基于卷积神经网络的自适应暖通空调系统节能控制方法  

Energy Saving Control Method for Adaptive HVAC Systems Based on Convolutional Neural Networks

作　　者：王天一 丁超 孙琳 郑成林 WANG Tianyi;DING Chao;SUN Lin;ZHENG Chenglin(Nanjing Stomatological Hospital,Affiliated Hospital of Medical School,Institute of stomatology Nanjing University,Nanjing 210008,China)

机构地区：[1]南京大学医学院附属口腔医院,南京市口腔医院,南京大学口腔医学研究所,江苏南京210008

出　　处：《微型电脑应用》2024年第10期38-41,共4页Microcomputer Applications

基　　金：南京市卫生科技发展专项资金资助项目(GAX22284)。

摘　　要：目前,在世界范围内的社会总耗能的一小半都是来自建筑行业,并且该行业中超过一半的能耗都是暖通空调(HVAC)系统造成的,怎样减少HVAC系统的能耗是一个亟待解决的问题。针对此问题,提出一种基于卷积神经网络的自适应HVAC系统节能控制方法,该方法通过卷积神经网络对建筑中接收到的数据进行分析,从而调整暖通系统中各个部分的工作状态,降低整个系统的能耗。实验结果表明,在迭代次数达到90时,反向传播算法、随机森林算法和卷积神经网络算法模型的平均绝对误差值分别为0.689、0.668和0.661,均方根误差值为0.884、0.882和0.879,所提出的算法模型能够有效地降低暖通系统的能量消耗,同时也能保证建筑内人员的舒适性,为指导建筑节能改造提供一定的参考。At present,less than half of the total energy consumption in society worldwide comes from the construction industry,and more than half of the energy consumption in this industry is caused by heating,ventilation and air conditioning(HVAC)systems.How to reduce the energy consumption of HVAC systems is an urgent problem to be solved.In response to this is-sue,this study proposes an energy saving control method for adaptive HVAC systems based on convolutional neural networks.This method analyzes the data received in buildings through convolutional neural networks to adjust the working status of vari-ous parts of the HVAC system and reduce the energy consumption of the entire system.The experimental results show that when the number of iterations reaches 90,the average absolute error values of the back propagation algorithm,random forest algorithm,and convolutional neural network algorithm models are 0.689,0.668,and 0.661 respectively,and the root mean square error values are 0.884,0.882,and 0.879.The research results indicate that the proposed algorithm model can effec-tively reduce the energy consumption of HVAC systems,and also ensures the comfort of personnel inside buildings,which pro-vides certain reference for guiding energy saving renovation of buildings.

关 键 词：暖通空调系统 卷积神经网络 节能控制 大型建筑 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]