Evolving smart meter data driven model for short-term forecasting of electric loads

Harri Niska, Pekka Koponen, Antti Mutanen

    Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

    5 Citations (Scopus)

    Abstract

    Short-term forecasting of electric loads is an essential function required by Smart Grids. Today increasing amount of smart metering data is available enabling the development of enhanced data-driven models for short-term load forecasting. Until now, a plethora of models have been developed ranging from simple linear regression models to more advanced models such as (artificial) neural networks (NNs) and support vector machines (SVMs). Despite the relatively high accuracy obtained, the acceptance of purely data-driven models such as NN models is still remained limited due to their complexity and nontransparent nature. Therefore it is important to develop optimization schemes, which can be used to facilitate the selection of appropriate model structure resulting good forecasting accuracy with low complexity. This study presents an optimization scheme based on multi-objective genetic algorithm (GA) for designing data-driven models for short-term forecasting of electric loads. The optimization scheme is demonstrated for designing the conventional NN/MLP model using real smart metering data and weather measurements. The optimal NN model structures are identified and analyzed in terms of model complexity and forecasting accuracy.
    Original languageEnglish
    Title of host publicationIntelligent Sensors, Sensor Networks and Information Processing (ISSNIP), 2015 IEEE Tenth International Conference on
    PublisherIEEE Institute of Electrical and Electronic Engineers
    Pages1-6
    ISBN (Electronic)978-1-4799-8054-3, 978-1-4799-8055-0
    DOIs
    Publication statusPublished - 2015
    MoE publication typeA4 Article in a conference publication
    EventIEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing - Singapore, Singapore
    Duration: 7 Apr 20159 Apr 2015
    Conference number: 10

    Conference

    ConferenceIEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing
    Abbreviated titleISSNIP
    CountrySingapore
    CitySingapore
    Period7/04/159/04/15

    Fingerprint

    Smart meters
    Electric loads
    Neural networks
    Model structures
    Linear regression
    Support vector machines
    Genetic algorithms

    Keywords

    • smart metering
    • data mining
    • load forecasting
    • genetic algorithms

    Cite this

    Niska, H., Koponen, P., & Mutanen, A. (2015). Evolving smart meter data driven model for short-term forecasting of electric loads. In Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP), 2015 IEEE Tenth International Conference on (pp. 1-6). IEEE Institute of Electrical and Electronic Engineers . https://doi.org/10.1109/ISSNIP.2015.7106966
    Niska, Harri ; Koponen, Pekka ; Mutanen, Antti. / Evolving smart meter data driven model for short-term forecasting of electric loads. Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP), 2015 IEEE Tenth International Conference on. IEEE Institute of Electrical and Electronic Engineers , 2015. pp. 1-6
    @inproceedings{fbd7260f90974195b0422959e4c1041b,
    title = "Evolving smart meter data driven model for short-term forecasting of electric loads",
    abstract = "Short-term forecasting of electric loads is an essential function required by Smart Grids. Today increasing amount of smart metering data is available enabling the development of enhanced data-driven models for short-term load forecasting. Until now, a plethora of models have been developed ranging from simple linear regression models to more advanced models such as (artificial) neural networks (NNs) and support vector machines (SVMs). Despite the relatively high accuracy obtained, the acceptance of purely data-driven models such as NN models is still remained limited due to their complexity and nontransparent nature. Therefore it is important to develop optimization schemes, which can be used to facilitate the selection of appropriate model structure resulting good forecasting accuracy with low complexity. This study presents an optimization scheme based on multi-objective genetic algorithm (GA) for designing data-driven models for short-term forecasting of electric loads. The optimization scheme is demonstrated for designing the conventional NN/MLP model using real smart metering data and weather measurements. The optimal NN model structures are identified and analyzed in terms of model complexity and forecasting accuracy.",
    keywords = "smart metering, data mining, load forecasting, genetic algorithms",
    author = "Harri Niska and Pekka Koponen and Antti Mutanen",
    note = "SDA: MIP: Intelligent Energy Grids Project : 101968",
    year = "2015",
    doi = "10.1109/ISSNIP.2015.7106966",
    language = "English",
    pages = "1--6",
    booktitle = "Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP), 2015 IEEE Tenth International Conference on",
    publisher = "IEEE Institute of Electrical and Electronic Engineers",
    address = "United States",

    }

    Niska, H, Koponen, P & Mutanen, A 2015, Evolving smart meter data driven model for short-term forecasting of electric loads. in Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP), 2015 IEEE Tenth International Conference on. IEEE Institute of Electrical and Electronic Engineers , pp. 1-6, IEEE Tenth International Conference on Intelligent Sensors, Sensor Networks and Information Processing, Singapore, Singapore, 7/04/15. https://doi.org/10.1109/ISSNIP.2015.7106966

    Evolving smart meter data driven model for short-term forecasting of electric loads. / Niska, Harri; Koponen, Pekka; Mutanen, Antti.

    Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP), 2015 IEEE Tenth International Conference on. IEEE Institute of Electrical and Electronic Engineers , 2015. p. 1-6.

    Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

    TY - GEN

    T1 - Evolving smart meter data driven model for short-term forecasting of electric loads

    AU - Niska, Harri

    AU - Koponen, Pekka

    AU - Mutanen, Antti

    N1 - SDA: MIP: Intelligent Energy Grids Project : 101968

    PY - 2015

    Y1 - 2015

    N2 - Short-term forecasting of electric loads is an essential function required by Smart Grids. Today increasing amount of smart metering data is available enabling the development of enhanced data-driven models for short-term load forecasting. Until now, a plethora of models have been developed ranging from simple linear regression models to more advanced models such as (artificial) neural networks (NNs) and support vector machines (SVMs). Despite the relatively high accuracy obtained, the acceptance of purely data-driven models such as NN models is still remained limited due to their complexity and nontransparent nature. Therefore it is important to develop optimization schemes, which can be used to facilitate the selection of appropriate model structure resulting good forecasting accuracy with low complexity. This study presents an optimization scheme based on multi-objective genetic algorithm (GA) for designing data-driven models for short-term forecasting of electric loads. The optimization scheme is demonstrated for designing the conventional NN/MLP model using real smart metering data and weather measurements. The optimal NN model structures are identified and analyzed in terms of model complexity and forecasting accuracy.

    AB - Short-term forecasting of electric loads is an essential function required by Smart Grids. Today increasing amount of smart metering data is available enabling the development of enhanced data-driven models for short-term load forecasting. Until now, a plethora of models have been developed ranging from simple linear regression models to more advanced models such as (artificial) neural networks (NNs) and support vector machines (SVMs). Despite the relatively high accuracy obtained, the acceptance of purely data-driven models such as NN models is still remained limited due to their complexity and nontransparent nature. Therefore it is important to develop optimization schemes, which can be used to facilitate the selection of appropriate model structure resulting good forecasting accuracy with low complexity. This study presents an optimization scheme based on multi-objective genetic algorithm (GA) for designing data-driven models for short-term forecasting of electric loads. The optimization scheme is demonstrated for designing the conventional NN/MLP model using real smart metering data and weather measurements. The optimal NN model structures are identified and analyzed in terms of model complexity and forecasting accuracy.

    KW - smart metering

    KW - data mining

    KW - load forecasting

    KW - genetic algorithms

    U2 - 10.1109/ISSNIP.2015.7106966

    DO - 10.1109/ISSNIP.2015.7106966

    M3 - Conference article in proceedings

    SP - 1

    EP - 6

    BT - Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP), 2015 IEEE Tenth International Conference on

    PB - IEEE Institute of Electrical and Electronic Engineers

    ER -

    Niska H, Koponen P, Mutanen A. Evolving smart meter data driven model for short-term forecasting of electric loads. In Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP), 2015 IEEE Tenth International Conference on. IEEE Institute of Electrical and Electronic Engineers . 2015. p. 1-6 https://doi.org/10.1109/ISSNIP.2015.7106966