Short-term Electricity Demand Forecasting and Key Feature Selection for Teaching, Experiments, Research Building

カテゴリ: 部門大会

論文No: 8

グループ名: 【B】令和5年電気学会電力・エネルギー部門大会

発行日: 2023/08/23

タイトル(英語): Short-term Electricity Demand Forecasting and Key Feature Selection for Teaching, Experiments, Research Building

著者名: 王維浩（東京大学），森勇貴（東京大学），佐藤正寛（東京大学），熊田亜紀子（東京大学）

著者名(英語): Wang Weihao (The Graduate School, the University of Tokyo), Mori Yuki (The Graduate School, the University of Tokyo), Sato Masahiro (The Graduate School, the University of Tokyo), Kumada Akiko (The Graduate School, the University of Tokyo)

キーワード: 短時間電力需要予測|エネルギーマネジメントシステム|LSTM|特徴量選択|Short-term electricity demand forecasting|Energy management system|LSTM|Key feature selection

要約(英語): Establishing a reliable energy management system (EMS) has become a feasible way to achieve carbon neutrality. The key of EMS is accurate electricity demand forecasting. However, electricity demand on campus is difficult to forecast accurately. In this study, we conduct a one-hour electricity demand forecasting for a comprehensive-function building on campus. The forecasting accuracy with different features are compared. Based on the compared models, we propose a key feature selection method. The method can filter out the important features. We also find the method can improve the forecasting accuracy and operational efficiency.

PDFファイルサイズ: 1,851 Kバイト