Revealing Generalized Load Characteristics with Variable Scale User Portrait

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

2 Scopus citations

Abstract

With the development of distributed energy technologies, the proportion of generalized load in electric load has greatly increased. Traditional research methods lack variation among different scales and efficiency in big data processing. This paper proposes variable scale user portrait to visually reveal temporal distribution characteristics of generalized load. We adopt it to the data of 3500 customers of Australian Smart Grid, Smart City(SGSC) program. We first set up the label system which is key to user portrait. With Euclidean distance based distributed clustering(EDBDC), the typical users are obtained quickly and efficiently. Then we construct the user portrait with variable time scale load curves for each customer category. The result pattern can provide exploitable information to utilities and retailers for efficient design and implementation of energy service targeted to various customer categories.

Original languageEnglish
Title of host publication2018 International Conference on Power System Technology, POWERCON 2018 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages4062-4070
Number of pages9
ISBN (Electronic)9781538664612
DOIs
StatePublished - 2 Jul 2018
Event2018 International Conference on Power System Technology, POWERCON 2018 - Guangzhou, China
Duration: 6 Nov 20189 Nov 2018

Publication series

Name2018 International Conference on Power System Technology, POWERCON 2018 - Proceedings

Conference

Conference2018 International Conference on Power System Technology, POWERCON 2018
Country/TerritoryChina
CityGuangzhou
Period6/11/189/11/18

Keywords

  • big data
  • EDBDC
  • Generalized load
  • label system
  • user portrait

Fingerprint

Dive into the research topics of 'Revealing Generalized Load Characteristics with Variable Scale User Portrait'. Together they form a unique fingerprint.

Cite this