Premium
Short‐term building load forecast based on a data‐mining feature selection and LSTM‐RNN method
Author(s) -
Sun Gaiping,
Jiang Chuanwen,
Wang Xu,
Yang Xiu
Publication year - 2020
Publication title -
ieej transactions on electrical and electronic engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 30
eISSN - 1931-4981
pISSN - 1931-4973
DOI - 10.1002/tee.23144
Subject(s) - redundancy (engineering) , computer science , feature selection , recurrent neural network , electrical load , term (time) , electric power system , artificial neural network , power consumption , reliability engineering , grid , smart grid , artificial intelligence , machine learning , data mining , power (physics) , engineering , voltage , physics , quantum mechanics , geometry , electrical engineering , mathematics
Short‐term load forecast for individual electric customers is becoming increasingly important in the grid operation, since the power system is becoming a more interactive and intelligent system. Accurate short‐term load forecast for industrial or commercial electric buildings is more challenging due to the complicated load characteristics and numerous influence variables. In this paper, we consider maximizing the relevancy and minimizing the redundancy criterion to select effectively feature variables, which influence the building load consumption, and then a deep learning technique—long‐short memory recurrent neural network is proposed to predict the load consumption. This novel strategy captures distinct load characteristics, choosing accurate input variables, and shows a great forecasting performance as demonstrated by three different types of city building load in China. © 2020 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.