Energy-efficient virtual sensor-based deep reinforcement learning control of indoor CO_(2)in a kindergarten

作     者：Patrick Nzivugira Duhirwe Jack Ngarambe Geun Young Yun 

作者机构：Department of Architectural EngineeringKyung Hee UniversityYongin 17104Republic of Korea 

出 版 物：《Frontiers of Architectural Research》 (建筑学研究前沿（英文版）)

年 卷 期：2023年第12卷第2期

页      面：394-409页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported by the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT)(No.2020R1A2C1099611) 

主　　题：Indoor air quality Indoor CO_(2)control Machine learning Virtual sensor Deep reinforcement learning 

摘      要：High concentrations of indoor CO_(2)pose severe health risks to building ***,mechanical equipment is used to provide sufficient ventilation as a remedy to high indoor CO_(2)***,such equipment consumes large amounts of energy,substantially increasing building energy *** the end,the issue becomes an optimization problem that revolves around maintaining CO_(2)levels below a certain threshold while utilizing the minimum amount of energy *** that end,we propose an intelligent approach that consists of a supervised learning-based virtual sensor that interacts with a deep reinforcement learning(DRL)-based control to efficiently control indoor CO_(2)while utilizing the minimum amount of energy *** data used to train and test the DRL agent is based on a 3-month field experiment conducted at a kindergarten equipped with a heat recovery *** results show that,unlike the manual control initially employed at the kindergarten,the DRL agent could always maintain the CO_(2)concentrations below sufficient ***,a 58%reduction in the energy consumption of the ventilator under the DRL control compared to the manual control was *** demonstrated approach illustrates the potential leveraging of Internet of Things and machine learning algorithms to create comfortable and healthy indoor environments with minimal energy requirements.