Residential smart thermostat use: An exploration of thermostat programming, environmental attitudes, and the influence of smart controls on energy savings

Abstract

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.

Introduction

The adoption rate of smart home technologies, particularly smart thermostats, in residential buildings is growing. Smart thermostats are thermostats which can be controlled remotely through the use of a phone or web application and/or through integration with other ‘internet-of-things’ devices. Other features, such as reactive- or predictive-occupancy based control, maybe be included in the device, however, these features vary between manufacturers. In 2016, 7.8 million homes in North America had smart thermostats, with this number expected to grow to 43.4 million by 2021 [1]. Smart thermostats allow for programming of temperature setpoint schedules, similar to their conventional, programmable thermostat counterparts (which do not provide online access options and must be programmed through the device itself), and offer new opportunities to implement alternative heating, ventilation and air conditioning (HVAC) control strategies, which can reduce energy use from these systems [2], [3], [4]. However, similar to the limitations found with conventional programmable thermostats [5], [6], [7], the potential energy savings of smart thermostat technologies are limited by the willingness and ability of users to program accurate thermostat settings and adopt energy-conscious thermostat temperature setpoints. Modern, smart thermostats may increase thermostat programming rates through improved user interfaces over those on conventional programmable thermostats, which have commonly been found to be confusing or inaccessible to users [7], [8], [17], [9], [10], [11], [12], [13], [14], [15], [16]. However, even if every household programs their thermostat with a setpoint schedule, energy savings may be limited by the temperature setpoints selected by users and how accurately programmed thermostat schedules reflect actual home occupancy and usage patterns. These limitations can, in part, be addressed through the use of alternative HVAC control methods. Occupancy-based control is a common feature in commercially-available smart thermostats [18], [19], [20] and offers opportunities to reduce the need for accurate programming of thermostat schedules. However, user choice of temperature setpoints for periods in which their home is unoccupied and suite-usage schedule programming may still limit their realized energy conservation.

In this paper, survey and thermostat data collected from two high-rise residential buildings in a cold climate are used to characterize occupant programming behaviours for smart thermostats; including temperature setpoints, scheduling, and schedule override use, and their relationship to self-reported environmental attitudes and technical skills. Finally, a preliminary assessment of how thermostat settings and occupant-override behaviours impact energy savings is presented.