Deriving knowledge of household behaviour from domestic electricity usage metering

Dent, Ian (2015) Deriving knowledge of household behaviour from domestic electricity usage metering. PhD thesis, University of Nottingham.

PDF (Thesis - as examined) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (3MB) | Preview


The electricity market in the UK is undergoing dramatic changes and requires a transformation of existing practices to meet the current and forthcoming challenges. One aspect of the solution is the deployment of demand side management (DSM) programmes to influence domestic behaviour patterns for the benefit of the overall network. Effective deployment of DSM requires segmentation of the population into a small number of groupings.

Using a database of electricity meter data collected at a frequency of five minutes over a year from several hundred houses, households are clustered based on the shape of the average daily electricity usage profile. A novel method, incorporating evaluation criteria beyond compactness, of evaluating the resulting groupings is defined and tested. The results indicate the potentially most useful algorithms for use with load profile clustering.

Patterns within the electricity meter data are approximated and symbolised to allow motifs (representing repeated behaviours) to be identified. Uninteresting motifs are automatically identified and discarded. The different possible parameters, including size of motif and number of symbols used in representing the data, are explored and the most appropriate values found for use with electricity meter data motif detection.

The concept of variability of regular behaviour within a household is introduced and methods of representing the variability are considered. The novel method of using variability in timing of motifs is compared to other techniques and the results tested using the previously defined evaluation criteria.

Combining the generated motif data with the meter data to produce a single set of archetypes does not produce more useful results for use with DSM. However, creating complementary sets of archetypes based on each set of data, provides a more complete understanding of the households and allows for better targeting of DSM initiatives.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Aickelin, U.
Rodden, T.A.
Subjects: H Social sciences > HB Economic theory
Q Science > QA Mathematics > QA 75 Electronic computers. Computer science
Faculties/Schools: UK Campuses > Faculty of Science > School of Computer Science
Item ID: 27972
Depositing User: Dent, Ian
Date Deposited: 12 Nov 2015 11:07
Last Modified: 15 Oct 2017 18:48

Actions (Archive Staff Only)

Edit View Edit View