Systems, Man & Cybernetics - October 2017 - 30

For example, the earliest expected arrival and latest
procedures often require building science and HVAC
expected departure times in a building can be identiexperts to be on site with auditing devices/equipment,
fied from the detections of a passive-infrared motion
such as thermal cameras, tracer gas testing equipment,
detector [29]-[31] or other forms of occupancy sensing
or blower door fans, to measure envelope air tightness
[32]. The temperature and indoor illuminance prefer[41], airflow capture hoods, and micromanometers. The
ences of occupants can be learned from their interacdisruption caused by ongoing commissioning procedures
tion patterns with thermostats and light switches and
can make it challenging to conduct such invasive tests in
concurrent indoor climatic conditions [33], [34]. For
occupied spaces.
Remote auditing is an emerging research field that
example, in a recent field--implementation study conemploys existing sensor networks in a building to moniducted in private offices [35], indoor temperature and
tor envelope health [42], [43]. It relies on various inverse
relative humidity measurements at the thermostat
modeling methods to assimilate the data from sensor
keypress instances were used to dynamically adapt
networks within the underlying physical phenomena
the set points. Adaptive control is best suited to build(e.g., heat, air, and moisture transfer within building
ings with flexible and intermittent occupancy schedenclosures). For example, in a few case studies, the
ules. In a ca se study, the deploy ment of adaptive
possibility to estimate envelope air leakage by lookoccupant-driven controls on a floor of an academic
ing at the after-hours CO 2 concentration decay patoff ice build ing (used by professors w ith f lex ible
schedules) achieved roughly 40% HVAC and lighting
terns in office spaces was explored [44]-[48]. In a
load reductions [2].
similar fashion, inverse modelThe dema nd response is a
ing for the indoor temperature
method to upgrade the controls
response (driven by the occupant
of a building to reduce its energy
and environment-driven thermal
Remote auditing is
costs [36]-[38]. For many comloads) can provide information
an emerging research
mercial buildings, a large fraction
about the thermal insulation perof the energy cost is due to their
formance [49].
field that employs
peak power dema nd (i.e., the
Aside from envelope health
existing sensor
highest power demand in a billing
monitoring, distributed and connetworks in a
cycle). Peak dema nd cha rges,
nected sensing can be used to
while varying in time and locacharacterize the equipment and
building to monitor
tion, can represent 10-30% of the
control infrastructure continuenvelope health.
total energy costs. Thus, demand
ously while the building is in serresponse in commercial buildvice [50]. This is also known as
ings looks at reducing the peak
virtual sensing/metering technolloads. However, for residential
ogy [10], [51], [52]. For example, a
buildings, the energy costs are affected by the time-ofvirtual sensor can monitor the fraction of outdoor air
use pricing. For example, using a baseboard heater
supplied by an AHU from the heat-mass balance relaafter 7 a.m. in Ontario, Canada, can cost three times as
tionships. In a similar fashion, the efficiency of a boiler
much as using it after 7 p.m. Consequently, the demand
can be monitored under different part-load and weather
response in the residential sector looks at load shifting
conditions, and the degradation of an AHU supply fan's
to reduce energy costs. Models driven by connected
efficiency can be estimated from the temperature rise
and distributed sensing and real-time utility pricing can
patterns across the fan at various air pressure levels
make energy demand forecasts, upon which peak load[10]. These pieces of information are expected to help
reducing or load-shifting decisions can be executed.
make better HVAC retrofit decisions.
Beyond individual buildings, the connectivity of sensors across a community will result in a communityNontechnical Barriers
scale demand response approach in which on-site
We have listed common methods that utilize the data colrenewables and thermal storage as well as energy
lected from sensor networks to optimize the operation of
demand from the utility grid can be harmonized. This is
our homes and buildings. Despite the scientific research
also known as the smart grid [39].
on the topic, two recent market surveys indicate that distributed and connected sensing has changed very little in
terms of the way we maintain, control, and retrofit our
The IoT and Decision-Making for Retrofits
buildings [6], [53]. Before delving into the technical barriOngoing commissioning in buildings is a process to ideners, we will discuss the nontechnical barriers against the
tify degrading building systems and components so that
wide adoption of the methods presented in the "Value
appropriate retrofit decisions can be made. However, it is
Proposition of the IoT to Facility Managers and Homeoften labor intensive and can be cost prohibitive (approxowners" section.
imately US$3 -10/m 2 [40]). Ongoing commissioning
30

IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE Oc tob e r 2017

Table of Contents for the Digital Edition of Systems, Man & Cybernetics - October 2017