Title: Energy consumption prediction method of energy saving building based on deep reinforcement learning

Authors: Chuan He; Ying Xiong; Yeda Lin; Lie Yu; Hui-Hua Xiong

Addresses: School of Architecture, Changsha University of Science & Technology, Changsha 410114, Hunan, China ' School of Architecture, Changsha University of Science & Technology, Changsha 410114, Hunan, China ' School of Architecture, Changsha University of Science & Technology, Changsha 410114, Hunan, China ' School of Hydraulic Engineering, Changsha University of Science & Technology, Changsha 410114, Hunan, China ' School of Architecture and Urban planning, Hunan City University, Yiyang 413000, Hunan, China

Abstract: In order to overcome the problems of low-prediction accuracy and long prediction time of traditional building energy consumption prediction methods, this paper proposes a new energy-saving building energy consumption prediction method based on deep reinforcement learning. Through the deep reinforcement learning algorithm, a number of energy consumption behaviour return information of specific value network and strategy network are calculated, respectively to build the energy consumption probability model of energy-saving building energy consumption equipment. The linear rectification function with leakage is used to update the probability model and parameters, and the linear relationship prediction function of energy consumption parameters is constructed by using the learning process and results to complete the dynamic prediction of energy consumption of energy-saving buildings. The experimental results show that the proposed method has fast prediction speed and high accuracy, which can provide reference for the implementation of energy-saving building.

Keywords: deep intensive learning; energy saving building; energy consumption forecasting; time series; multivariate linear regression.

DOI: 10.1504/IJGEI.2022.125406

International Journal of Global Energy Issues, 2022 Vol.44 No.5/6, pp.524 - 536

Received: 14 Jan 2021
Accepted: 10 Sep 2021
Published online: 08 Sep 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article