Analysis and Prediction of Building Electricity Energy Consumption

カテゴリ: 部門大会

論文No: OS9-6

グループ名: 【C】平成22年電気学会電子・情報・システム部門大会講演論文集

発行日: 2010/09/02

タイトル(英語): Analysis and Prediction of Building Electricity Energy Consumption

著者名: Dandan Liu(Tongji University),Hao Zhang(Tongji University),Qijun Chen(Tongji University),森 一之(三菱電機),木田 幸夫(三菱電機)

著者名(英語): Dandan Liu(Tongji University),Hao Zhang(Tongji University),Qijun Chen(Tongji University),kazuyuki mori(Mitsubishi Electric),yukio kida(Mitsubishi Electric)

キーワード: 電力消費|データマイニング|統計解析|線形回帰|決定木|electricity energy consumption|data mining|statistical analysis|linear regression|decision trees

要約(日本語): This paper presented a method of analyzing and predicting building energy consumption. Based on collected electricity energy consumption data of a building the relationship between electricity energy consumption and impact factors was studied, then predicting model was developed using statistical and data mining algorithm. The model can forecast the building hourly electricity consumption and assess the impact of office building management mode on power demands. In addition, this study discussed an approach for automatically detecting abnormal building energy consumption. It can compare the model predicted results and actual results and report the outlier and will improve the current situation that manually detecting abnormal energy consumption by visualization program.

PDFファイルサイズ: 4,473 Kバイト