Smart home technologies are becoming prevalent in residential buildings. Despite interface and control improvements offered by smart thermostats, previous work has found energy savings from this technology to be variable. Further, little is known about how smart thermostat operation differs from conventional programmable thermostat operation, which has been studied extensively. In this paper, data from 54 smart thermostats installed in two high-rise residential buildings are used to characterize smart thermostat programming and occupant interaction behaviours. Survey data are also used to examine how environmental attitudes and technical abilities influence thermostat usage behaviours. Smart thermostat programming rates were higher than those previously found for conventional programmable thermostats, likely due to improved thermostat interfaces. Further, programmed schedule accuracy was high (88% occupancy during scheduled unoccupied periods). Participants programmed substantial setforwards (cooling season average of 2.8 °C) and setbacks (heating season average of 3.8 °C) to be used when their home is unoccupied. Override behaviours were generally not problematic: most overrides had short durations or were frequently adjusted by users. Strong environmental attitudes and technical skills also appeared to result in more energy-conscious thermostat use behaviours, however, due to the population demographics, specific relationships between energy conservation attitudes and thermostat use behaviours were unclear. An analysis of thermostat schedules and setpoints considering measured occupancy data indicated that participant-programmed schedules achieved significant HVAC load reductions when compared to a scenario in which no schedule was programmed. Further, the use of occupancy-based controls, whether alone or in combination with a schedule, was shown to improve energy savings.

智能家居技术在住宅建筑中正变得越来越普遍。尽管智能恒温器提供了界面和控制方面的改进，但以前的工作发现，这项技术所节省的能源是可变的。此外，关于智能恒温器操作与常规可编程恒温器操作有何不同，人们对此知之甚少，传统的可编程恒温器操作已被广泛研究。在本文中，使用安装在两座高层住宅楼中的54个智能恒温器的数据来表征智能恒温器编程和乘员交互行为。调查数据还用于检查环境态度和技术能力如何影响恒温器的使用行为。智能恒温器的编程速率比以前发现的传统可编程恒温器更高，这可能是由于改进了恒温器接口所致。此外，计划的日程安排准确性很高（在计划的无人入住期间占用率为88％）。参加者设定了当他们的房屋闲置时要使用的大幅降温（平均制冷季节为2.8°C）和降温（平均降温季节为3.8°C）。覆盖行为通常没有问题：大多数覆盖持续时间短或经常由用户进行调整。强烈的环境态度和技术技能似乎也会导致人们对节温器的使用更加自觉，但是由于人口统计的原因，节能态度与节温器使用行为之间的具体关系尚不清楚。根据测得的占用数据对恒温器时间表和设定点进行的分析表明，与没有计划时间表的情况相比，参与者编程的时间表可显着降低HVAC负荷。此外，无论是单独使用还是与计划结合使用基于占用率的控件，都可以改善节能效果。