•  
  •  
 

Keywords

power consumption data analysis, big data, clustering, Graph, cloud computing

Abstract

As the deepen development of smart grids construction, more and more monitoring data are collected in power distribution network, and gradually forming big data on the user side of smart grids. The traditional data analysis model does no longer meet performance requirements of huge data processing. Thereby new storage and data analysis models are urgently needed to be established. Under this background, this paper proposes a cluster analysis method for massive electricity consumption data based on the Alibaba Cloud big data analysis platform MaxCompute. The characteristics of power consumption data are fully considered, and a multilevel partition tablebased power consumption data storage mode are designed by adopting the threephase voltage, threephase current, threephase power factor, etc. to establish multidimensional data characteristics. Furthermore, the MaxCompute Graph framework design is employed to achieve an efficient clustering and partitioning algorithm for massive electricity consumption data. Experimental results show that the designed storage mode can effectively improve the retrieval efficiency of electricity consumption data. The clustering accuracy rate reaches 88% for clustering of users with different electricity consumption types. The effectiveness and high efficiency of clustering is verified.

DOI

10.19781/j.issn.16739140.2020.06.019

First Page

144

Last Page

151

Download

Share

COinS