Understanding electricity consumption patterns in households through data fusion of smart meters and door-to-door surveys

Panel: 5. Energy use in buildings: projects, technologies and innovation

This is a peer-reviewed paper.

Authors:
João Pedro Gouveia, FCT-UNL - Faculdade de Ciências e Tecnologia - Universidade Nova de Lisboa, Portugal
Júlia Seixas, CENSE, Center for Environmental and Sustainability Research, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, Portugal
Shimming Luo, CENSE, Center for Environmental and Sustainability Research, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, Portugal
Nuno Bilo, Évora Municipality, Portugal
António Valentim, Évora Municipality, Portugal

Abstract

This paper contributes to a comprehensive understanding on how electricity consumption patterns are explained in a Southwest European city. Household electricity consumption drivers and profiles of different type of consumers are analysed. They are built upon data gathered from a sample of 230 households with daily electricity consumption from smart meters of InovCity project, the first of its kind in Portugal consisting of a massive smart metering system (31 000 smart meters) with door-to-door 110–question surveys for the same houses in Évora. The survey encompassed information on socio economic data, equipment’s ownership and use and physical characteristics of the dwellings. Ten clusters were derived using Ward’s method hierarchical clustering to identify similar types of consumers based on their means and standard deviations, and three of them are deeply analysed and compared. Based on the surveys, a socioeconomic characterization of each cluster was made in order to capture the main similarities and differences within each cluster and compared to the others. We conclude that three major groups of determinants influence residential electricity consumption segmentation: physical characteristics of a dwelling especially year of construction and total floor area; electrical heating/cooling equipment and fireplaces ownership and use; and occupants profiles (mainly number of occupants and monthly income). Urbanisation levels, bearing structure, type of tariff and contracted power are not variables that distinct the clusters grouping. This consumer profiling allows deriving insights to support utilities for marketing segmentation and policies for effective energy reduction (e.g. tariff design, demand side management strategies, peak shaving). This work is being developed under the EU project InSMART, that involves four European cities targeting innovative methods to integrative city planning, including buildings, transport, and utilities networks.