ASHRAE Journal - January 2021 - 34

TECHNICAL FEATURE

popsicles on top of locked thermostats to taping shiny
objects on windowsills to reflect daylight onto sensors,
occupants are smart, and they know how to beat the system. In one study, we interviewed 170 office occupants in
their offices and found that more than half the offices in
some buildings had plug-in space heaters-some of which
were running in the cooling season.7 Often these occupant
interventions are more energy-intensive than providing
occupant control in the first place would have been.
The Standard 55-2017 adaptive thermal comfort model
requires as one necessary condition that buildings have
operable windows. Let us think about the inherent traits
of operable windows that give occupants greater tolerance for a range of indoor air temperature. Operable
windows have a fast-acting response (i.e., immediate
sensation of air movement and fresh air) and are highly
intuitive to occupants. When applying the framework of
Figure 1 to operable windows, we see that the interface,
control logic and actuator are all combined without any
advanced digital technology. This sets them apart from
the majority of modern central HVAC systems. The question emerges: what traits can we borrow from operable
windows to improve other interfaces?

Reevaluating Default Settings

The default setting that an interface and underlying
system uses is critical to performance. It is likely to be
the predominant state that the system is left in; thus it
is ideally optimal for both energy and occupant satisfaction. In a study of lighting use in a daylit office building,8
the researchers discovered that the lighting was set to
turn on automatically upon occupancy of the room. A
post-occupancy evaluation revealed that occupants were
generally dissatisfied with this setting. Meanwhile, the
overhead lights turned on upon occupant entry and
remained on for 15 minutes, leading to an enormous
waste of energy considering the mobile in-and-out
occupancy patterns of office workers. The problem was
so severe (in relative terms) that about 20% of occupants
had independently devised methods to re-engineer
the controls-primarily by permanently or temporarily covering the occupancy sensors. When the controls
were changed to be manual-on and vacancy-off, the
year-over-year energy savings was 62%. Meanwhile, 24
of 25 surveyed occupants were more satisfied with the
new control logic. We note that this particular issue is
now addressed by building codes and standards (e.g.,

ASHRAE JOURNAL

ashrae.org

JAN UARY 2021

Standard 90.1-2019), thus illustrating their important role in regulating such subtle but important
specifications.

Emerging Technologies

Despite the examples used in this article, we note that
several technologies are emerging or reemerging to
improve occupant feedback and usability. These include
desk or ceiling fans,9 smart phone-operated HVAC terminal units with voting10 and personalized controls (e.g.,
chairs with personal control systems11). Meanwhile,
ubiquitous sensing is making it possible to provide much
richer input to BASs.12 Moreover, mobile/smart phone-
based interfaces could create a major paradigm shift to
address some of the challenges expressed in this article.

The Path Forward

This article has provided a framework for understanding the key characteristics of building interfaces and
their relationship to established heuristics from the field
of human factors. We have used numerous anecdotes to
argue that there is significant room for improvement in
design practice, with many areas for further research
and development. In closing, we provide several concrete steps to improve the status quo of building interfaces-to simultaneously improve usability, occupant
comfort and energy performance:
We need to close the loop between human input and
feedback to improve design for future buildings. This
statement has been touted by post-occupancy evaluation (POE) enthusiasts (e.g., Bordass and Leaman13) for
decades. As noted above, an interface is only the surface
of human-building interaction; usability depends on
all seven components outlined in the conceptual model
shown in Figure 1. While we feel there is significant room
for improvement in interfaces, occupant experience is
only as good as the way the interface is integrated into
the building.
While POE surveys and laboratory-based testing (e.g.,
think-aloud and eye-tracking methods) can obtain subjective occupant feedback, a newer promising method is
via in-situ data collection.14 An interface can automatically collect and analyze the sequence and rapidity of
interactions with the interface to help understand the
efficiency and effectiveness with which occupants are
using it. For example, one study established a set of metrics such as " button mashing " and " confusion " to assess

https://www.ashrae.org/

ASHRAE Journal - January 2021

Table of Contents for the Digital Edition of ASHRAE Journal - January 2021