8th International Conference on Advanced Power System Automation and Protection (APAP2019)

1055 / 2019-05-19 21:01:23

Analysis of Residents' Differential Electricity Use Behavior Based on Load Decomposition

Differential analysis; Load decomposition; Load characteristics; Improved k-means algorithm; Cluster analysis.

In the process of smart grid construction, the impact of residential electricity consumption on the distribution network is increasing. In order to accurately grasp residential electricity consumption characteristics, the residential load is decomposed into daily basic load and holiday load. Aiming at the problem that traditional K-means algorithm is sensitive to clustering centers, this paper proposes an initial center selection method based on high-density data sets and data heterogeneity, which is based on an improved algorithm，the daily basic load data and holiday load of residents in a certain community are cluster analysis. The load characteristics of the classification results are analyzed. Based on this, a new user classification method is proposed to realize the differential analysis of the residential electricity behavior characteristics. The experimental results show that the proposed classification method can not only accurately describe the behavior of residential electricity, but also provide more effective data support for demand response.