基于贝叶斯MCMC方法的城市建筑供热能耗基准  

Urban building heating energy consumption benchmark based on Bayesian MCMC method

作　　者：那威 邴佳乐 刘品妍 NA Wei;BING Jia-le;LIU Pin-yan(School of Environment and Energy Engineering,Beijing University of Civil Engineering and Architecture,Beijing 100044,China;School of Architecture and Urban Planning,Beijing University of Civil Engineering and Architecture,Beijing 100044,China)

机构地区：[1]北京建筑大学环境与能源工程学院,北京100044 [2]北京建筑大学建筑与城市规划学院,北京100044

出　　处：《浙江大学学报（工学版）》2023年第10期2106-2115,共10页Journal of Zhejiang University：Engineering Science

基　　金：北京市社会科学基金资助项目(20GLB025).

摘　　要：城市尺度建筑能耗基准的研究模型需解决建筑样本数据有限、输入不确定性强、建模复杂度和计算成本高的问题,据此提出贝叶斯框架的混合效应模型.利用城市气候、建筑分布和建造状况等特征指标作为模型输入,采集223座居住建筑的热力站能耗监测数据作为测试集样本,通过马尔科夫链蒙特卡洛(MCMC)方法进行重要性采样,并预测城市尺度建筑供热能耗强度(EUIH).混合效应模型通过48座居住建筑的热力站能耗监测数据进行验证,预测精度指标归一比平均偏差(NMBE)为0.400%,均方根误差(RMSE)为0.034,均方根误差变异系数(CVRMSE)为13.100%.结果显示模型能够有效地预测和校准建筑供热能耗预测和校准,模型的EUIH点估计和区间估计结果可以作为城市建筑供热能耗基准依据.北京市城市建筑的供热能耗基准与国际能源署(IEA)国家的相关基准进行对比,反映出北京市居住建筑供热能耗较低及节能潜力较小.The model of benchmarking building energy use in city-scale should tackle the challenges on reliable evaluation and calibration for building energy consumption,regarding issues of insufficient building samples,input uncertainties,modelling complexity,and computing cost in the development.A mixed-effect probability model was proposed in accordance with Bayesian theory.The characteristic indicators such as urban climate,building distribution and construction status were used as model inputs.The metering data of energy consumption from 223 heating substation samples were used as training data.Markov Chain Monte Carlo(MCMC)simulation was utilized to conduct importance sampling,and building energy use intensity for space heating(EUIH)in city-scale were calibrated.The mixed-effect probability model was validated by a new dataset with 48 heating substations,which the values of normalized mean bias error(NMBE),root mean square error(RMSE),and cofficient of variation root mean squared error(CVRMSE)indices were 0.400%,0.034,13.100%.The results show that the model can be effectively used to calibrate building energy consumption for heating.The EUIH point and interval estimation results of the model can serve as the benchmark for the urban building heating energy consumption.The proposed benchmark of city-scale building heating energy was compared with that of international energy agercy(IEA)member countries,and the reveal resident building in Beijing had lower EUIH and less potential on the energy conservation for heating.

关 键 词：建筑能耗基准 贝叶斯理论 马尔科夫链蒙特卡洛方法 城市尺度 建筑能耗预测 

分 类 号：TU111.195[建筑科学—建筑理论]