通过ASHRAE赞助的一个项目（ASHRAE RP-1092），为常用暖通空调系统开发了简化建筑能耗模拟模型的校准程序。该程序适用于五栋建筑。这五栋建筑包括一栋44层办公楼、一栋低层办公楼、一座教堂、一座大学教学楼和一座大学研究楼。五座建筑中有三座位于内布拉斯加州奥马哈（寒冷气候）。两栋建筑位于德克萨斯州大学站（湿热气候）。本文介绍了五个案例研究的校准程序和总结结果。两级校准程序为模型校准提供了一种很好的方法。案例研究强烈表明，所开发的简化模型校准程序可用于精确计算长时间模型- 使用不同系统的不同类型建筑的短期现场能量测量数据的短期能耗数据。第一级校准程序非常重要，可以显著提高模型精度。一级校准后，这些建筑模型的NMBE平均绝对值从41%降至6%，CV（RMSE）从63%降至26%。二级校准程序进一步将这些值分别降低至2.5%和22.4%。在使用四周或更短时间内测量的小时数据校准模拟后，年度消耗值的预测误差绝对值的平均值为2.5%，所检查案例研究的最大误差为9%。 建筑物和HVAC系统的一般信息是简化模型校准的最关键输入参数。短期每小时冷水消耗量和热水消耗量是校准最关键的能源数据。引文:内华达州拉斯维加斯ASHRAE Transactions第117卷第1部分

A calibration procedure for simplified building energy simulation models for commonly used HVAC systems has been developed through an ASHRAE-sponsored project (ASHRAE RP-1092). The procedure is applied to five buildings. These five buildings include a 44-story office building, a low-rise office building, a church, a university teaching building, and a university research building. Three of the five buildings are located in Omaha, Nebraska (cold climate). Two buildings are located in College Station, Texas (hot and humid climate). This paper presents the calibration procedure and summary results of five case studies. A two-level calibration procedure provides a good approach for model calibration. The case studies strongly indicate that the simplified model calibration procedure developed can be used to accurately calculate long-term energy consumption data using short-term field energy measurement data for different types of buildings with different systems. The first-level calibration procedure is very important and improves the model accuracy significantly. The average absolute value of NMBE decreased from 41% to 6%, and CV(RMSE) decreased from 63% to 26% for these building models after the first-level calibration. The second-level calibration procedure further reduced these values to 2.5% and 22.4%, respectively. After calibration of the simulation using four weeks or less of measured hourly data, the average of the absolute values of the prediction error for annual consumption values was 2.5%, with a maximum error of 9% for the case studies examined. General information on the building and HVAC systems are the most critical input parameters for the simplified model calibration. Short-term hourly chilled-water consumption and hot-water consumption are the most critical energy data for calibration.