基于移动人流数据的商业楼宇负荷用电预测

COMMERCIAL BUILDING LOAD PREDICTION BASED ON MOBILE PEOPLE DATA

建筑能耗的准确预测为电网系统进行高效管理和合理分配电能资源奠定了重要基础,而结合外因数据预测商业楼宇的负荷用电量是一个难点问题。根据建筑内人流量对用电负荷的影响,提出一种将融合移动人流数据的多维时间序列预测模型LMO(Linear Model with Occupancy)。利用楼宇负荷用电和移动人流时间序列的耦合性,LMO融合了多维特征,提高楼宇用电量预测的准确性。实验结果表明,LMO能够引入更多的先验知识,减少不确定性。因此,相比于传统方法,该预测方法具有更高的预测精度。

Accurate prediction of building energy consumption lays an important foundation for efficient management of power grid systems and rational allocation of power resources.It is a difficult problem to predict the load power consumption of commercial buildings combining external data.According to the influence of people flow on electricity load in buildings,this paper proposed a multi-dimensional time series prediction model which integrated mobile people data---linear model with occupancy(LMO).Using the coupling of building load power and mobile people flow time series,LMO combined multi-dimensional features to improve the accuracy of building electricity consumption prediction.Experimental results show that LMO can introduce more prior knowledge and reduce uncertainty.Therefore,compared with the traditional method,the prediction method has higher prediction accuracy.