On a training-less solution for non-intrusive appliance load monitoring using graph signal processing

Research output: Contribution to journalArticle

57 Citations (Scopus)

Abstract

With ongoing large-scale smart energy metering deployments worldwide, disaggregation of a household’s total energy consumption down to individual appliances using analytical tools, aka. non-intrusive appliance load monitoring (NALM), has generated increased research interest lately. NALM can deepen energy feedback, support appliance retrofit advice and support home automation. However, despite the fact that NALM was proposed over 30 years ago, there are still many open challenges with respect to its practicality and effectiveness at low sampling rates. Indeed, the majority of NALM approaches, supervised or unsupervised, require training to build appliance models, and are sensitive to appliance changes in the house, thus requiring regular re-training. In this paper, we tackle this challenge by proposing a NALM approach that does not require any training. The main idea is to build upon the emerging field of graph signal processing to perform adaptive thresholding, signal clustering and pattern matching. We determine the performance limits of our approach and demonstrate its usefulness in practice. Using two open access datasets - the US REDD dataset with active power measurements downsampled to 1min resolution and the UK REFIT dataset with 8sec resolution, we demonstrate the effectiveness of the proposed method for typical smart meter sampling rate, with state-of-the-art supervised and unsupervised NALM approaches as benchmarks.
LanguageEnglish
Pages1784-1799
Number of pages15
JournalIEEE Access
Volume4
DOIs
Publication statusPublished - 21 Apr 2016

Fingerprint

Signal processing
Monitoring
Sampling
Smart meters
Pattern matching
Energy utilization
Automation
Feedback

Keywords

  • non-intrusive appliance load monitoring
  • load disaggregation
  • graph signal processing

Cite this

@article{1fd926a799394ed2aafcc96267448da3,
title = "On a training-less solution for non-intrusive appliance load monitoring using graph signal processing",
abstract = "With ongoing large-scale smart energy metering deployments worldwide, disaggregation of a household’s total energy consumption down to individual appliances using analytical tools, aka. non-intrusive appliance load monitoring (NALM), has generated increased research interest lately. NALM can deepen energy feedback, support appliance retrofit advice and support home automation. However, despite the fact that NALM was proposed over 30 years ago, there are still many open challenges with respect to its practicality and effectiveness at low sampling rates. Indeed, the majority of NALM approaches, supervised or unsupervised, require training to build appliance models, and are sensitive to appliance changes in the house, thus requiring regular re-training. In this paper, we tackle this challenge by proposing a NALM approach that does not require any training. The main idea is to build upon the emerging field of graph signal processing to perform adaptive thresholding, signal clustering and pattern matching. We determine the performance limits of our approach and demonstrate its usefulness in practice. Using two open access datasets - the US REDD dataset with active power measurements downsampled to 1min resolution and the UK REFIT dataset with 8sec resolution, we demonstrate the effectiveness of the proposed method for typical smart meter sampling rate, with state-of-the-art supervised and unsupervised NALM approaches as benchmarks.",
keywords = "non-intrusive appliance load monitoring, load disaggregation, graph signal processing",
author = "Bochao Zhao and Lina Stankovic and Vladimir Stankovic",
year = "2016",
month = "4",
day = "21",
doi = "10.1109/ACCESS.2016.2557460",
language = "English",
volume = "4",
pages = "1784--1799",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "IEEE",

}

TY - JOUR

T1 - On a training-less solution for non-intrusive appliance load monitoring using graph signal processing

AU - Zhao, Bochao

AU - Stankovic, Lina

AU - Stankovic, Vladimir

PY - 2016/4/21

Y1 - 2016/4/21

N2 - With ongoing large-scale smart energy metering deployments worldwide, disaggregation of a household’s total energy consumption down to individual appliances using analytical tools, aka. non-intrusive appliance load monitoring (NALM), has generated increased research interest lately. NALM can deepen energy feedback, support appliance retrofit advice and support home automation. However, despite the fact that NALM was proposed over 30 years ago, there are still many open challenges with respect to its practicality and effectiveness at low sampling rates. Indeed, the majority of NALM approaches, supervised or unsupervised, require training to build appliance models, and are sensitive to appliance changes in the house, thus requiring regular re-training. In this paper, we tackle this challenge by proposing a NALM approach that does not require any training. The main idea is to build upon the emerging field of graph signal processing to perform adaptive thresholding, signal clustering and pattern matching. We determine the performance limits of our approach and demonstrate its usefulness in practice. Using two open access datasets - the US REDD dataset with active power measurements downsampled to 1min resolution and the UK REFIT dataset with 8sec resolution, we demonstrate the effectiveness of the proposed method for typical smart meter sampling rate, with state-of-the-art supervised and unsupervised NALM approaches as benchmarks.

AB - With ongoing large-scale smart energy metering deployments worldwide, disaggregation of a household’s total energy consumption down to individual appliances using analytical tools, aka. non-intrusive appliance load monitoring (NALM), has generated increased research interest lately. NALM can deepen energy feedback, support appliance retrofit advice and support home automation. However, despite the fact that NALM was proposed over 30 years ago, there are still many open challenges with respect to its practicality and effectiveness at low sampling rates. Indeed, the majority of NALM approaches, supervised or unsupervised, require training to build appliance models, and are sensitive to appliance changes in the house, thus requiring regular re-training. In this paper, we tackle this challenge by proposing a NALM approach that does not require any training. The main idea is to build upon the emerging field of graph signal processing to perform adaptive thresholding, signal clustering and pattern matching. We determine the performance limits of our approach and demonstrate its usefulness in practice. Using two open access datasets - the US REDD dataset with active power measurements downsampled to 1min resolution and the UK REFIT dataset with 8sec resolution, we demonstrate the effectiveness of the proposed method for typical smart meter sampling rate, with state-of-the-art supervised and unsupervised NALM approaches as benchmarks.

KW - non-intrusive appliance load monitoring

KW - load disaggregation

KW - graph signal processing

UR - http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=6287639

UR - https://github.com/loneharoon/GSP_energy_disaggregator

U2 - 10.1109/ACCESS.2016.2557460

DO - 10.1109/ACCESS.2016.2557460

M3 - Article

VL - 4

SP - 1784

EP - 1799

JO - IEEE Access

T2 - IEEE Access

JF - IEEE Access

SN - 2169-3536

ER -