Clustering district heat exchange stations using smart meter consumption data

Alexander Martin Tureczek*, Per Sieverts Nielsen, Henrik Madsen, Adam Brun

*Corresponding author for this work

Research output: Contribution to journalJournal articleResearchpeer-review

Abstract

Contrary to electricity smart meter data analysis, little research regarding district heat smart meter data has been published. Previous papers on smart meter data analytics have not investigated autocorrelation in smart meter data. This paper examines district heat smart meter data from the largest district heat supplier in Denmark and autocorrelation is identified in the data. The K-Means algorithm is not able to take autocorrelation into account when clustering. We propose different data transformation methods to enable K-Means to account for this autocorrelation information in the data by using wavelet transformation and autocorrelation features. Our results show that the K-Means yield acceptable clustering results for district heat data when clustering normalized data, inclusion of autocorrelation improves the clustering. The clusters on normalized data are similar to the wavelet transformed clusters, where the autocorrelation has been accounted for. The clustering achieved with the autocorrelation transformation yields finer clusters through accounting for autocorrelation. We are not able to statistically show a difference between the transformations. All transformations result in shadowing clusters, but the autocorrelation transformation generates fewer shadow clusters and reduce the number of dimensions from 744 to 24, resulting in a dramatic reduction in K-Means runtime.
Original languageEnglish
JournalEnergy and Buildings
Volume182
Pages (from-to)144-158
ISSN0378-7788
DOIs
Publication statusPublished - 2019

Keywords

  • Clustering
  • Feature extraction
  • Autocorrelation
  • Wavelet analysis
  • Smart meter data
  • Load pattern

Cite this

@article{d7f033cd28b8467686a16386d72b0770,
title = "Clustering district heat exchange stations using smart meter consumption data",
abstract = "Contrary to electricity smart meter data analysis, little research regarding district heat smart meter data has been published. Previous papers on smart meter data analytics have not investigated autocorrelation in smart meter data. This paper examines district heat smart meter data from the largest district heat supplier in Denmark and autocorrelation is identified in the data. The K-Means algorithm is not able to take autocorrelation into account when clustering. We propose different data transformation methods to enable K-Means to account for this autocorrelation information in the data by using wavelet transformation and autocorrelation features. Our results show that the K-Means yield acceptable clustering results for district heat data when clustering normalized data, inclusion of autocorrelation improves the clustering. The clusters on normalized data are similar to the wavelet transformed clusters, where the autocorrelation has been accounted for. The clustering achieved with the autocorrelation transformation yields finer clusters through accounting for autocorrelation. We are not able to statistically show a difference between the transformations. All transformations result in shadowing clusters, but the autocorrelation transformation generates fewer shadow clusters and reduce the number of dimensions from 744 to 24, resulting in a dramatic reduction in K-Means runtime.",
keywords = "Clustering, Feature extraction, Autocorrelation, Wavelet analysis, Smart meter data, Load pattern",
author = "Tureczek, {Alexander Martin} and Nielsen, {Per Sieverts} and Henrik Madsen and Adam Brun",
year = "2019",
doi = "10.1016/j.enbuild.2018.10.009",
language = "English",
volume = "182",
pages = "144--158",
journal = "Energy and Buildings",
issn = "0378-7788",
publisher = "Elsevier",

}

Clustering district heat exchange stations using smart meter consumption data. / Tureczek, Alexander Martin; Nielsen, Per Sieverts; Madsen, Henrik; Brun, Adam.

In: Energy and Buildings, Vol. 182, 2019, p. 144-158.

Research output: Contribution to journalJournal articleResearchpeer-review

TY - JOUR

T1 - Clustering district heat exchange stations using smart meter consumption data

AU - Tureczek, Alexander Martin

AU - Nielsen, Per Sieverts

AU - Madsen, Henrik

AU - Brun, Adam

PY - 2019

Y1 - 2019

N2 - Contrary to electricity smart meter data analysis, little research regarding district heat smart meter data has been published. Previous papers on smart meter data analytics have not investigated autocorrelation in smart meter data. This paper examines district heat smart meter data from the largest district heat supplier in Denmark and autocorrelation is identified in the data. The K-Means algorithm is not able to take autocorrelation into account when clustering. We propose different data transformation methods to enable K-Means to account for this autocorrelation information in the data by using wavelet transformation and autocorrelation features. Our results show that the K-Means yield acceptable clustering results for district heat data when clustering normalized data, inclusion of autocorrelation improves the clustering. The clusters on normalized data are similar to the wavelet transformed clusters, where the autocorrelation has been accounted for. The clustering achieved with the autocorrelation transformation yields finer clusters through accounting for autocorrelation. We are not able to statistically show a difference between the transformations. All transformations result in shadowing clusters, but the autocorrelation transformation generates fewer shadow clusters and reduce the number of dimensions from 744 to 24, resulting in a dramatic reduction in K-Means runtime.

AB - Contrary to electricity smart meter data analysis, little research regarding district heat smart meter data has been published. Previous papers on smart meter data analytics have not investigated autocorrelation in smart meter data. This paper examines district heat smart meter data from the largest district heat supplier in Denmark and autocorrelation is identified in the data. The K-Means algorithm is not able to take autocorrelation into account when clustering. We propose different data transformation methods to enable K-Means to account for this autocorrelation information in the data by using wavelet transformation and autocorrelation features. Our results show that the K-Means yield acceptable clustering results for district heat data when clustering normalized data, inclusion of autocorrelation improves the clustering. The clusters on normalized data are similar to the wavelet transformed clusters, where the autocorrelation has been accounted for. The clustering achieved with the autocorrelation transformation yields finer clusters through accounting for autocorrelation. We are not able to statistically show a difference between the transformations. All transformations result in shadowing clusters, but the autocorrelation transformation generates fewer shadow clusters and reduce the number of dimensions from 744 to 24, resulting in a dramatic reduction in K-Means runtime.

KW - Clustering

KW - Feature extraction

KW - Autocorrelation

KW - Wavelet analysis

KW - Smart meter data

KW - Load pattern

U2 - 10.1016/j.enbuild.2018.10.009

DO - 10.1016/j.enbuild.2018.10.009

M3 - Journal article

VL - 182

SP - 144

EP - 158

JO - Energy and Buildings

JF - Energy and Buildings

SN - 0378-7788

ER -