Thermal properties of building envelopes are often described using thermal conductivity or thermal resistance. And the opposite task involves the identification of thermal parameters of building envelopes based on the measurements of their cooling process. In this article, the authors proposed a method of identifying thermal parameters of a building envelope based on cooling measurements, using a multiple regression model for this purpose. To satisfy the research objectives, two basic experiments were carried out. The first experiment was performed in laboratory conditions. The research model was a cube of the dimensions of 1.1 m × 1.1 m × 1.1 m. The second experiment was carried out in semi-real conditions, and the used model was a small house of the dimensions of 6.0 m × 4.15 m × 5.2 m. The measurement results were also used to calibrate numerical models made in the ESP-r program. The research studies have demonstrated that the model can be used to identify thermal parameters of a building envelope. Based on the measurements and simulations, the cooling equations of the object were determined and the 95% confidence interval for the heat retention index was estimated. On that basis, using the multiple regression model, such parameters of the model as density, specific heat, and thermal conductivity were estimated. It turned out that using the Gauss–Newton approximation, we obtained the correlation of the measurement results and the analytical model with the correlation coefficient of 0.9971 (for the laboratory scale). And the multiple regression improved not only the correlation between the measurement and the analytical model, but it also allowed to obtain “almost identical” results. Similarly, promising results were obtained for the semi-real scale.

中文翻译：

---

基于建筑物体冷却过程的建筑围护结构热参数识别

建筑围护结构的热性能通常使用热导率或热阻来描述。而相反的任务涉及根据其冷却过程的测量来识别建筑围护结构的热参数。在本文中，作者提出了一种基于冷却测量识别建筑围护结构热参数的方法，为此使用了多元回归模型。为了满足研究目标，进行了两个基本实验。第一个实验是在实验室条件下进行的。研究模型是一个尺寸为 1.1 m × 1.1 m × 1.1 m 的立方体。第二个实验是在半真实条件下进行的，所使用的模型是一个尺寸为6.0 m × 4.15 m × 5.2 m的小房子。测量结果还用于校准在 ESP-r 程序中制作的数值模型。研究表明，该模型可用于识别建筑围护结构的热参数。基于测量和模拟，确定了物体的冷却方程，并估计了保温指数的 95% 置信区间。在此基础上，利用多元回归模型，估计了模型的密度、比热、导热系数等参数。结果表明，使用高斯-牛顿近似，我们获得了测量结果与分析模型的相关性，相关系数为 0.9971（实验室规模）。多元回归不仅改善了测量与分析模型之间的相关性，但它也允许获得“几乎相同”的结果。同样，在半真实尺度上也获得了有希望的结果。