Energy disaggregation based on smart metering data via semi-binary nonnegative matrix factorization

Ayumu Miyasawa*, Yu Fujimoto, Yasuhiro Hayashi

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

31 Citations (Scopus)


The information measured by smart electricity meters in households enables consumers to understand the relationship between their behavior and the power consumption. This paper proposes an energy disaggregation method to estimate the individual-appliance power consumption utilizing the total power consumption data collected by smart meters without requiring additional sensors for supporting sustainable demand-side energy management. In addition, in order to enable appliance labeling and highly accurate estimation without submetered data of the target households, the utilization of auxiliary information such as consumer feedback is suggested. This scheme provides users the opportunity to explicitly point out incorrect current disaggregation results so as to improve forthcoming future decomposition accuracy. The accuracy of the proposed energy disaggregation framework is evaluated using real-world power consumption data sets.

Original languageEnglish
Pages (from-to)547-558
Number of pages12
JournalEnergy and Buildings
Publication statusPublished - 2019 Jan 15


  • Energy disaggregation
  • Non-intrusive appliance load monitoring
  • Non-negative matrix factorization
  • Smart electricity meters

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Building and Construction
  • Mechanical Engineering
  • Electrical and Electronic Engineering


Dive into the research topics of 'Energy disaggregation based on smart metering data via semi-binary nonnegative matrix factorization'. Together they form a unique fingerprint.

Cite this