当为输入变量指定新值时，模拟模型估计的建筑能耗会发生变化。在最近的一项研究中，测试了建筑能源模型对25个输入变量变化的敏感性。某些能量变化比其他能量变化更显著，这表明在建模过程中应特别注意选定的输入。这项研究集中在太平洋西北部的两座建筑上。其中一栋建筑是小型零售建筑的代表，另一栋是大型办公建筑的代表。使用五个不同的软件包对两栋建筑中的每栋重复进行敏感性分析，得出10个建筑能量模型。两种建筑类型中，任何一种测量（输入变化）的敏感度都不同。总体而言，模型对影响建筑围护结构和照明的措施相对不太敏感，而对涉及占用率、天气、空气供应、系统和工厂的措施更敏感。 关键词:建筑、计算、能耗、测试、美国、办公室、商店、计算机程序、比较、节能、准确度:ASHRAE Trans。1992年，第98卷，第1部分，论文编号AN-92-1-3，618-626，图9。，4个标签。，参考文献6。

The building energy estimated by a simulation model changes when new values are assigned to input variables. In a recent study, building energy models were tested for sensitivity to changes in 25 input variables. Certain energy changes were more significant than others, suggesting that selected inputs should be given particular attention during modelling. The study focused on two buildings in the Pacific Northwest. One building was representative of small retail buildings and the other was an example of large office buildings. The sensitivity analysis was repeated for each of the two buildings using five different software packages, resulting in 10 building energy models. The degree of sensitivity noted for any one measure (input change) differed between the two building types. In general, the models were comparatively less sensitive to measures affecting the building envelope and lighting and more sensitive to measures involving occupancy, weather, air supply, systems, and plant.KEYWORDS: buildings, calculating, energy consumption, testing, USA, offices, shops, computer programs, comparing, energy conservation, accuracy