The energy performance of residential buildings significantly depends on the building occupants’ behavior, which can be highly variable. When the heating, ventilation and air conditioning (HVAC) system is controlled based on the presence or absence of occupants in a building, occupant behavior is of even further importance to its energy performance. In current practice, building energy simulation tools generally use a single occupancy profile to represent the building’s occupancy schedule, the schedule of which is considered to be the same, regardless of the type of household being modeled. Thus, there is significant potential for improvement to allow for more flexibility and accuracy in calculation of occupancy. The objective of this study is to assess the variations in the typical types of occupancy schedules followed by the U.S. population using cluster analysis. American Time Use Survey data, which statically represents the overall U.S. population’s activities, across 12 years (2006–2017), is used. The ATUS data is segregated into smaller groups based on age and weekday/weekend, then divided into activities that are considered “at home” and “away from home”, which are mapped to the presence or non-presence of occupants in the home. Cluster analysis is then used to identify common types of occupancy schedule patterns for each age group. Three main types of patterns are obtained from cluster analysis for each age group, which together represent approximately 88% of people in the United States. The output of the cluster analysis is further analyzed to evaluate the variation in characteristics, including the number of times leaving home, time of day when leaving the home, and the timespan of absence from the home. The results of this study provide detailed insights on how typical occupants in the United States spend their time in residential spaces which can be used to create occupancy profiles for residential buildings. These occupancy profiles could be utilized inform an assessment of the energy use impact of occupancy-based controls of energy consuming systems and technologies.

住宅建筑物的能源性能在很大程度上取决于建筑物居民的行为，行为可能会高度变化。当根据建筑物中居住者的存在或不存在来控制供暖，通风和空调（HVAC）系统时，居住者的行为对其能源性能甚至更为重要。在当前的实践中，建筑物能源模拟工具通常使用单个占用概况来表示建筑物的占用时间表，无论所建模的家庭类型如何，该时间表均被视为相同。因此，存在很大的改进潜力，以允许在占用率计算中具有更大的灵活性和准确性。这项研究的目的是评估美国紧随其后的典型入住时间表类型的差异。人口使用聚类分析。使用美国时间使用情况调查数据，该数据静态表示12年（2006-2017年）内美国总体人口活动。根据年龄和工作日/周末将ATUS数据分成较小的组，然后分为被认为是“在家”和“在家外”的活动，这些活动被映射为在家中有无居住者。然后，使用聚类分析来确定每个年龄组的常见的入住时间表类型。从每个年龄组的聚类分析中可以获得三种主要类型的模式，这些模式合在一起代表了美国约88％的人。进一步分析聚类分析的输出，以评估特征变化，包括出门次数，出门时间，以及出门在外的时间。这项研究的结果提供了有关美国典型居住者如何在住宅空间中度过的时间的详细见解，这些空间可用于创建住宅建筑物的居住情况。可以利用这些占用概况来评估基于能耗的系统和技术的基于占用的控制对能源使用的影响。