Skip to main content
placeholder image

A two-level clustering strategy for energy performance evaluation of university buildings

Conference Paper


Abstract


  • This paper presents a clustering strategy to evaluate the

    energy performance and identify typical daily load profiles of

    buildings. The cluster analysis included intra-building

    clustering and inter-building clustering. The intra-building

    clustering used Gaussian mixture model clustering to identify

    the typical daily load profiles of each individual building. The

    inter-building clustering used hierarchical clustering to further

    identify the typical daily load profiles of a stock of buildings

    based on the typical daily load profiles identified for each

    individual building. The performance of this strategy was

    tested and evaluated using the two-year hourly electricity

    consumption data collected from 40 buildings on a university

    campus in Australia. The result showed that this strategy

    could discover the information related to building energy

    usage. The results obtained from this study could be

    potentially used to assist in decision making for energy

    performance enhancement initiatives of university buildings.

Publication Date


  • 2018

Citation


  • Li, K., Ma, Z., Robinson, D. & Ma, J. (2018). A two-level clustering strategy for energy performance evaluation of university buildings. In K. Inthavong, C. P. Cheung, G. Yeoh & J. Tu (Eds.), Proceedings of the 4th International Conference on Building Energy & Environment (pp. 168-173). Melbourne, Australia: Conference On Building Energy & Environment.

Start Page


  • 168

End Page


  • 173

Place Of Publication


  • Melbourne, Australia

Abstract


  • This paper presents a clustering strategy to evaluate the

    energy performance and identify typical daily load profiles of

    buildings. The cluster analysis included intra-building

    clustering and inter-building clustering. The intra-building

    clustering used Gaussian mixture model clustering to identify

    the typical daily load profiles of each individual building. The

    inter-building clustering used hierarchical clustering to further

    identify the typical daily load profiles of a stock of buildings

    based on the typical daily load profiles identified for each

    individual building. The performance of this strategy was

    tested and evaluated using the two-year hourly electricity

    consumption data collected from 40 buildings on a university

    campus in Australia. The result showed that this strategy

    could discover the information related to building energy

    usage. The results obtained from this study could be

    potentially used to assist in decision making for energy

    performance enhancement initiatives of university buildings.

Publication Date


  • 2018

Citation


  • Li, K., Ma, Z., Robinson, D. & Ma, J. (2018). A two-level clustering strategy for energy performance evaluation of university buildings. In K. Inthavong, C. P. Cheung, G. Yeoh & J. Tu (Eds.), Proceedings of the 4th International Conference on Building Energy & Environment (pp. 168-173). Melbourne, Australia: Conference On Building Energy & Environment.

Start Page


  • 168

End Page


  • 173

Place Of Publication


  • Melbourne, Australia