Predicting winners and losers under time-of-use tariffs using smart meter data. (1st December 2021)

• Record Type: Journal Article
• Authors: Kiguchi, Y.; Weeks, M.; Arakawa, R.
• Is Part Of: Energy. Volume 236(2021)
• Publication Date: 2021-12-01
• Subjects: Time-of-use pricing -- Demand-side management -- Smart meters -- Electricity consumption modelling -- Load shifting -- Residential electricity demand; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Energy consumption -- Periodicals; 333.7905
• DOI: 10.1016/j.energy.2021.121438
• Languages: English
• ISSNs: 0360-5442

Abstract

Abstract: Time-of-use electricity tariffs may become more widespread as smart meters are installed across deregulated domestic electricity markets. Time-of-use tariffs and other methods of time-dependant pricing can be mutually beneficial, realising a cost reduction for both energy companies and customers if the customer responds to the price signalling. However, such tariffs are likely to create positive and negative financial outcomes for individuals because of customer engagement and potential peak shifting capacity. Identifying potential reducers or non-reducers beforehand can optimise a time-of-use programme design, in turn maximising the outcome of the programme. This paper provides a statistical model to identify the characteristics of so-called winners and losers - or households that would be better or worse off under a time-of-use tariff - using only ex ante information. The model's accuracy reaches a reliable level using historical electricity load and basic household characteristics. This accuracy can be further improved if online activity data is available - providing justification for digital interaction and gamification in time-of-use programmes. This paper also publishes a new public dataset of 1423 households in Japan, including historical smart meter data, household characteristics and online activity variables during the time-of-use intervention period in 2017 and 2018. Highlights: A model to identify household outcomes under a Time-of-Use tariff. Reliable model accuracy using historical electricity load and basic characteristics. Created public dataset of energy consumption with online activity variables.