ASHRAE Journal - March 2023 - 27

FEATURE
A
rtifi cial intelligence (AI) has the potential to operate buildings more energyeffi
ciently, sustainably and comfortably. Many potential applications exist for AI in the built
environment sector, including building design, building performance optimization, energy
management and monitoring and building maintenance and operations. One might be
able to understand the concept of integrating AI technologies in self-driving cars, but when
it comes to the integration of AI in building design and operations, the topic seems much
more complicated. In this column, we will try to explain the signifi cance of AI in optimizing
the performance of buildings and discuss its applications and limitations.
Background
Although businesses have used AI for many years,
it has only been in recent years that it has started to
become widespread in consumer applications. This is
due to improvements made to and the abundance of
hardware and software technologies underlying AI, such
as the availability of faster processing speeds and the
reduced cost of storage. As artifi cial intelligence continues
to evolve and improve, there is increasing discussion
about how AI may be used in the built environment.
To understand AI and its applications in buildings, let's
fi rst defi ne AI. It is a technology that aims to emulate
human intelligence, including understanding natural
language, recognizing images and making decisions
through learning from past experiences. AI enables
computers and machines to complete complex tasks that
typically require human intelligence.
AI for High-Performance Building Design
According to the U.S. Energy Information
Administration, building energy consumption accounts
for about 30% of global energy consumption.1 To support
energy-effi cient built environment design, planners,
architects, engineers and designers began exploring the
energy performance of their designs using a new set of
building performance simulation (BPS) tools that enable
architects and engineers to examine the thermal comfort,
daylighting performance and energy performance
of various building confi gurations. Developing energyeffi
cient building designs is crucial in the early design
phase when decisions have the most signifi cant impact
on the fi nal building's energy consumption and costs.2 - 4
Despite the great potential of BPS tasks to facilitate
high performance buildings' design, they are usually
time-consuming and computationally expensive.4,5
Practitioners in the construction industry usually
adhere to strict project deadlines that prevent them
from performing lengthy simulation tasks, and they end
up abandoning these simulations. To mitigate this issue,
researchers have examined the use of different AI algorithms
to automate BPS tasks.
Researchers have demonstrated that AI algorithms,
specifi cally neural networks (NNs), a deep learning
method, can accurately predict energy consumption
and other performance aspects of buildings.6 - 9 For
example, Wong, et al., examined the use of an NN model
to predict the energy and daylighting performance of
an offi ce building. The researchers used a parametric
building model that had nine variables as the input
parameters: four related to external weather conditions
(daily average dry-bulb temperature, daily average
wet-bulb temperature, daily global solar radiation and
daily average clearness index), four related to building
envelope design (solar aperture, daylight aperture, overhang
and side-fi ns projections), and a day-type variable
(i.e., weekdays, Saturdays and Sundays). The NN model
estimated daily electricity use for cooling, heating and
lighting. The accuracy of the model's cooling, heating,
electric lighting and total building electricity use metrics
was 99.4%, 94.0%, 99.3% and 99.6%, respectively, indicating
excellent predictability.6
Other studies have successfully used NN models to
accurately predict the thermal performance of buildings
for the ultimate energy-effi cient and comfortable
Rania Labib, Ph.D., is assistant professor and director of the Artifi cial Intelligence for High Performance Buildings Lab at Prairie View A&M University. Zoltan Nagy, Ph.D., is an
assistant professor at the University of Texas at Austin.
M A R C H 2 0 2 3
ashrae.org
a s h r a e . o r g
ASH AE JOURNAL
A S H R A E J O U R N A L
27
27
ISTOCK.COM/PHONLAMAIPHOTO
http://www.ISTOCK.COM/PHONLAMAIPHOTO http://ashrae.org
ASHRAE Journal - March 2023

Table of Contents for the Digital Edition of ASHRAE Journal - March 2023
