公共建筑能耗监测平台数据缺陷辨识方法和在线诊断研究  

Data Defect Identification Methodology and Fault Diagnosis of Public Building Energy Consumption Monitoring Hatform

作　　者：李怀[1] 于震[1] 吴剑林[1] LI Huai;YU Zhen;WU Jian-lin(Institute of Building Environment and Energy, China Academy of Building Research, Beijing 100013, China)

机构地区：[1]中国建筑科学研究院建筑节能与环境研究院,北京100013

出　　处：《建筑节能》2017年第1期119-124,128,共7页BUILDING ENERGY EFFICIENCY

基　　金："十二五"国家科技支撑计划项目:公共机构能源管理现代信息技术研究与应用示范(2013BAJ15B02-3)

摘　　要：数据问题多、发现和处理不及时是目前公共机构能耗监测平台普遍存在的问题,在对公共机构能耗监测平台常见的数据缺陷和问题进行研究的基础上,按照数据缺陷的不同程度,对数据缺陷进行了模糊归类,并对不同数据问题进行数学模型抽象,并结合不同数据属性,提出了一套简单数据纠错数学方法。以某实际办公建筑能耗监测平台为基础,在数据库层面,应用该数据纠错和辨识算法对能耗监测平台出现的数据问题及时响应并在线缺陷辨识。该方法能够快速对数据问题进行响应和辨识,有效避免能耗监测平台因数据问题带来的不同程度的后果。It is very common that public building has large volume,more energy consumption equip-ments,and some devices sharing the same circuit and equipment,complex circuit construction,so it is inevitable that data loss and error occur in energy monitoring platform of these buildings.Currently,public building energy monitoring platform industry is not optimistic.Due to low quality of monitoring data,many platforms face a short life,and some even paralysis.These serious consequences are partly caused by response not in time when data error occurs,which leads more serious data problem.Many domestic and foreign researchers engaged in building energy efficiency,mostly focusing on existing problems of platform by data mining techniques,and clustering analysis method to address and fix data error or missing data.The energy consumption data error response is focused in a timely manner.A simple mathematic method is proposed for online error diagnostics,and data problems'resolving.The article firstly distinguishes the fuzzy category common data problems appearing in energy consumption data platform,and abstracts simple mathematic models based on data problems,analyzes and identifies data errors on line on an opera-ted energy consumption platform.There are various data errors and defects in energy monitoring platform,according to source defects.They can be classified into the following types:hardware defect,communication defect and software defect,according to the way of handling;it can be grouped into data needs rectification and rectification,accord-ing to the defect of data;it can be classified into no data,data error,data serious error,data missing.To handle the problem more abstractly,data defects can be described as“Normal”,Defect“normal”Normal“defects”and“Defects”.“Normal”:it includes the following phenomenon and causes:1)the real operation data is collected,but the data is not the real data of what should be monitored due to incorrect equipment cabling;2)the equip-ment works in special working condition

关 键 词：能耗监测平台 数据缺陷 在线诊断 数学模型 

分 类 号：TU20[建筑科学—建筑设计及理论]