Authors:
Bo Yao
1
;
Yiming Xu
1
;
Yue Pang
1
;
Chaoyi Jin
1
;
Zijing Tan
1
;
Xiangdong Zhou
1
and
Yun Su
2
Affiliations:
1
Fudan University, China
;
2
State Grid Shanghai Municipal Electric Power Company, China
Keyword(s):
Electricity Time Series, Sparse Principal Components Analysis, Clustering and Categorization.
Abstract:
The well-being of people, industry and economy depends on reliable, sustainable and affordable energy. The analysis on energy consumption model, especially on electricity consumption model, plays an important role in providing guidance that makes energy system stable and economical. In this paper, clustering based on electricity consumption model is imposed to categorize consumers, and Sparse Principal Components Analysis (SPCA) is employed to analyse electricity consumption model for each group clustered. Experimental results show that our methods can automatically divide a day into peak times and off-peak times, so as to reveal in detail the electricity consumption model of different types of consumers. Additionally, we study the relationships between social background of consumers and their electricity consumption model. Our experimental results show that social background of consumers has impact on their consumption model, as expected, but cannot fully determine it.