11th International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services

Research Article

Can Smart Plugs Predict Electric Power Consumption? A Case Study

Download751 downloads
  • @INPROCEEDINGS{10.4108/icst.mobiquitous.2014.257991,
        author={Andreas Reinhardt and Delphine Christin and Salil Kanhere},
        title={Can Smart Plugs Predict Electric Power Consumption? A Case Study},
        proceedings={11th International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services},
        publisher={ICST},
        proceedings_a={MOBIQUITOUS},
        year={2014},
        month={11},
        keywords={distributed appliance metering power consumption prediction smart buildings smart grids},
        doi={10.4108/icst.mobiquitous.2014.257991}
    }
    
  • Andreas Reinhardt
    Delphine Christin
    Salil Kanhere
    Year: 2014
    Can Smart Plugs Predict Electric Power Consumption? A Case Study
    MOBIQUITOUS
    ICST
    DOI: 10.4108/icst.mobiquitous.2014.257991
Andreas Reinhardt1,*, Delphine Christin2, Salil Kanhere3
  • 1: Technische Universität Clausthal
  • 2: University of Bonn and Fraunhofer FKIE
  • 3: The University of New South Wales
*Contact email: andreasr@cse.unsw.edu.au

Abstract

The Internet of Things will encompass a rich variety of sensing systems including mobile phones, embedded sensor and actuator platforms, and even smart electricity meters. Through their collaborative operation, billions of such devices will realize the vision of smart homes, smart cities, and beyond. Smart electricity meters can, e.g., already help utilities monitor the stability of the power grid by periodically reporting each connected household’s energy consumption. Even more sophisticated services can be realized when data is available at higher spatio-temporal resolutions, e.g., when distributed smart power meters (sometimes referred to as smart plugs) are deployed. In this paper, we present one such novel service for the smart home, namely making projections of an appliance’s future energy demand based on previously observed power consumption data. In a first step, our system identifies and isolates unique characteristic signatures from collected power consumption traces. Subsequently, time series pattern matching is applied to detect these signatures in real-time data. Based on the occurrences of the extracted signatures in real-time data, the appliance’s future power demand is predicted. We evaluate our approach with more than 2,500 appliance activity segments collected from 15 different appliance types, and show that accurate forecasts can be made in many cases.