质量指标作为聚类变量,然后利用统计指标确定聚类个数,最后对聚类结果中的电能质量指标特征及样本特征进行综合分析,确定有效的聚类分析结果并提取出某些典型用户类型的电能质量特征。算例分析结果表明了方法的有效性。研究结果对电能质量问题的监管、分析和有针对性地提出治理策略有重要的辅助作用。

With the increasing number of power quality monitors （PQMs）, the monitoring data are growing quickly. It is necessary to analyze the big data of PQM and extract some important information for the power quality （PQ） solutions. This paper fully absorbs the thinking of big data, and applies the middle distance clustering into the analysis of PQM da- ta from certain region. Firstly, the combinations of PQ indicators are selected as the clustering variables. Secondly, the statistical indicators are used to determine the number of clustering categories. Finally, the valid clustering results are used to extract the characteristics of different power users. The case study shows the effectiveness of the proposed ap- proach. The clustering results are helpful in assisting the supervision, analysis and solution of the PQ problems.