Energy consumption prediction for buildings and architectural complexes has become increasingly important. However, existing methods based on simulation software or statistical algorithms have distinct disadvantages. Consequently, this study introduces a novel prediction method based on models between occupant behavior and energy consumption. A new parameter––the occupant behavior characteristic parameter, is defined by coupling factors, including occupant behavior probability, occupant numbers, and building operation hours. Correlation models between this characteristic parameter (Parameter A) and electricity consumption (Parameter B) are established. Moreover, because of the difficulty in obtaining occupant numbers, models between environment, time influence factors (parameter C) and occupant numbers are proposed for occupant number prediction. In summary, models of the three parameters (A, B, and C) can be combined to obtain an electricity consumption prediction model. On the one hand, this model can reflect energy use laws with a higher accuracy than prediction methods based on temperature or other factors; on the other hand, it can be applied online without complex algorithms and software simulation. Moreover, it needs only model one equation for the same building types, which obviates modeling building-by-building like software simulation. Thus, the proposed model offers strong operability, especially the prediction for architectural complexes.

建筑物和建筑群的能耗预测变得越来越重要。然而，现有的基于仿真软件或统计算法的方法具有明显的缺点。因此，本研究引入了一种基于乘员行为和能源消耗之间模型的新预测方法。一个新的参数——居住者行为特征参数，由耦合因素定义，包括居住者行为概率、居住者人数和建筑物运行时间。建立了该特征参数（参数A）与耗电量（参数B）的相关模型。此外，由于难以获得乘员人数，环境之间的模型，时间影响因素（参数 C ) 和乘员人数被提议用于乘员人数预测。综上所述，可以将三个参数（A、B、C）的模型组合起来，得到电力消耗预测模型。一方面，该模型可以比基于温度或其他因素的预测方法更准确地反映能源使用规律；另一方面，它可以在线应用，无需复杂的算法和软件模拟。而且，它只需要为相同的建筑类型建模一个方程，这避免了像软件模拟那样逐个建筑建模。因此，所提出的模型提供了很强的可操作性，尤其是对建筑群的预测。