基于能耗监测数据的校园建筑贝叶斯能耗预测模型
Bayesian energy consumption prediction model of campus buildings based on energy consumption monitoring data
摘要：
对于建筑类型众多、用能规律复杂的校园建筑群，建立区域能耗模型是分析其能源使用情况的高效途径。自下而上的区域能耗模型不仅可以实现对区域内建筑能耗的预测，还可以评估新技术与新能源的节能情况，其建模方法目前普遍基于面积扩展的方式。提出了一种基于能耗监测数据的贝叶斯能耗预测方法，对寒冷地区某高校建筑群分别建立了简单面积扩展预测模型和贝叶斯能耗预测模型。以混合类科研建筑为例，面积扩展模型月误差在-30%～-7%之间，而贝叶斯模型月误差在-5%～15%之间，月误差绝对值明显降低。贝叶斯能耗预测模型对校园建筑群总能耗的预测月误差在-2%～6%之间，年误差仅为1.07%。
Abstract：
For campus buildings with numerous types and complex energy use rules, establishing a regional energy consumption model is an efficient way to analyse their energy use. The bottom-up regional energy consumption model can not only predict the building energy consumption in the region, but also evaluate the energy conservation of new technology and new energy resources. At present, its modeling method is generally based on area expansion method. Proposes a Bayesian energy consumption prediction method based on energy consumption monitoring data. Establishes a simple area expansion prediction model and a Bayesian energy consumption prediction model for a campus building complex in cold zone. Taking the mixed scientific research buildings as an example, the monthly error range of the area expansion model is between -30% and -7%, while that of the Bayesian model is between -5% and 15%, and the absolute value of the monthly error is significantly reduced. The monthly error range of the Bayesian energy consumption prediction model for the total energy consumption of the campus building complex is between -2% and 6%, and the annual error is only 1.07%.
Keywords:campusbuilding,energyconsumptionprediction,bottom-upregionalenergyconsumptionmodel,Bayesianestimation,energyconsumptionmonitoringplatform,areaexpansionmodel