Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Identifying representative types of residential electricity consumers - a cluster analysis of hourly smart meter data

 
: Klingler, Anna-Lena; Schuhmacher, Florian; Wohlfarth, Katharina

Lopes, Marta A.R. (Ed.):
4th European Conference on Behaviour and Energy Efficiency, Behave 2016 : INESC Coimbra - Institute for Systems Engineering and Computers at Coimbra & ADENE - Portuguese Agency for Energy, Coimbra, Portugal, September 8-9, 2016
Coimbra(Portugal): University of Coimbra, 2016
ISBN: 978-989-95055-9-9
11 S.
European Conference on Behaviour and Energy Efficiency (BEHAVE) <4, 2016, Coimbra>
Englisch
Konferenzbeitrag
Fraunhofer ISI ()
cluster analysis; smart meter data; standard load profile; electricity consumption

Abstract
Electricity consumption is characterized not only by its total amount, but also its temporal course, the so-called load or consumption profile. To address future-oriented questions, like in the context of the rising trend towards the adoption of rooftop Photovoltaic panels for onsite generation and self-consumption in the residential sector, it is important to better understand the differences between individual residential load profiles and the underlying behaviour. In this study, we propose a set of residential load profiles which better reflect the sector’s heterogeneity than the current standard load profile. The load profiles are created through a cluster analysis of a comprehensive set of hourly measured residential load data by means of their hourly electricity consumption behaviour only, using the household characteristics as explanatory variables. In our analysis, four in itself homogenous cluster groups could be identified. The resulting profiles feature distinctive characteristics concerning the level and time of peak load for different weekdays and seasons. As the clustering is solely based on consumption data, conclusions regarding consumer types are possible. Using binary logit regression and cross table analysis, we found that the household size, average age of the household members and the presence of children have significant influence on the electricity consumption behaviour. Whether the household members have employment or not, was especially relevant, as it determines the presence of people at home during the day. Concerning household appliances, only the absolute number of end-uses is significant, provided that the appliances are not time controlled.Concerning the dominant attributes, the resulting four cluster groups can be described as the single household with an irregular daily schedule, the households with old and retired couples, the high educated, working performers and the households representing young classic families.

: http://publica.fraunhofer.de/dokumente/N-426336.html