公共建筑能耗监测预警体系研究  

Research on Monitoring and Early Warning System for Energy Consumption in Public Buildings

作　　者：朱小六 刘殿凡 ZHU Xiao-liu;LIU Dian-fan(China Construction Sixth Engineering Bureau Fifth Construction Co.,Ltd.;Tianjin University of Technology)

机构地区：[1]中建六局第五建设有限公司 [2]天津理工大学

出　　处：《智能建筑与智慧城市》2025年第3期5-8,共4页Intelligent Building & Smart City

摘　　要：文章致力于增强公共建筑的能源利用效率及精细化管理能力,通过集成能源管理系统(EMS)与机器学习技术,针对运维阶段中的能耗监控、评价及异常消耗诊断进行了深入探讨。借助全面实地考察与物联网传感器收集的数据,建立详尽的能耗信息库。基于此数据库,运用机器学习算法构建能耗预测模型和异常消耗检测机制,以期为建筑能耗管理提供坚实的科学依据和技术支撑。实验结果表明,本文所提出的能耗预测模型在测试数据集上的拟合优度达到了94.26%,证明了其较高的预测精度。此外,本文基于上述预测模型提出了一种能量使用评估与异常诊断策略,不仅为建立建筑能耗监测预警系统提供了理论支持,也指明了实践路径,有利于促进建筑物能源结构优化及节能潜力的进一步挖掘。This study is devoted to enhancing the energy utilization efficiency and fine management ability of public buildings.Through the integration of energy management system(EMS)and machine learning technology,the energy consumption monitoring,evaluation and abnormal consumption diagnosis in the operation and maintenance stage are discussed in depth.With the help of comprehensive field visits and data collected by IoT sensors,a detailed energy consumption information database is established.Based on this database,machine learning algorithms are used to construct energy consumption prediction models and abnormal consumption detection mechanisms,in order to provide a solid scientific basis and technical support for building energy consumption management.The experimental results show that the goodness of fit of the proposed energy consumption prediction model on the test data set reaches 94.26%,which proves its high prediction accuracy.In addition,based on the above prediction model,this paper proposes an energy use assessment and anomaly diagnosis strategy,which not only provides theoretical support for the establishment of building energy consumption monitoring and early warning system,but also points out the practical path,which is conducive to promoting the optimization of building energy structure and the further exploration of energy-saving potential.

关 键 词：能源管理 能耗预测 机器学习 用能评价 异常诊断 

分 类 号：TU111.195[建筑科学—建筑理论]