TU Darmstadt / ULB / TUbiblio

Modelling Energy Consumption Profiles with Data Mining Methods based on Smart Meter Survey Data with Additional User Characteristics

Rüppel, Uwe ; Schwöbel, Christian
Li, Haijiang ; de Wilde, Pieter ; Rafiq, Yaqub (eds.) :

Modelling Energy Consumption Profiles with Data Mining Methods based on Smart Meter Survey Data with Additional User Characteristics.
[Online-Edition: http://egice2014.engineering.cf.ac.uk/images/BIN/papers/18.p...]
In: 21st EG-ICE International Workshop, 16.-18. Juli 2014, Cardiff, England. Proceedings of the 21st International Workshop on Intelligent Computing in Engineering
[ Konferenzveröffentlichung] , (2014)

Offizielle URL: http://egice2014.engineering.cf.ac.uk/images/BIN/papers/18.p...

Kurzbeschreibung (Abstract)

Among the many measures in the field of energy efficiency, predicting energy consumption and a better understanding of how and when energy is consumed are key factors. Possible ways to predict energy consumption would be to apply simulation models or to make use of statistical methods. However, most of these methods do not include user and building properties in their models. In this paper, a data mining method is shown which uses the questionnaire of a smart metering study and its metering values as an input. First, clustering is applied on a subset of the survey questions to form user groups and then a link to each cluster’s appropriate consumption characteristics is established. In this work two clustering processes are shown as examples. In the first example, clusters are formed using characteristics about the residents. Then the clusters are being linked to their electricity consumption. In the second example, building characteristics are used to form the clusters and these are being connected to their gas consumption.

Typ des Eintrags: Konferenzveröffentlichung
Erschienen: 2014
Herausgeber: Li, Haijiang ; de Wilde, Pieter ; Rafiq, Yaqub
Autor(en): Rüppel, Uwe ; Schwöbel, Christian
Titel: Modelling Energy Consumption Profiles with Data Mining Methods based on Smart Meter Survey Data with Additional User Characteristics
Sprache: Englisch
Kurzbeschreibung (Abstract):

Among the many measures in the field of energy efficiency, predicting energy consumption and a better understanding of how and when energy is consumed are key factors. Possible ways to predict energy consumption would be to apply simulation models or to make use of statistical methods. However, most of these methods do not include user and building properties in their models. In this paper, a data mining method is shown which uses the questionnaire of a smart metering study and its metering values as an input. First, clustering is applied on a subset of the survey questions to form user groups and then a link to each cluster’s appropriate consumption characteristics is established. In this work two clustering processes are shown as examples. In the first example, clusters are formed using characteristics about the residents. Then the clusters are being linked to their electricity consumption. In the second example, building characteristics are used to form the clusters and these are being connected to their gas consumption.

Buchtitel: Proceedings of the 21st International Workshop on Intelligent Computing in Engineering
Fachbereich(e)/-gebiet(e): 13 Fachbereich Bau- und Umweltingenieurwissenschaften
13 Fachbereich Bau- und Umweltingenieurwissenschaften > Institut für Numerische Methoden und Informatik im Bauwesen
Veranstaltungstitel: 21st EG-ICE International Workshop
Veranstaltungsort: Cardiff, England
Veranstaltungsdatum: 16.-18. Juli 2014
Hinterlegungsdatum: 22 Jan 2015 10:26
Offizielle URL: http://egice2014.engineering.cf.ac.uk/images/BIN/papers/18.p...
Zusätzliche Informationen:

ISBN: 978-0-9930807-0-8

Export:

Optionen (nur für Redakteure)

Eintrag anzeigen Eintrag anzeigen