建筑能耗分析用气象数据的构成研究

建筑能耗分析需要首先构成有代表性的全年气象数据，气象数据可以根据一定规则从以往数据中选取，也可以用以往数据建立气象数学模型，生成可以代表未来的气象数据。对F检验法与AIC准则建模进行了比较，表明F检验法建模是合理的，用F检验逐步回归法建立了13维疏系数混合回归气象数学模型，对模型残差进行了残差独立性、正态性及零均值检验，结果表明模型是成功的，并且维数增高可以为模型提供更多的信息，用所得模型构成了建筑能耗计算用气象数据。

Constitute of Meteorological Data in Building Energy Analysis

Building energy analysis needs representative yearly meteorological data,the data could be selected from the past years' data according to specific regulation,it can also be produced from the meteorological math model which was built with the past data. A comparison is made between F testing method and AIC criterion method, which leads a result that the F testing method is reasonable. Mixed regression model of thirteen dimensions is established using F testing stepwise method. Test is conducted on independence and distribution of model residuals, and the result shows that the model is successful and more information can be provided for establishing model with dimension number increased. Using the above established model future meteorological data of building energy analysis can be predicted.