基于不同数据采集方法的人员行为表征对学生宿舍能耗模拟预测的影响  

the Influence of Occupant Behavior Characterization Based on Different Data Acquisition Methods on Simulation and Prediction of Energy Consumption in Dormitories

作　　者：车铭江 徐象国[1,2] 李平[2] 丁德[2] CHE Mingjiang;XU Xiangguo;LI Ping;DING De(Institute of Refrigeration and Cryogenics,Key Laboratory of Refrigeration and Cryogenic Technology of Zhejiang Province,Zhejiang University,Hangzhou 310027,China;Center for Balance Architecture,Zhejiang University,Hangzhou 310027,China)

机构地区：[1]浙江大学制冷与低温研究所浙江省制冷与低温技术重点实验室,杭州310027 [2]浙江大学平衡建筑研究中心,杭州310027

出　　处：《建筑科学》2024年第8期24-33,共10页Building Science

基　　金：浙江大学平衡建筑研究中心自主立项科研项目(K横20203317C);中央高校基本科研业务费专项资金(2022FZZX01-09)。

摘　　要：人员行为模型在建筑能耗模拟中发挥着重要的作用。为获取建筑内的人员行为,可以采用不同的数据采集方法。不同方法所获得的人员行为模型可能存在差异,特别是在对人员行为随机性的表征方面,不同方法会获取不同的随机行为时间颗粒度。为比较不同数据采集方式下,人员行为表征的差异性以及该差异对最终能耗模拟预测准确性的影响,本文选取学生宿舍,针对其住户展开相关问卷调研与全球定位系统(Global Positioning System,GPS)数据采集并对其实际用电数据进行测量。结果表明,传感器采集的方式能够更准确的表征人员行为的随机性。然而,尽管在随机性细节上存在差异,将两种方法获取的人员行为数据转化为以均值体现的在室率后,差异却显著减小。这也证明了人员行为随机性确实存在着相互抵消的效果。通过进一步将标准中采用的人员行为数据以及采集获得的人员行为数据导入能耗模拟软件EnergyPlus开展能耗模拟,并与实测值进行对照分析发现,不同时间尺度下的模拟精度存在极大差异,从周精确到天再到小时的过程中,模拟的相对误差均值分别提升了近20倍和近200倍。以上结果反映出,需要根据使用需求来选择合理的能耗模拟时间尺度,并匹配合适的数据采集方法。Occupant behavior plays an important role in building energy consumption simulation.Different data collection methods are used to understand the occupant's behavior in the building.Different methods may cause differences in the occupant's model,especially in the representation of the randomness of the occupant's behavior,different methods will get different random behaviors in different time scales.To compare the differences in occupant behavior and the impact of it on the accuracy of energy consumption simulation in different data acquisition methods,this paper selected the student dormitories,using the questionnaire and Global Positioning System(GPS)to gain the relevant occupant data and measuring the actual electricity consumption data.The results show that using the GPS can more accurately characterize the randomness of occupant behavior.However,there are some differences in details about randomness,the differences reduced rapidly when converting the data into the room ratio based on the average.It shows that the randomness can cancel each other.By further simulation based on inputting the occupant behavior data used in the standard and data collected to Energyplus,and comparing the real data and the simulation's results,the results revealed that in different time scales,the relative error can improve by 20 times to 200 times from week to hour.This shows that the reasonable energy consumption time scale should be selected based on the usage and be matched with a suitable data acquisition method.

关 键 词：能耗模拟 人员行为 问卷统计 传感器采集 随机性 

分 类 号：TU831[建筑科学—供热、供燃气、通风及空调工程] TU111.19