sustainability

Article
Artificial Intelligence (AI)-Based Occupant-Centric Heating
Ventilation and Air Conditioning (HVAC) Control System for
Multi-Zone Commercial Buildings
Alperen Yayla 1 , Kübra Sultan Świerczewska 2 , Mahmut Kaya 3 , Bahadır Karaca 4 , Yusuf Arayici 5 ,
Yunus Emre Ayözen 6 and Onur Behzat Tokdemir 7, *

1 Department of Civil and Environmental Engineering, Imperial College London, Skempton Building,
London SW7 2AZ, UK
2 Cundall Polska, 00-582 Warszawa, Poland
3 KPD Engineering & Consultancy, Bursa 16090, Türkiye
4 Nuclear Islands Department, Akkuyu Nuclear Power Plant, Mersin 33715, Türkiye
5 Department of Architecture and Built Environment, Northumbria University,
Newcastle upon Tyne NE1 8ST, UK
6 Strategy Development Department, Ministry of Transport and Infrastructure, Ankara 06338, Türkiye
7 Department of Civil Engineering, Istanbul Technical University, Istanbul 34467, Türkiye
* Correspondence: otokdemir@itu.edu.tr

Abstract: Buildings are responsible for almost half of the world’s energy consumption, and approx-
imately 40% of total building energy is consumed by the heating ventilation and air conditioning
Citation: Yayla, A.; Świerczewska,
(HVAC) system. The inability of traditional HVAC controllers to respond to sudden changes in
K.S.; Kaya, M.; Karaca, B.; Arayici, Y.;
occupancy and environmental conditions makes them energy inefficient. Despite the oversim-
Ayözen, Y.E.; Tokdemir, O.B.
Artificial Intelligence (AI)-Based
plified building thermal response models and inexact occupancy sensors of traditional building
Occupant-Centric Heating automation systems, investigations into a more efficient and effective sensor-free control mechanism
Ventilation and Air Conditioning have remained entirely inadequate. This study aims to develop an artificial intelligence (AI)-based
(HVAC) Control System for occupant-centric HVAC control mechanism for cooling that continually improves its knowledge
Multi-Zone Commercial Buildings. to increase energy efficiency in a multi-zone commercial building. The study is carried out using
Sustainability 2022, 14, 16107. two-year occupancy and environmental conditions data of a shopping mall in Istanbul, Turkey. The
https://doi.org/10.3390/su142316107 research model consists of three steps: prediction of hourly occupancy, development of a new HVAC
Academic Editors: Luis control mechanism, and comparison of the traditional and AI-based control systems via simulation.
Hernández-Callejo, Sergio After determining the attributions for occupancy in the mall, hourly occupancy prediction is made
Nesmachnow and Sara using real data and an artificial neural network (ANN). A sensor-free HVAC control algorithm is
Gallardo Saavedra developed with the help of occupancy data obtained from the previous stage, building characteristics,
and real-time weather forecast information. Finally, a comparison of traditional and AI-based HVAC
Received: 11 October 2022
Accepted: 12 November 2022
control mechanisms is performed using IDA Indoor Climate and Energy (ICE) simulation software.
Published: 2 December 2022 The results show that applying AI for HVAC operation achieves savings of a minimum of 10%
energy consumption while providing a better thermal comfort level to occupants. The findings of
Publisher’s Note: MDPI stays neutral
this study demonstrate that the proposed approach can be a very advantageous tool for sustainable
with regard to jurisdictional claims in
development and also used as a standalone control mechanism as it improves.
published maps and institutional affil-
iations.
Keywords: artificial intelligence (AI); automatic HVAC control; occupant behavior; model predictive
control; energy efficiency

Copyright: © 2022 by the authors.

Licensee MDPI, Basel, Switzerland.
This article is an open access article 1. Introduction
distributed under the terms and Due to high demand and the need for an increasing energy supply, energy efficiency
conditions of the Creative Commons becomes crucial. Restricted energy markets have wide effects in areas ranging from house-
Attribution (CC BY) license (https://
hold budgets to international relations. Thus, due to high energy consumption, buildings
creativecommons.org/licenses/by/
are on the front line of energy efficiency research. Buildings compose approximately 40%
4.0/).

Sustainability 2022, 14, 16107. https://doi.org/10.3390/su142316107 https://www.mdpi.com/journal/sustainability

Sustainability 2022, 14, 16107 2 of 29

of the total global energy consumption (International Energy Agency, 2019 [1]), and al-
most 40% of this goes towards heating, ventilation, and air conditioning (HVAC) systems
(Yang et al., 2014 [2]). Clearly, the development and implementation of efficient building
energy control systems is essential for economic and environmental sustainability. The
HVAC system is a commonly used tool to maintain thermal comfort in buildings. It also
serves as an essential demand-response source for peak load reduction and system-wide
activity stabilization through effective demand-side energy management strategies. Until
today, this energy demand in the buildings has been measured with sensors. Since heat-
ing and cooling in large masses do not occur rapidly, the inability to respond to sudden
changes in occupancy and environmental conditions makes traditional HVAC control
systems energy-inefficient, especially in large commercial buildings.
An HVAC is a dynamic mechanism that includes multiple input and output vari-
ables and is subject to various fluctuations and uncertainties, including occupant be-
havior, external air temperature, humidity, air volume, and regulated air temperature
(Alcalá et al., 2003 [3]; Mirinejad et al., 2008 [4]). These specific features and characteristics
all need to be taken into consideration to operate the HVAC system effectively. Thus, the
research question of this paper is “how HVAC systems can be made efficient in meeting
the sudden changes in demand responses in large commercial buildings by taking into
consideration the occupancy patterns and prediction?” The following section provides a
critical review of the literature with related studies in energy management with HVAC
control systems to establish the setting for the research.

2. Related Studies
2.1. Traditional and Advanced Control Strategies
HVAC control strategies can be examined in general terms under two headings:
traditional control strategies (TCSs) and advanced control strategies (ACSs). This section
presents a review of related studies focusing on ACSs. Different control mechanisms are
examined, and then a limited number of occupancy-based control approaches are discussed.
TCSs generally include sequencing, on-off, process, and proportional-integral-derivative
(PID) controls. Their simple structure, quick response, easy implementation, and low initial
costs are the main advantages of TCSs. They also have many disadvantages, such as low
accuracy, quality, and performance, and (thus) energy efficiency. Furthermore, they do
not interact with the external environment or regulate and adapt to the input variables
accurately, in terms of their setpoints, schedules, and working modes, among others
(Gholamzadehmir et al., 2020 [5]). Thus, the diversity and complexity of variables make it
impossible to create accurate and reliable mathematical HVAC models for TCSs.
ACTs effectively obtain superior results in HVAC applications. These can be divided
into four categories: (i) soft-computing, (ii) hard-computing, (iii) hybrid, and (iv) adaptive-
predictive control strategies.

2.1.1. Soft Computing Strategies

Reinforcement learning (RL), artificial neural network (ANN)-based deep learning,
fuzzy logic (FL), and agent-based controls together comprise the soft-computing control
strategies. As a control mechanism, this enables solutions to more complex problems by
generating more accurate and statistical responses for unclear and uncertain inputs. The
key benefit of fuzzy logic controllers is that no mathematical simulation is needed for
controller design (Mizumoto, 1995 [6]; Mirinejad et al., 2008 [4]; Soyguder et al., 2009 [7]).
The knowledge-based methodology is the fundamental aspect of a fuzzy controller. This
consists of if-then rules, membership functions, and scaling factors constructed based
on expert experience or learning and self-organization methods that do not involve the
system’s mathematical model forms.
Since the human sensation of thermal comfort is subjective, and self-reporting can
vary among occupants and over time, linguistic rules, on which fuzzy logic is based, are
well suited to characterize HVAC systems and thus ideal for increasing thermal comfort
Sustainability 2022, 14, 16107 3 of 29

(Chiou and Lan, 2005 [8]; Mirinejad et al., 2008 [4]). There are two different approaches to
the automation of rule-based construction in fuzzy systems, which can be used for opti-
mizing the fuzzy system parameters (Mirinejad et al., 2012 [9]): one involves evolutionary
techniques and the other soft-computing methods and technologies, such as ANNs.
Soft-computing methods with ANNs can integrate the learning ability of neural
networks with the knowledge representation of fuzzy logic. They are frequently used
when the aim is to decrease the error between the fuzzy system output and the target
value, as characterized by the general term “neurofuzzy system” (Mirinejad et al., 2012 [9]).
ANNs can also be applied to optimize the fuzzy database, including membership functions
and scaling factors in a fuzzy system (Egilegor et al., 1997 [10]; Kruse et al., 1997 [11];
Wu et al., 2011 [12]). Some studies have utilized advanced fuzzy methods to optimize
the function of existing, traditional PID controllers (Malki et al., 1994 [13]; Ying, 1994 [14];
Wu et al., 1996 [15]; Patel and Mohan, 2002 [16]; Li et al., 2005 [17]), while others have
used them more directly in the development of new HVAC control mechanisms (Fanger,
1972 [18]; Alcalá et al., 2003 [3]; Liang and Ru, 2008 [19]; Gacto et al., 2011 [20]; Nowak and
Urbaniak, 2011 [21]).
Together with model predictive control (MPC) algorithms, fuzzy control algorithms
are implemented in the hierarchical framework for HVAC device control (Nowak and
Urbaniak, 2011 [21]). Wei et al. (2017) [22] presented a deep reinforced learning (RL)
method to develop an HVAC system that they found to be energy-efficient compared with
the traditional rule-based approach. Du et al. (2021) [23] presented a model-free deep
RL framework for an optimized control approach for a multi-zone residential building.
This proposed RL model was reported to provide substantial energy savings and 98% less
comfort violation than a rule-based HVAC control strategy

2.1.2. Hard Computing Strategies

Hard-computing control strategies, which include auto-tuning PID control, gain-
scheduling control, self-tuning control, supervisory/optimal control, MPC, and robust
control, benefit from a mathematical/analytical model that needs real input variables to
respond accurately and rapidly. Some important hard-computing control strategy examples
are summarized below, with a focus on MPC applications as these are more important here.
Pasgianos et al. (2003) [24] applied a non-linear feedback approach for climate control
in greenhouses, especially for ventilation, cooling, and moisturizing. A non-linear multi-
input and multi-output model has been used for an air-handling unit (AHU) control
(Moradi et al., 2010 [25]). Robust control was applied to control the temperature in a
multi-zone HVAC mechanism (Al-Assadi et al., 2004 [26]) and to supply air temperature
(Anderson et al., 2008 [27]). Optimal control strategies were used to manage both single
zone heating in buildings (Dong, 2010 [28]) and a multi-zone air conditioning system
(Mossolly et al., 2009 [29]). An adaptive optimal control approach was also employed to
optimize HVAC system control using a genetic algorithm (Yan et al., 2008 [30]).
MPC is an optimization technique that involves the construction of an objective func-
tion and an input sequence considering both specified and forced constraints. Serale et al.
(2018) [31] aimed to describe the problem formulation, applications, and advantages of
an MPC framework for improving building and HVAC energy efficiency. MPC has four
functions in buildings, related to weather, user behavior, grid, and thermal mass. Kusiak
et al. (2011) [32] created a predictive model with a data-mining approach to optimize HVAC
mechanisms using information gathered from an experiment performed at a research facil-
ity. Kusiak et al. (2014) [33] presented an HVAC optimization approach with data-driven
models and an interior-point method. The Poisson and uniform distributions modeled
the uncertainty of occupant behavior, and the internal heating gain was measured with
the stochastic mechanism of the building’s occupancy. The results showed that the future
performance of HVAC was estimated precisely.
Another data-driven approach for optimizing HVAC energy consumption was pro-
posed by Wei et al. (2015) [34]. For this, a quad-objective optimization problem was built
Sustainability 2022, 14, 16107 4 of 29

to balance energy usage and occupancy comfort and solve a modified particle swarm
optimization algorithm. A substantial amount of energy savings was obtained. Biyik et al.
(2015) [35] and Kelman et al. (2013) [36] suggested an MPC solution in a standard commer-
cial building for two traditional HVAC setups to maximize energy efficiency and increase
occupant comfort by using weather forecasting data. The effect of occupants on internal
load prediction and learning from occupant activity is one of the key features of MPC,
which can have a major impact on energy efficiency (Serale et al., 2018 [31]).

2.1.3. Hybrid Strategies

Huang et al. (2015a) [37] carried out a study that proposed a hybrid MPC framework.
This integrated a classical MPC with a neural network feedback linearization method to
reduce the cost and energy of HVAC in commercial buildings. The results indicated that
a significant level of energy-saving could be achieved without compromising thermal
comfort. Garnier et al. (2015) [38] implemented predictive control for a multi-zone HVAC
mechanism in non-residential buildings using EnergyPLus software for the building model
and ANN-based models for the controller’s internal models. This took the predicted
mean vote index as a measure of thermal comfort. Basic scheduling techniques and the
proposed HVAC system using a genetic algorithm for optimization were compared, and
the importance of the predictive approach demonstrated. Barzin et al. (2016) [39] carried
out an experimental study using weather prediction and a price-based control system for
passive solar buildings, with up to 90% energy savings achieved.
Alibabaei et al. (2016) [40] explored a Matlab-TRNSYS co-simulator development
for control of the TRNSYS software, which was previously designed and balanced based
on a real case-study building and used an advanced predictive controller. This study is
important here in terms of the co-simulation application. For various other studies, Afram
and Sharifi (2013) [41] supplied a detailed literature review including control techniques that
focused on the theory, and implementation of MPC approaches for the HVAC mechanism;
Afram et al. (2017) [42] presented another comprehensive MPC review focusing on artificial
neural network applications with a case study involving ANN models built and calibrated
with the on-site data of a residential house. Trčka and Hensen (2010) [43] and Afroz et al.
(2017) [44] presented a critical review of the latest simulation and modeling techniques,
used in HVAC, focusing on their benefits, limitations, implementations, and efficiency.

2.1.4. Adaptive-Predictive Control Strategies

The APCS (adaptive-predicted control strategy) method can be adapted to a con-
trolled system with time-dependent variables through online variation of its control gains.
Huang et al. (2015b) [45] presented an ANN model-based system identification approach to
model multi-zone buildings. This showed the thermal interactions between the zones to be
well captured by the ANN model, incorporating the energy input from mechanical cooling,
ventilation, changes in the weather, and the convective heat transfer between adjacent
zones. Thus, more precise outcomes are obtained than a single-zone model. Javed et al.
(2017) [46] introduced a random neural network (RNN)-based controller on an Internet of
Things (IoT) platform combined with cloud computing to carry out RNN that estimated
the number of occupants inside the area and sent information to the central RNN-based
occupancy calculator placed in the sensor node.
Cardoso et al. (2018) [47] introduced a study of HVAC power-demand forecasting
based on occupant activity. This influences our study in terms of the use of real data
from a research building for estimation. Estimation of HVAC demand plays a vital role in
developing a more efficient HVAC system. Yang et al. (2019) [48] proposed an adaptive,
robust MPC and compared its performance with predictive model controllers. This study
showed that adaptive modeling and robust optimization minimize unsuitable indoor
conditions because of uncertainties. Zhou et al. (2019) [49] developed a non-linear MPC
by MATLAB using production control systems and weather forecasts and reported a
substantial decrease in energy consumption. Finally, Gholamzadehmir et al. (2020) [5]
Sustainability 2022, 14, 16107 5 of 29

presented a review of the adaptive-predictive control strategy for HVAC systems in smart
buildings focusing on advanced control approaches and their effect on buildings according
to energy consumption and cost. This study indicated that although adaptive control
strategies eliminate the shortcomings of model predictive approaches, such as uncertainty
and unpredictable data, a high degree of inconsistency is observed in the literature.

2.2. Occupancy Related Studies

Since the primary focus of our study is the occupancy pattern and prediction, the
following paragraphs look at occupancy-related studies. Erickson et al. (2009) [50] indicate
that a 14% reduction in HVAC energy usage can be provided with occupancy prediction
and usage patterns. They created a wireless camera sensor network for occupancy data and
estimated occupancy with an accuracy of 80%. Erickson and Cerpa (2010) [51] proposed
a strategy for HVAC systems using real-time occupancy monitoring and estimation of
occupancy with a sensor network of cameras, indicating energy savings of up to 20%.
Oldewurtel et al. (2013) [52] developed an MPC framework using occupancy information
to investigate the effect of occupancy patterns to achieve a more energy-efficient HVAC
mechanism. Furthermore, an RFID-based occupancy detection was presented by Li et al.
(2012) [53] to decrease the consumption of the HVAC. The study shows how demand-driven
HVAC operation is efficient by integrating an occupancy detection system.
A clustering-based iterative evaluation algorithm for eliminating when and how
occupants occupy a building was introduced by Yang et al. (2016) [54], who evaluated
energy implications at the building level with building information modeling that provided
the building geometries, HVAC system configurations, and spatial information as inputs
for the computation of possible energy consequences. Capozzoli et al. (2017) [55] applied an
occupancy-related HVAC operation schedule that focused on shifting groups of occupants
with similar activity in the same thermal zone. As a result of the new schedule approach,
HVAC-related energy use decreased by almost 14%.
Another occupant-centric model, the predictive control approach, was developed
by Aftab et al. (2017) [56], who created and applied an occupancy-predictive HVAC
mechanism using real-time occupancy recognition, predicting user activity, and building
thermal simulation. Aftab et al. (2017) [56] focused on a single-zone mosque area whereas
the research in this paper focuses on multi-zone commercial buildings and adopts the use
of AI for the prediction of occupancy activity. With these advancements, the research in the
paper differs from the one by Aftab et al. (2017) [56].
Shi et al. (2017) [57] used a change-point logistic regression model for precise occu-
pancy estimation to create an occupant-centric model predictive algorithm. Their findings
indicated that an HVAC control strategy with real-time occupancy estimation provides
energy-saving and increases building occupant comfort. Peng et al. (2018) [58] found
that 52% energy saving is possible with occupancy prediction-based cooling control us-
ing machine learning in office buildings. A demand-responsive method was developed
based on energy-related occupant activity. Nikdel et al. (2018) [59] estimated the benefits
of occupancy centric HVAC controls in small office buildings based on programmable
thermostats; when compared with no thermostat control, their proposed HVAC control
approach reduced electricity and natural gas use by up to 50% and 87%.
Ahmadi-Karvigh et al. (2019) [60] presented an automation system that continually
learns occupant behavior to help service system control by determining the set of rules
according to the user’s preferences and behaviors. Adaptive automation gave better results
than inquisitive automation in terms of benefits and occupant satisfaction. Pang et al.
(2020) [61] determined the energy efficiency potential of the new HVAC system combined
with occupancy sensing methods. Their study involved an energy simulation with three
different occupancy scenarios, with occupancy presence sensor and occupant counting
sensor providing energy savings in office buildings.
Azuatalam et al. (2020) [62] developed a reinforced learning (RL) framework to opti-
mize and control the HVAC for a whole commercial building. Simulations showed that,
Sustainability 2022, 14, 16107 6 of 29

compared to a handcrafted baseline controller, an energy saving of up to 22% could be

reached. Deng and Chen (2020) [63] developed a smart HVAC control mechanism for
multi-occupant offices using the physiological signals of occupants. They applied an ANN
model to predict indoor conditions and physiological signals, such as clothing level, (wrist)
skin temperature, relative skin humidity, and heart rate. The heating and cooling loads in
interior offices were reduced by 90% and 30%, respectively, following coupling with the
occupancy-based control through lighting sensors and wristband Bluetooth. This study
was vital for our research in terms of its development of occupancy-related HVAC and
direct measurement of occupant comfort level. Jung and Jazizadeh (2019) [64] presented
a structured literature review examining the user-centric operations and human dynam-
ics of HVAC systems. This study focused on occupancy, comfort, and energy savings
aspects. Finally, Jazaeri et al. (2019) [65] analyzed the complex relationships among lo-
cal climates, building characteristics, and occupancy patterns with the annual and peak
HVAC demand of residential buildings. These studies are important for us in terms of
occupancy, but as mentioned before there is no study that predicts occupancy without
real-time detection tools.

2.3. IDA Indoor Climate and Energy (ICE) Software Background

The IDA Indoor Climate and Energy (ICE) simulation software is one of the four
primary building-energy simulation tools used in research (Ryan and Sanquist, 2012 [66])
and one of the twenty main building-energy simulation software packages (Crawley et al.,
2008 [67]). As with many other simulation software packages, this uses building geom-
etry as the foundation for accurate measurements of solar radiation distribution in and
between spaces. The program dynamically measures energy balances while considering
climatic changes and a changing time-step. Heat balance equations are solved by the
program using building geometry, design, HVAC conditions, and internal heat loads. The
effectiveness and validness of the IDA-ICE software are proved in several studies over
recent years (Bring et al., 1999 [68]; Achermann and Zweifel, 2003 [69]; ISO, 2003 [70];
Karlsson et al., 2007 [71]; Loutzenhiser et al., 2009 [72]; Hilliaho et al., 2015 [73]; Salvalai,
2012 [74]; Mazzeo et al., 2015 [75]; Milić et al., 2018 [76]).

3. Aim of the Research

The advanced prediction ability of AI methods can be employed with sensors to deter-
mine occupant behavior, which offers an excellent opportunity to minimize the weakness
of the traditional HVAC systems. The aim of the paper is to develop an AI-based, occupant-
centric HVAC control mechanism that uses actual weather predictions and continually
improves its knowledge to increase energy efficiency in a commercial building. Since the
cooling problem has gained importance in recent years, the focus will be on the cooling
function of the HVAC systems.
The novelty of the work is twofold. Firstly, a new HVAC control algorithm is proposed,
based on forecasted weather and occupancy information to establish a sensor-free mech-
anism. Secondly, an artificial intelligence-based occupancy forecast system is presented,
which considers all parameters (weather information, time indicators, social situations) and
provides year-round usage with accurate prediction. Although there are limited examples
for real-time occupancy detection in multi-zone buildings, any research study involving
occupancy prediction without a camera or sensors has not been performed yet. There is also
no other study that constructs the relationship between occupancy prediction, real-time
weather, and indoor temperature to manage HVAC control via an algorithm. While a
sensor-free algorithm allows both low installation cost and high energy efficiency, AI-based
occupancy forecasting provides a system that improves itself as these data increase to use
control mechanism standalone and obtain better energy savings.
Sustainability 2022, 14, 16107 7 of 29

4. Research Methodology
Paper adopts the design science research (DSR) methodology that facilitates the de-
velopment of innovative solutions for industry and organizations driven by information
(Vaishnavi, et al., 2019 [77]). Its characteristics involve iterative design processes leading
to development of innovative solutions in the problem domain (Wieringa, 2014 [78]). The
DSR methodology integrates both social context and knowledge base technical capability to
achieve the aim of the research (Markus et al., 2002 [79]). Wieringa (2014) [78] described that
there are two types of DSR: These include “problem-oriented research—evaluation research”,
and “solution-oriented—technical research”. The problem-oriented research looks at what
causes/effects a problem has, or how to solve a problem, whereas the solution-oriented
research design and validate a system, or a requirement (Peffers et al., 2006) [80].
With DSR, this paper promotes the adoption of AI-based and occupant-centric HVAC
control systems in commercial buildings to address the research problem around inefficient
energy management of the existing HVAC systems. The DSR features with social context
that are relevant to the paper are given in Table 1 and the overall research methodology is
illustrated in Figure 1.

Table 1. Design science research (DSR) features.

Design Science Research (DSR) Features

AI-based occupant-centric HVAC control
Design science research
system in commercial buildings
Remove inefficiencies in energy management Design an artifact: Development of AI
via HVAC systems algorithm for occupant-centric HVAC
Enhance the prediction ability with AI for the control system
determination of occupant behavior for Research instruments-tools: Iterative design
effective energy management in and development process for knowledge
commercial buildings capture and development
Answering the knowledge questions: How
Development of an innovative artifact to
Sustainability 2022, 14, x FOR PEER REVIEW 8 of 32
does the artifact adopt smart heritage
achieve and enhance the social context.
project principles?

Figure 1. The
Figure 1. The flowchart
flowchart of the design
of the design science
science research
research (DSR)
(DSR) methodology.
methodology.

Thethe
In design science research
information systems (DSR) methodology
(IS) science, the DSRenabled developing
methodology the HVAC
is highly control
preferred to
system for accurate prediction of energy supply in commercial buildings
solve identified organizational problems by developing information technologies. Thethat will serve
pa-
to meet
per human in
is designed needs for energy
accordance with demand.
the DSRThe system developed
methodology. via DSR
The research is innovative
problem domain
with the key beingare
and opportunities theelaborated
embedment byof artificial
means intelligence
of literature thatthrough
review processesthethe occupancy
relevance cy-
cle. The design cycle is evaluated in Section 5 regarding the development of the artifact
(AI-based occupant-centric HVAC control system) that is extended with the test and
demonstration of the proposed artifact in Section 6 through the rigor cycle. This then leads
to the accumulation of findings into the new knowledge base, articulated in Sections 6 and
Sustainability 2022, 14, 16107 8 of 29

and these weather data without the use of sensors. This paper with DSR implementation
brings not only the novelty but also provides important solution to the practice for energy
management in commercial buildings.
In the information systems (IS) science, the DSR methodology is highly preferred to
solve identified organizational problems by developing information technologies. The
paper is designed in accordance with the DSR methodology. The research problem domain
and opportunities are elaborated by means of literature review through the relevance
cycle. The design cycle is evaluated in Section 5 regarding the development of the artifact
(AI-based occupant-centric HVAC control system) that is extended with the test and demon-
stration of the proposed artifact in Section 6 through the rigor cycle. This then leads to the
accumulation of findings into the new knowledge base, articulated in Sections 6 and 7.

5. Design and Development of the AI-Based Occupant-Centric HVAC Control System

The artifact, which includes novelty about the research problem, is created. This
artifact in this paper is the AI-based occupant-centric HVAC control system. Since the
purpose of the study is to reveal energy efficiency potential of the proposed HVAC control
mechanism, energy analysis according to different scenarios constitute the central part of
this section. The research focuses on a specific site to obtain realistic results, using two-year
occupancy and environmental conditions data of a shopping mall in Istanbul. Figure 2
shows the architecture of the system, consisting of three steps: predicting hourly occupancy,
a new HVAC control mechanism, and comparison of the traditional and AI-based control
systems via simulation according to different scenarios.
In the first step, building properties and real occupancy information are collected.
In the second step, after determining the attributions for occupancy in the mall, hourly
occupancy predictions are made using real data and ANNs, and a sensor-free HVAC control
algorithm is developed with the help of occupancy data obtained from the previous stage,
building characteristics, and real-time weather forecast information.
ANN is considered one of the traditional and most used artificial intelligence methods,
and is still one of the most accurate and effective. This enables traditional and AI-based
sensor-free HVAC control mechanism comparison to be performed in the final step, using
IDA-ICE 4.8 software developed by EQUA Simulation AB based in Stockholm, Sweden.

5.1. Building Properties, Occupancy, and Environmental Information

According to the Association of Real Estate and Real Investment Companies of Turkey
(2019) [81], there are currently 454 shopping malls in Turkey; across Europe, there are
more than 9500 malls, with over 1000 in France and more than 1500 in the UK (STATISTA,
2021 [82]). Worldwide, there is a huge number of shopping malls, which makes them a
significant target for energy savings and important in the development of sustainable energy
policy. These buildings tend not to have good energy efficiency strategies because they are
mostly constructed for consumption and entertainment purposes. It is commonplace for
them to use varied and excessive lighting to attract people and make them feel good inside
the building.
Poor heating and cooling settings disrupt the comfort area for people as well as causing
energy inefficiencies. As a complicating factor and unlike office buildings, shopping malls
do not have certain daily occupancy distributions, thus accurately gauging correct heating
and cooling settings is not easy. It is for these reasons that this study takes a shopping
mall as its case study. For more accurate energy analysis, a realistic model of the building
is used that incorporates the real properties of the building elements. Furthermore, the
solar radiation and weather data for the building location are obtained automatically from
IDA-ICE software for energy simulations. Figure 3 shows the model of the building story;
Table 2 shows the properties of the building elements.
Sustainability 2022, 14,
Sustainability 2022, 14, 16107
x FOR PEER REVIEW 9 9of
of 32
29

Figure 2. AI-based occupant-centric HVAC control system design.

Figure 2. AI-based occupant-centric HVAC control system design.

ANN is considered one of the traditional and most used artificial intelligence meth-
ods, and is still one of the most accurate and effective. This enables traditional and AI-
based sensor-free HVAC control mechanism comparison to be performed in the final step,
 malls do not have certain daily occupancy distributions, thus accurately gauging correct
heating and cooling settings is not easy. It is for these reasons that this study takes a shop-
ping mall as its case study. For more accurate energy analysis, a realistic model of the
building is used that incorporates the real properties of the building elements. Further-
more, the solar radiation and weather data for the building location are obtained automat-
Sustainability 2022, 14, 16107 10 of 29
ically from IDA-ICE software for energy simulations. Figure 3 shows the model of the
building story; Table 2 shows the properties of the building elements.

Figure
Figure 3.
3. Sample
Sample 2D
2D drawings
drawings and
and 3D
3D models
models of
of the building story.
the building story.

Table 2. Building components.

Table 2. components.

Construction
Construction Material Layers
Material (from
Layers Outside
(from toInside)
Outside to Inside)
PressPress
brickbrick
and and
supporters—0.088
supporters—0.088 m|Gypsum
m|Gypsumboard—0.02 m|
board—0.02 m|
External wall Light steel wall and XPS Insulation—0.15 m|Gypsum board—0.02m|
External wall Light steel wall and XPS Insulation—0.15 m|Gypsum board—0.02 m|
Plywood
Plywood wallwall panel
panel andand supporters—0.07 m
supporters—0.07 m
Reinforced concrete—1.5 m|Concrete—0.05 m|
Ground floor XPS Thermal Insulation—0.06 m Concrete—0.03 m |
Screed—0.005 m|Floor Covering—0.008 m
Standing seam roof sheet metal|OSB sheet—0.015 m|
Roof
Corrugated steel sheet|Steel roof supporters and XPS insulation—0.12 m
Window Low-e glass double—4 mm + 12 mm argon + 4 mm

5.2. Occupancy Prediction with ANNs

5.2.1. ANN Parameters
Many factors affect the occupancy numbers and distribution of a shopping mall. They
can be divided into two categories: social and environmental. When the collected real
entrance data are examined, temperature, humidity, and weather conditions along with
type and time of a day come to the forefront as significant parameters. Determined as
attributes in the ANN calculation, these parameters are thus:
1. Temperature: This is one of the most critical factors affecting the number of people;
there are fewer visitors to the shopping mall in winter than in summer days;
2. Humidity: This affects the temperature feel; when the humidity in the air is high,
warm moisture stays on people’s skin longer and makes them feel hotter;
3. Weather condition: This also affects the occupant number significantly; on rainy or
snowy days, shopping malls attract fewer visitors;
4. Time indicators: Days are also significant for shopping mall occupancy; on non-
working days, the number of visitors is higher than on working days. In our study,
the days are not separated into working and non-working days, as in some studies,
but each day of the week is included in the calculation separately; furthermore, month
and year information are considered as separate parameters since they are essential
variables in the long-term use of the shopping mall;
Sustainability 2022, 14, 16107 11 of 29

5. Special days: Public (state) and religious holidays significantly affect occupancy;
the number of visitors increases on national holidays and decreases considerably
on religious holidays; furthermore, the first day of religious holidays is separately
considered because there are far fewer visitors on these days than others;
6. Time of day: This is the most critical factor for sudden changes in the number of
people visiting; for example, the occupancy number increases rapidly at the start of
the lunch break and decreases rapidly when it finishes.
Table 3 illustrates the detailed categories, variables, and unit/index of attributes used
in the ANN model; Table 4 shows a sample of the actual data. Furthermore, the histograms
of the temperature, humidity, weather conditions, and occupancy variables are presented
in Figure 4 to show the distribution of the collected data.

Table 3. Detailed category, variable, and unit/index information of attributes.

Category Variables Unit/Index

Environmental Temperature ◦F
Humidity %
1: Fair|2: Partly cloudy|3: Mostly cloudy|4: Light rain
Weather Conditions
5: Rain|6: Heavy rain|7: Fog|8: Snow|9: Thunder
Weekday 1–7 (1: Monday ··· 7: Sunday)
Social and Time Indicators
Month 1–12 (1: January ··· 12: December)
Day 1–31
Year 2017–2018–2019
Time 10–21 (10: 10:00 a.m. ··· 21: 09:00 p.m.)
0: Normal day|1: Public holiday
Day Type 2: First day of religious holidays
3: Other days of religious holidays

Table 4. Real data for ANN (sample).

Occupancy
Day
Month Day Year Time Day Type Temp. ◦ F Hum. % Weather Conditions (Number
of Week
of People)
Sunday 8 18 2019 1000– 1100 Normal 70 83 Partly cloudy 661
Sunday 8 18 2019 1100– 1200 Normal 68 88 Light rain 1346
Sunday 8 18 2019 1200– 1300 Normal 73 78 Partly cloudy 1448
Sunday 8 18 2019 1300– 1400 Normal 72 78 Mostly cloudy 2547
Sunday 8 18 2019 1400– 1500 Normal 77 50 Partly cloudy 2921
Sunday 8 18 2019 1500– 1600 Normal 79 47 Partly cloudy 3353
Sunday 8 18 2019 1600– 1700 Normal 79 47 Partly cloudy 3181
Sunday 8 18 2019 1700– 1800 Normal 77 47 Partly cloudy 2455
Sunday 8 18 2019 1800– 1900 Normal 77 50 Partly cloudy 2339
Sunday 8 18 2019 1900– 2000 Normal 75 50 Partly cloudy 2126
Sunday 8 18 2019 2000– 2100 Normal 72 60 Partly cloudy 1644
Sunday 8 18 2019 2100– 2200 Normal 70 68 Fair 777
Monday 8 19 2019 1000– 1100 Normal 77 54 Mostly cloudy 463
Monday 8 19 2019 1100– 1200 Normal 75 57 Mostly cloudy 906
Monday 8 19 2019 1200– 1300 Normal 77 54 Mostly cloudy 1418
Monday 8 19 2019 1300– 1400 Normal 81 48 Partly cloudy 1690
Monday 8 19 2019 1400– 1500 Normal 81 45 Partly cloudy 1643
Monday 8 19 2019 1500– 1600 Normal 79 51 Partly cloudy 1379
Monday 8 19 2019 1600– 1700 Normal 77 50 Partly cloudy 1547
Monday 8 19 2019 1700– 1800 Normal 79 51 Partly cloudy 1494
Monday 8 19 2019 1800– 1900 Normal 77 54 Partly cloudy 1907
Monday 8 19 2019 1900– 2000 Normal 75 61 Partly cloudy 1806
Monday 8 19 2019 2000– 2100 Normal 72 73 Fair 1496
Monday 8 19 2019 2100– 2200 Normal 70 78 Fair 727
Sustainability 2022, 14, 16107 12 of 29

Table 4. Cont.

Occupancy
Day
Month Day Year Time Day Type Temp. ◦ F Hum. % Weather Conditions (Number
of Week
of People)
.. .. .. .. .. .. .. .. .. ..
. . . . . . . . . .
Friday 8 30 2019 1000– 1100 Pub. Hol. 77 69 Partly cloudy 1269
Friday 8 30 2019 1100– 1200 Pub. Hol. 81 54 Partly cloudy 1406
Friday 8 30 2019 1200– 1300 Pub. Hol. 81 48 Partly cloudy 1738
Friday 8 30 2019 1300– 1400 Pub. Hol. 82 48 Partly cloudy 2562
Friday 8 30 2019 1400– 1500 Pub. Hol. 82 48 Partly cloudy 2601
Friday 8 30 2019 1500– 1600 Pub. Hol. 81 54 Partly cloudy 2990
Friday 8 30 2019 1600– 1700 Pub. Hol. 81 51 Partly cloudy 2518
Friday 8 30 2019 1700– 1800 Pub. Hol. 79 54 Partly cloudy 2428
Friday 8 30 2019 1800– 1900 Pub. Hol. 77 54 Partly cloudy 2701
Friday 8 30 2019 1900– 2000 Pub. Hol. 75 65 Partly cloudy 2262
Sustainability
Friday 2022, 14,
8 x FOR 30
PEER REVIEW
2019 2000– 2100 Pub. Hol. 73 65 Fair 1805 13 of 32
Friday 8 30 2019 2100– 2200 Pub. Hol. 72 69 Fair 818

Figure 4. Histograms of the temperature, humidity,

humidity, weather
weather conditions,
conditions, and
and occupancy
occupancy variables.
variables.

5.2.2. ANN
5.2.2. ANN Models
Models
Due to their strong logic, error tolerance, versatility, and generalization capabilities, AI
Due to their strong logic, error tolerance, versatility, and generalization capabilities,
methods are used in various applications. The ANN, a mathematical model that imitates
AI methods are used in various applications. The ANN, a mathematical model that imi-
tates the biological nervous system, is one of the most widely used types of AI and has
been implemented to solve a variety of practical challenges in many fields of study.
The fundamental biological unit of the nervous system is the neuron, a fundamental
processing factor that receives and integrates signals from other neurons through dendrite
Sustainability 2022, 14, 16107 13 of 29

the biological nervous system, is one of the most widely used types of AI and has been
implemented to solve a variety of practical challenges in many fields of study.
The fundamental biological unit of the nervous system is the neuron, a fundamental
processing factor that receives and integrates signals from other neurons through dendrite
input paths. The neuron generates an output signal along the axon that links to the dendrites
of several other neurons if the combined input signal is sufficiently high. An attempt to
model the behavior of biological neural systems was made that led to the development
of ANNs, in which artificial neurons model the components of a real neuron. An ANN is
thus a set of independently linked processing units that function as parallel-distributed
computing networks.
Unlike traditional computers, which are programmed to perform particular tasks,
ANNs may learn from examples and eliminate the need for complicated mathematical
formulas or costly physical models by acting as (human) brain-like mathematical models.
They are fault-tolerant and can work with noisy data, allowing for quick generalization
of unknown inputs (Wijayasekara et al., 2011 [83]). They also have specific adaptation
Sustainability 2022, 14, x FOR PEER REVIEW 14 of 32
abilities that enable them to solve highly non-linear problems in which finding analyt-
ical formulations that relate the input data to the output data is especially challenging
(Hagan et al., 2014 [84]). Unlike other statistical or parametric approaches, ANNs can
explicitnon-explicit
extract relationships from a massive
relationships fromvolume
a massiveof correlated
volume of data usingdata
correlated the high
usingcomputa-
the high
tional capabilities
computational of currentofcomputers;
capabilities thus, ANNs
current computers; have
thus, ANNsbecome
havea become
prevalent problem-
a prevalent
solving strategy in
problem-solving a diverse
strategy in arange
diverseof study
range areas.
of study areas.
The architecture of ANN models is formed bybylayers
The architecture of ANN models is formed layers with
with complete
complete or or random
random con-
connec-
nections between them. There is a connection between each neuron,
tions between them. There is a connection between each neuron, and information exchange and information ex-
change is performed. The network receives data from the input layer.
is performed. The network receives data from the input layer. The nodes in this layer do The nodes in this
layer
not do not
have anyhave
weightsanyor weights or activation
activation functions,functions, thus
thus it is not it is notcomputing
a neural a neural computing
layer. The
layer. The
hidden hidden
layer layer or intermediate
or intermediate layer includes layer
dataincludes
processingdataandprocessing
computing and computing
steps and the
stepsresponse
final and the final response
to a given to awhich
input, given isinput,
calledwhich is called
the output the(see
layer output layer
Figure 5).(see
TheFigure
ANN
5). TheisANN
model modelusing
developed is developed using Keras
TensorFlow’s TensorFlow’s
API and Keras
the AdamAPI and the Adam
algorithm algorithm
is used to train
is used
the to train
model. the model.
Five-fold Five-fold cross-validation
cross-validation was applied for was appliedthe
splitting fordata
splitting
into the
twodata into
subsets,
two subsets,
namely, namely,
training trainingNinety
and testing. and testing.
percentNinety
of casespercent of cases
were used for were used
training in for
each training
trivial,
and the trivial,
in each remaining andwere utilized to were
the remaining test the modelto
utilized accuracy.
test the All
modelequations are All
accuracy. adopted from
equations
the book Artificial
are adopted Neural
from the bookNetworks byNeural
Artificial Springer US (2021)
Networks [85].
by Springer US (2021) [85].

Figure 5. ANN structure for the study.

Figure 5. ANN structure for the study.

Generally, the net input of a neuron—activation potential 𝐴 — is equivalent to the

product 𝑤 𝑥 , where 𝑤 is the weight of the corresponding connection on the i-th
postsynaptic neuron and 𝑥 is the input signal (Equation (1)). Connection weights can be
considered as storage of the knowledge that underlies the processing. Thus,
Sustainability 2022, 14, 16107 14 of 29

Generally, the net input of a neuron—activation potential Ai —is equivalent to the

product wij x j , where wij is the weight of the corresponding connection on the i-th post-
synaptic neuron and x j is the input signal (Equation (1)). Connection weights can be
considered as storage of the knowledge that underlies the processing. Thus,

Ai = ∑ wij x j − ai (1)
j

where ai is the threshold activation constant of the neuron. An output can only be obtained
by propagating through a specific activation function. After the signal has been thus
propagated, an output can be found thus:
!
yi = ϕ ( Ai ) = ϕ ∑ wij x j − ai (2)
j

where yi is the output of a layer and ϕ(•) is the transfer function.

The sigmoid activation function has been a common activation function for neural
networks for a long time. Its input is converted to a value of between 0.0 and 1.0, with
inputs that are significantly greater than 1.0 being converted to 1.0, and inputs that are
significantly smaller than 0.0 snapped to 0.0. However, due to the vanishing gradient
problem, usage of the sigmoid and hyperbolic tangent activation functions in networks
with many layers is not true. This problem can be overcome by using the rectified linear
activation function, allowing ANN structures to learn faster and increase performance. The
formula of the rectifier or rectified linear unit (ReLU) is as follows:

f ( x ) = x + = max ( x ) (3)

where x is the input to a neuron. This is also known as a “ramp function” and is analogous
to half-wave rectification in electrical engineering. Connection weights are modified by the
ANN model using a suitable learning method during the training phase. The network uses
a learning mode to obtain the desired output by adjusting the weights. This is executed by
introducing input and desired output to the network. The difference between the expected
output and the network’s output is then used to determine the error value. In the training
phase, recalculations are carried out to decrease the error to an acceptable value. Due to
zero occupancy on some days, mean absolute error (MAE) is used to calculate the error
value, thus:
MAE = |yi − ŷi | (4)
where ŷi is the corresponding desired output value. An error of close to zero shows that
the ANN output values match the expected values very well and the network is well-
trained. Backpropagation training is accomplished by assigning random weights to all
nodes. Equation (5) is used to measure the variation quantity of the connection weights:

∆wij (t) = λδi − yi + α∆wij (t − 1) (5)

where the training rate is λ, the momentum coefficient is α, and the error of the i-th output
layer is δi , which is calculated thus:

δi = yi (1 − yi ) MAEi (6)

MAE and mean absolute percentage error (MAPE) are also calculated as indices to
evaluate the performance of the ANN model, thus:

N N
1 yi − ŷi 1
MAPE =
N ∑ yi
=
N ∑ | Relative Errori | (7)
i =1 i =1
Sustainability 2022, 14, 16107 15 of 29

where N is the total number of data sequences.

5.2.3. HVAC Control Scenarios for Energy Simulation

The primary goal of establishing HVAC control scenarios in terms of the level of
development here is to measure the amount of energy to be saved with the proposed
AI-based control approach. Although great progress has been made in air conditioning
systems, a large proportion of commercial buildings have the most traditional type of
control system, which is one that is operated manually by an attendant (janitor or similar)
responsible for turning the system on and off. The most common HVAC control is based
on the measurements of environmental conditions via sensors, generally temperature,
humidity, and pressure sensors. The most serious deficiency of sensors in terms of energy
consumption is the failure to facilitate a quickly responsive control system.
Many shopping malls serve as a lunch-places for people working near the building,
which causes short-term occupancy densities during the lunch-break period. The rise in
temperature due to sudden increase of people density is a slower process and, by the
time this reaches the sensor, the control system responds, and the appropriate ambient
temperature is provided, most people will already have left the building to return to work.
Moreover, traditional building automation systems depend on quite imperfect occupancy
sensors, which retards system responsiveness. Passive infrared and ultrasonic occupancy
sensors, for example, are low-functioning devices for this usage since they are unable to
accurately assess the occupancy condition, especially when people are stationary for an
extended period and have a limited range, which especially affects their effectiveness in
large areas.
AI prediction technology offers significantly more accurate occupancy information
and improved energy efficiency than traditional building automation systems. Accordingly,
our HVAC control mechanism takes predicted occupancy information and the maximum
number of people per day and adjusts its power according to occupancy rate over time.
Furthermore, new schedule algorithms are developed based on occupancy information,
and weather forecasts for the scenarios (S3 and S4) explained below. The on-off status of
the HVAC is determined according to these setpoint schedule algorithms. The maximum
setpoint value is determined as 24 ◦ C for all scenarios since we focus on the summer period
in this study. Finally, four different scenarios—showing a level of development (from
traditional to advanced)—are determined, as follows:
1. S1: The S1 scenario represents the full-powered HVAC at all times.
2. S2: The S2 scenario represents the most common traditional HVAC control mechanism
based on temperature and occupancy sensors, where the HVAC control system is
automatically (de)activated according to the temperature setpoints and temperature
measurements from the sensors. In this scenario, occupancy is measured by CO2
sensors, which record the level of CO2 in the air. If the number of people in a space
exceeds the amount of CO2 allowed, the sensor triggers the HVAC mechanism to
turn on. This type of sensor is more accurate than a standard motion sensor for the
measurement of occupancy.
3. S3: The S3 scenario represents the proposed AI-based HVAC control system, which
uses predicted occupancy numbers as produced by the ANN model. In this scenario,
the HVAC system responds automatically to changes in the occupancy with no lag
time, contrary to sensor-based systems. The control algorithm provides an HVAC
setpoint schedule to control the system according to real weather conditions (as sup-
plied by weather prediction services) and predicted occupant numbers. The existing
sensors can still be used to monitor the real-time indoor temperature, humidity, and
amount of CO2 . If the actual thermal comfort parameters exceed the desired values,
the control system adjusts itself according to the sensors until thermal comfort is
provided. In the energy simulations performed by IDA-ICE, the effect of the real-time
sensors is not used to examine the no-sensor control mechanism. For this reason, in
the illustrations and graphs for S3 and S4, dashed lines are used to show this potential.
 sensors can still be used to monitor the real-time indoor temperature, humidity, and
amount of CO2. If the actual thermal comfort parameters exceed the desired values,
the control system adjusts itself according to the sensors until thermal comfort is pro-
vided. In the energy simulations performed by IDA-ICE, the effect of the real-time
Sustainability 2022, 14, 16107 sensors is not used to examine the no-sensor control mechanism. For this reason, 16 of in
29
the illustrations and graphs for S3 and S4, dashed lines are used to show this poten-
tial.
4.
4. S4:
S4: The
TheS4
S4scenario
scenariorepresents
representsthe
theHVAC
HVACcontrol
controlsystem
systemininthe
theS3S3scenario
scenariowith
witha
pre-cooling ability along with a quick response. The control algorithm provides
a pre-cooling ability along with a quick response. The control algorithm provides pre-
cooling time to control the system according to predicted weather conditions
pre-cooling time to control the system according to predicted weather conditions and and
occupant numbers. All other features are the same as for
occupant numbers. All other features are the same as for S3. S3.
Figure 66 provides
Figure provides basic
basic illustrations
illustrations for
for the
the four
four energy
energy simulation
simulation scenarios.
scenarios.

Figure 6. HVAC control scenarios for energy simulations.

Algorithm 1 and Algorithm 2, shown in below, explain the proposed HVAC control
schedule algorithms in terms of cooling for S3 and S4. The control algorithm takes occu-
pancy prediction results from the ANN analysis and real weather forecast information
from provider websites as inputs (see lines 1–3). In the time intervals when the occu-
pancy volume increases, the HVAC control activates according to the maximum setpoint
(lines 4–6); otherwise, the algorithm checks the forecasted temperature and compares it
with the maximum setpoint value.

Algorithm 1 HVAC Schedule Algorithm of S3 for Cooling

1 train ANN model
2 make day-ahead prediction for occupancy
3 take day-ahead local weather forecast information
4 if occupancyt < occupancyt+1
5 setpointmax ®Ttarget
6 set HVAC setpoint to Ttarget
7 end
8 else
9 if weather forecast temp.t > setpointmax
10 setpointmax ®Ttarget
11 set HVAC setpoint to Ttarget
12 else
13 deactivate the cooling • deactivation
14 end if
15 end if

If the weather forecast temperature for time t is greater than the maximum, the HVAC
control uses the maximum setpoint (lines 8–10); if not, while the HVAC deactivates cooling
automatically for S3 (lines 12 in Algorithm 1), the algorithm checks the occupancy trend of
one hour later for S4. If there is a sudden increase (determined at 250 visitors), it activates
the pre-cooling 30 min before the upward trend begins for S4 (line 15 in Algorithm 2).
Due to fluctuations in the occupancy numbers, sudden changes can cause comfort
limit values to be exceeded, especially in situations where the number of visitors will
increase too much one or two hours later, even if the occupancy trend is downward for
the current time. To prevent this, S4 presents a 30-min pre-cooling. If there is no such
Sustainability 2022, 14, 16107 17 of 29

increase, the HVAC control deactivates cooling (see line 17 in Algorithm 2), just as for the
S3 algorithm.

Sustainability 2022, 14, x FOR PEER REVIEW

Algorithm 2 HVAC Schedule Algorithm of S4 for Cooling 18 of 32

1 train ANN model

2 make day-ahead prediction for occupancy
3 19take day-ahead
end if local weather forecast information
4 if occupancyt < occupancyt+1
5 20 setpoint
end if ®T
max target
6 21 setend HVAC
if setpoint to Ttarget
7 end
8 else
If the weather forecast temperature for time t is greater than the maximum, the
9 if weather forecast temp.t > setpointmax
HVAC control uses the maximum setpoint (lines 8–10); if not, while the HVAC deactivates
10 setpointmax ®Ttarget
cooling
11 automatically
set HVAC setpoint to Ttarget 12 in Algorithm 1), the algorithm checks the occupancy
for S3 (lines
trend
12 of one hour later for S4. If there is a sudden increase (determined at 250 visitors), it
else
activates
13 ifthe pre-cooling
occupancy 30 min t+1
t+2 - occupanc before
> 250the upward trend begins for S4 (line 15 in Algo-
rithm
14 2). deactivate the cooling for first t/2 • deactivation
15 Due to setpoint max ®Ttarget
fluctuations in the occupancy numbers, sudden changes can cause comfort
16 values
limit settoHVAC setpoint to
be exceeded, Ttarget for in
especially t/2 • start
lastsituations wherepre-cooling
the number of visitors will in-
17 else
crease too much one or two hours later, even if the occupancy trend is downward for the
18 deactivate the cooling • deactivation
current time. To prevent this, S4 presents a 30-min pre-cooling. If there is no such increase,
19 end if
the HVAC control deactivates cooling (see line 17 in Algorithm 2), just as for the S3 algo-
20 end if
rithm.
21 end if

6. Demonstration and Evaluation of the AI-Based Occupant-Centric HVAC Control

6. Demonstration and Evaluation of the AI-Based Occupant-Centric HVAC
System System
Control
Inthis
In thisstage,
stage,the
thedesigned
designedandanddeveloped
developedsystem
systemisistested
testedin inrelation
relationto tothe
thescenarios
scenarios
for energy analysis. According to design science research, research can exploit
for energy analysis. According to design science research, research can exploit experimenta- experimen-
tation,
tion, simulation,
simulation, casecase study,
study, proof,
proof, or other
or other activities
activities to demonstrate
to demonstrate the the proposed
proposed solu-
solution
tion to the research problem. Hence, experimentation of the scenarios
to the research problem. Hence, experimentation of the scenarios via simulation using via simulation using
IDA-ICE4.8
IDA-ICE 4.8software
softwareisisperformed
performed forfor
thethe computational
computational energy
energy analysis.
analysis. TheThe modelmodel of
of the
the shopping
shopping mallmall
was was created
created usingusing
RevitRevit software
software and and imported
imported to IDA-ICE
to IDA-ICE in IFC
in IFC for-
format.
mat. different
Four Four different
energy energy analyses
analyses are carried
are carried out according
out according to the to thescenarios.
four four scenarios. For
For each
each scenario,
scenario, the HVAC
the HVAC controlcontrol
systemsystem corresponding
corresponding to thetocharacteristics
the characteristics
of theofscenario
the sce-
isnario is created
created in the in the simulation
simulation software
software using using
macros.macros.
FigureFigure
7 shows 7 shows
the MPC the MPC algo-
algorithm
rithm framework.
framework. Simulations
Simulations are carried
are carried out daily,outwith
daily,
the with
resultsthefor
results for two
two days days ex-
explained in
plained
detail in in
thedetail in section.
results the results section.

Figure7.7.Model
Figure Modelpredictive
predictivecontrol
control(MPC)
(MPC)algorithm
algorithmframework.
framework.

The weather data of the software are used in the energy analysis. It is seen that the
day-ahead predictions give almost the same values as real values. Therefore, the fore-
casted weather data for the algorithm are not used (to avoid repeating the results and
graphics). In scenarios S4 and S5, both the estimated occupancy numbers found as a result
of ANN calculation and the real occupancy numbers are used in different simulations to
Sustainability 2022, 14, 16107 18 of 29

The weather data of the software are used in the energy analysis. It is seen that
the day-ahead predictions give almost the same values as real values. Therefore, the
forecasted weather data for the algorithm are not used (to avoid repeating the results and
graphics). In scenarios S4 and S5, both the estimated occupancy numbers found as a result
of ANN calculation and the real occupancy numbers are used in different simulations
to show how the small difference between the real data and ANN prediction affects the
energy simulation.
The research artifact, the AI-based occupant-centric HVAC control system, is assessed,
and evaluated to conceive how well the developed and demonstrated artifact is considered
as a solution to the research problem. At this stage, research can benefit from surveys,
feedback, and simulations. If the solution rate, which corresponds to the research problem,
or functionality of the solution is not at an acceptable level, the iterative process is performed
by turning stage 2 and 3.

6.1. ANN Results

In addition to the initial network settings (attributes, layer design, training algo-
rithm, etc.), ANN parameters (hidden layer size, number of neurons in the hidden layer,
batch size, which refers to the number of training examples utilized in one iteration, number
of epochs, the number of complete passes through the training dataset, etc.) have a highly
significant influence on the network output during the training and prediction phases.
While a model with too few neurons has poor predictive performance because it cannot
handle a complex model structure, if too many are selected, weak prediction performance
follows as overfit too easily results from a minor fluctuation in the data.
Therefore, it is crucial to test the model’s output with different design parameters.
Different ANN models were trained for this study; as a result of the trials grid search
methodology using the number of neurons in each layer and the number of epochs as
variables by keeping the number of hidden layers constant at 8. Figure 8 shows the
MAPE and R-squared results of the ANN models created using a grid search. Since the
computational times are not too long and do not change much between them, they are not
considered as parameters. The ANN models with 8 neurons in each hidden layer and with
500 epochs (ANN-1), with 8 neurons in each hidden layer and with 750 epochs (ANN-2),
with 16 neurons in each hidden layer and with 250 epochs (ANN-3), and with 16 neurons
in each hidden layer and with 500 epochs (ANN-4) give the best results with overall MAPE
values of 0.1323, 0.1344, 0.1315, and 0.1335, respectively.
Figure 9 shows the learning curves of the different ANN models. It is clear from the loss
curves that training and validation loss values for ANN-1 and ANN-3 (Figures 9a and 10b,
respectively) are in the ideal range for model complexity. However, the distance between
the training loss line and validation loss line gradually increases after a certain point
because of overfitting in the ANN model with 64 neurons in each hidden layer and with
1000 epochs (Figure 9d).
The comparison between actual occupancy numbers and predicted occupancy values
as given by ANN-1 and ANN-2 is illustrated in Figure 10. Although the prediction values
are naturally far from the real values at some peak points, the prediction trend follows the
real numbers in a general fashion. Thus, as the quantity of data increases in the future,
more accurate results will be obtained.
Furthermore, to examine the results in more detail, four days (18, 19, 20, and
30 August 2019) were removed from the training data set and used as prediction values.
Prediction values for these days were obtained using the ANN-1 features because of the
model accuracy and processing time; a comparison with actual occupancy numbers is
shown as a list (Table 5) and graphically (Figure 11).
Additionally, 18 August was a Sunday, while 30 August is a national holiday. These
days are important in examining the ANN algorithm for weekdays, weekends, and special
days. More people are expected to visit the shopping center on weekends and national
holidays than on weekdays.
 and R-squared results of the ANN models created using a grid search. Since the compu-
tational times are not too long and do not change much between them, they are not con-
sidered as parameters. The ANN models with 8 neurons in each hidden layer and with
500 epochs (ANN-1), with 8 neurons in each hidden layer and with 750 epochs (ANN-2),
with 16 neurons in each hidden layer and with 250 epochs (ANN-3), and with 16 neurons
Sustainability 2022, 14, 16107 19 of 29
in each hidden layer and with 500 epochs (ANN-4) give the best results with overall
MAPE values of 0.1323, 0.1344, 0.1315, and 0.1335, respectively.

Sustainability 2022, 14, x FOR PEER REVIEW 20 of 32

Figure 9 shows the learning curves of the different ANN models. It is clear from the
loss curves that training and validation loss values for ANN-1 and ANN-3 (Figure 9a and
10b, respectively) are in the ideal range for model complexity. However, the distance be-
tween the training loss line and validation loss line gradually increases after a certain
point because of overfitting in the ANN model with 64 neurons in each hidden layer and
with
Figure1000
Figure epochs
8.8.Grid
Gridsearch(Figure
search results9d).
results of the
of the ANN models
modelsaccording
accordingtotoMAPE
MAPEand
andR-squared
R-squaredvalues.
values.

Figure9.9.Learning
Figure Learningcurves
curves of
of the different
different ANN
ANNmodels.
models.

The comparison between actual occupancy numbers and predicted occupancy values
as given by ANN-1 and ANN-2 is illustrated in Figure 10. Although the prediction values
are naturally far from the real values at some peak points, the prediction trend follows the
real numbers in a general fashion. Thus, as the quantity of data increases in the future,
more accurate results will be obtained.
Sustainability 2022, 14, 16107
x FOR PEER REVIEW 2120of
of 32
29

Figure
Figure 10.
10. Actual and ANN-predicted occupancy results (ANN-1 and ANN-2).

TableFurthermore, to examine theoccupancy

5. Actual and ANN-predicted results inresults
more (hourly).
detail, four days (18, 19, 20, and 30 Au-
gust 2019) were removed from the training data set and used as prediction values. Predic-
Sunday,tion values for these Monday,
days were obtained using Thursday,
the ANN-1 features because Friday,
of the model
Time 18 August 2019 19 August 2019 29 August 2019 30 August 2019
accuracy and processing time; a comparison with actual occupancy numbers is shown as
Real aPred.
list (Table 5) Real Pred.
and graphically Real
(Figure 11). Pred. Real Pred.
10:00 a.m. 661 963Additionally,
463 18 August881
was a Sunday,642 while 30 August
697 is a national
1269 holiday.769These
11:00 a.m. 1346 days
1480are important
906 in examining
1322 the ANN 1240algorithm 1345
for weekdays, weekends, and
1406 1602 spe-
12:00 p.m. 1448 1594
cial days. More1418 2139
people are expected 2194
to visit the shopping2290 1738
center on weekends and 1980
national
01:00 p.m. 2547 2412
holidays 1690
than on weekdays.1751 1981 1657 2562 2445
02:00 p.m. 2921 3373 1643 1680 1489 1491 2601 3452
03:00 p.m. 3353 3384 1379 1648 1732 1557 2990 2870
Table 5. Actual and ANN-predicted occupancy results (hourly).
04:00 p.m. 3181 3156 1547 1749 1722 1817 2518 2598
05:00 p.m. 2455 Sunday, 2833 1494 Monday, 1787 1622 Thursday, 1565 2428 Friday, 2498
06:00 p.m. 2339 2351 1907 2134 2034 1793 2701 2411
Time
07:00 p.m. 212618 August1892
2019 19 August 1833
1806 2019 29 August 2019
2362 1821 30 August 2019
2262 1930
08:00 p.m. Real
1644 Pred.
1491 Real
1496 Pred.
1519 Real
2102 Pred.
1518 Real
1805 Pred.
1305
09:00
10:00p.m.
a.m. 777661 704963 727463 635
881 887
642 903
697 818
1269 553
769
11:00 a.m. 1346 1480 906 1322 1240 1345 1406 1602
12:00 p.m. 1448 From Table 1418
1594 5 and Figure2139
11, it is clear2194
that the prediction
2290 values1738
show a harmonious
1980
01:00 p.m. 2547 performance
2412 against
1690 time parameters.
1751 Moreover,
1981 although
1657 30 August
2562was a Friday,
2445this
02:00 p.m. 2921 analysis
3373 managed to
1643 approximate
1680 actual values
1489 with an
1491accuracy of
2601about 87% as an
3452
important measure of the success of predictions.
03:00 p.m. 3353 3384 1379 1648 1732 1557 2990 2870
04:00 p.m. 3181 3156 1547 1749 1722 1817 2518 2598
05:00 p.m. 2455 2833 1494 1787 1622 1565 2428 2498
 08:00 p.m. 1644 1491 1496 1519 2102 1518 1805 1305
09:00 p.m. 777 704 727 635 887 903 818 553

From Table 5 and Figure 11, it is clear that the prediction values show a harmonious
performance against time parameters. Moreover, although 30 August was a Friday, this
Sustainability 2022, 14, 16107 analysis managed to approximate actual values with an accuracy of about 87% as an 21 of 29
im-
portant measure of the success of predictions.

Figure 11.
Figure 11. Actual and ANN-predicted
ANN-predicted occupancy
occupancy results
results (hourly).
(hourly).

6.2.
6.2. Energy
Energy Analysis
Analysis Results
Results
The
The energy analyses
energy analyses aim
aim to
to demonstrate
demonstrate the
the effectiveness
effectiveness ofof the
the proposed
proposed HVAC
HVAC
control algorithm and compare it with traditional control systems. For this purpose,
control algorithm and compare it with traditional control systems. For this purpose, indoor
in-
temperature
door temperature results and daily energy consumption values for the four scenarios were
results and daily energy consumption values for the four scenarios were
obtained from the IDA-ICE software. Two days are selected for detailed energy analysis
obtained from the IDA-ICE software. Two days are selected for detailed energy analysis
and comparison of indoor temperatures according to energy simulation scenarios, Monday,
and comparison of indoor temperatures according to energy simulation scenarios, Mon-
29 August 2019, and Saturday, 7 June 2019; these are illustrated in Figures 11 and 12,
day, 29 August 2019, and Saturday, 7 June 2019; these are illustrated in Figures 11 and 12,
respectively.
respectively.
Since the S1 scenario represents the full-powered HVAC at all times, the indoor
Since the S1 scenario represents the full-powered HVAC at all times, the indoor tem-
temperature remains constant with small fluctuations at 24 ◦ C for both 29 August and
perature remains constant with small fluctuations at 24 °C for both 29 August and 7 June
7 June (Figures 12a and 13a, respectively), as expected. When the indoor temperature
results of the S2 scenario, which represents the sensor-based traditional control approach,
are examined for 29 August (Figure 12b), the temperature is found to vary between 23 and
25 ◦ C across wide intervals. This is because basic thermostats allow the temperature to
fluctuate a few degrees from the fixed temperature to reduce the frequency with which
the cooling device is turned on and off. Consequently, it is seen that the HVAC control
mechanism fails to respond to the rapid increase in outdoor temperature and occupancy
numbers between 10 and 11 o’clock, and when the maximum occupancy number is reached,
the indoor temperature values stay outside the comfort limits. Additionally, although the
Sustainability 2022, 14, 16107 22 of 29

Sustainability 2022, 14, x FOR PEER REVIEW 24 of 32

fluctuations for 7 June (Figure 13b) show a similar pattern, low outdoor temperatures cause
the indoor temperatures to return to comfort limit values more quickly.

Figure 12.
Figure 12. Comparison
Comparison of indoor temperatures
of indoor temperatures of
of scenarios
scenarios for
for Monday,
Monday, 29
29 August
August 2019.
2019.
Sustainability 2022,
Sustainability 2022, 14,
14, 16107
x FOR PEER REVIEW 25
23 of 32
of 29

Figure 13.
Figure 13. Comparison
Comparison of
of indoor
indoor temperatures
temperatures of
of scenarios
scenarios for
for Saturday,
Saturday,77June
June2019.
2019.

The S3 scenario represents the energy simulation according according toto our
our new
new control
control ap-
ap-
proach without pre-cooling, which is Algorithm 1, with the HVAC control adjusting
proach without pre-cooling, which is Algorithm 1, with the HVAC control adjusting ac- accord-
ing to thetooccupancy
cording rate. rate.
the occupancy Although therethere
Although was awas
decrease in occupancy
a decrease between
in occupancy 01:0001:00
between p.m.
and
p.m.02:00 p.m. and
and 02:00 p.m.between 04:00 p.m.
and between 04:00 and
p.m.05:00
and p.m.
05:00on 29 August,
p.m. the cooling
on 29 August, status
the cooling
remained on as the
status remained outdoor
on as temperature
the outdoor was higher
temperature than than
was higher the setpoint (Figure
the setpoint 12c).12c).
(Figure At
07:00 p.m.,
At 07:00 as as
p.m., occupancy
occupancy started
startedtotodecrease
decreaseand
andthe
theair
airtemperature
temperature dropped,
dropped, the cool-
cool-
ing went off, and the indoor temperature increased due to the occupancy. As a result of
the
the dramatic
dramatic decrease
decrease inin the
the number
number of of people,
people, this
this increase
increase ended
ended before
before exceeding
exceeding the
the
Sustainability 2022, 14, 16107 24 of 29

comfort level. Furthermore, since the HVAC became operational in response to the rise in
occupancy, the indoor air temperature remained mostly at the comfort level.
The scenario S4 for 29 August (Figure 12d) produces the same result as does S3 because
conditions that would activate the pre-cooling did not arise on this day. Similarly, there
is no difference in the application of the algorithm in the simulation with the estimated
occupancy values on 29 August (Figure 12e,f) because the increase and decrease trends are
captured correctly by ANN. Depending on the difference in occupancy values between real
and predicted, small changes are observed in temperature changes and fluctuations. For
instance, indoor temperatures do not rise as high for the simulations with predicted values
as for the simulation with real values after the cooling is off because the predicted values
are smaller than others for those time intervals.
In the S3 scenario for 7 June (Figure 13c), the cooling is switched off between 11:00 a.m.
and 01:00 p.m. because the outdoor temperature was below the setpoint with a decrease in
the number of people between these hours. An indoor temperature increase is observed to
occur naturally in this period, but the low outdoor temperature prevents this increase from
reaching significant levels.
Likewise, with the increase in the number of people, the cooling becomes active
again from 01:00 p.m. Similar to the scenario for 29 August, cooling is deactivated by
the algorithm in the hours close to the shopping mall closing time. While the actual
occupancy numbers increase, estimated occupancy values decrease between 03:00 p.m.
and 04:00 p.m. However, the S3 scenario simulation with estimated occupancy (Figure 13e)
follows the same cooling status as the simulation with actual occupancy (Figure 13c) since
air temperature is above the setpoint between these hours. This situation is critical to
minimize the failures due to inaccurate estimation by ANN.
The main difference between S3 (Figure 11c,e) and S4 (Figure 11d,f) is that cooling
status is active at 12:30 p.m. for S4. The reason for this is that the pre-cooling algorithm
is activated under suitable conditions at S4. While there is a decrease in both real and
predicted occupancy numbers between 12:00 p.m. and 01:00 p.m., there is an increase of
more than 250 people between 01:00 p.m. and 02:00 p.m.. The algorithm starts the cooling
30 min before this increase in occupancy to prevent comfort disturbances caused by the
rapid increase. As a result of pre-cooling, the indoor temperature falls to the setpoint level
at the beginning of the occupancy increase, contrary to the S3 scenarios. Similar to the
simulations performed for 29 August, the actual and predicted occupancy numbers lead to
slight differences in the simulations.
When the daily energy consumption results are examined for 29 August (Figure 14),
scenario S1 has the greatest consumption with 4090.61 kWh, as expected. Scenario 2 pro-
vides an energy saving of approximately 30% compared to S1; with a consumption value
of 2279.26 kWh; however, scenarios S3 and S4 consume 22% less energy than S2. When
the daily energy consumption results are analyzed for 7 June (Figure 14), the energy con-
sumption trends generally show a similar pattern to that of 29 August. The S2 scenario
uses almost 30% less energy than the S1, while S3 and S4 provide an energy saving of
approximately 10% over S2. The savings presented by the HVAC control scenario are lower
in June than August because delays resulting from the sensor-based approaches at low air
temperatures affect the energy efficiency less.
There are also some minor naturally based differences between the simulations per-
formed according to actual and estimated occupancy numbers because the simulation tool
adjusts the HVAC power depending on the occupancy. Regarding the ANN values, it
is natural to obtain a lower energy consumption because of simulations with predicted
values for 29 August, since the average of the predicted occupancy number, 1485.27, is
lower than the real occupancy average, 1613.57. Similarly, energy consumption values of
simulations with predicted occupancy are greater than simulations with real occupancy
because the average of the predicted occupancy, 1755.32, is greater than the real occupancy
average, 1679.64.
 x FOR PEER REVIEW
Sustainability 2022, 14, 16107 2725of
of 32
29

Figure
Figure 14.
14. Comparison
Comparison of
of energy
energy consumption values of
consumption values of scenarios.
scenarios.

7. Conclusions
There are also some minor naturally based differences between the simulations per-
formedThisaccording to actualan
paper presented and estimated
analysis occupancy
of different numbers
HVAC controlbecause the simulation
approaches according tool
to
adjusts the HVAC power depending on the occupancy. Regarding
their level of development using energy simulations in which ANN features as the focus the ANN values, it of
is
natural
the study to due
obtain a lower
to the needenergy
for theconsumption because
sensor-free control of simulations
mechanism. The with
ANNpredicted
analysis wasval-
ues for 29 August,
performed since
using real the average
occupancy andof the predicted
weather occupancy
information number,
collected for each1485.27,
day andis lower
hour,
than the energy
with real occupancy average,
simulations 1613.57.
performed forSimilarly, energy
four scenarios consumption
using values of simu-
IDA-ICE software.
lations
Thewith
ANN predicted
results occupancy
showed that arethe
greater than simulations
prediction of occupancy with real occupancy
numbers accordingbe- to
cause the average
time intervals of the
could be predicted
calculatedoccupancy,
with almost 1755.32, is greater than
87% accuracy. This the real occupancy
accuracy rate was
achieved1679.64.
average, with a limited dataset, and estimation precision should be expected to increase
with stronger datasets developed over time. Further, the ANN prediction responded
7.
to Conclusions
different parameters, such as special days. This allows the proposed HVAC control
algorithm to be used
This paper year-round,
presented without
an analysis exceptions.
of different HVAC control approaches according to
According to Wong and Li (2010),
their level of development using energy simulations “total energyinuse”
whichis ANN
the top selection
features criterion,
as the focus
followed by “system reliability and stability”, “operating and maintenance
of the study due to the need for the sensor-free control mechanism. The ANN analysis cost”, and
“control of indoor humidity and temperature”. Since our control
was performed using real occupancy and weather information collected for each day and strategy is based on
data, with
hour, not real-time
the energy detection tools,
simulations while it reduces
performed energy consumption,
for four scenarios using IDA-ICE it software.
also very
positively affects reliability and operating cost. Different scenarios
The ANN results showed that the prediction of occupancy numbers according to varying according to
level of development were used to measure the effectiveness of
time intervals could be calculated with almost 87% accuracy. This accuracy rate was our new HVAC control
mechanism. A detailed examination of energy simulation results has revealed that the
achieved with a limited dataset, and estimation precision should be expected to increase
scenarios representing our AI-based occupant-centric control approach (S3 and S4) save
with stronger datasets developed over time. Further, the ANN prediction responded to
a minimum of 10% energy consumption as compared to the traditional sensor-based
different parameters, such as special days. This allows the proposed HVAC control algo-
approach (S2) and a minimum of 35% on those with full-powered HVAC at all times (S1).
rithm to be used year-round, without exceptions.
In the months when the outside temperature is high, these rates reach approximately 20%
According to Wong and Li (2010), “total energy use” is the top selection criterion,
and 40%, respectively, because traditional approaches allow the indoor temperature to
followed by “system reliability and stability”, “operating and maintenance cost”, and
fluctuate excessively, causing an increase in the power consumed for cooling.
“control of indoor humidity and temperature”. Since our control strategy is based on data,
Another significant result is that there were only very slight differences in indoor
not real-time detection tools, while it reduces energy consumption, it also very positively
temperature and energy consumption results between simulations performed with pre-
affects reliability and operating cost. Different scenarios varying according to level of de-
dicted and real occupancy numbers. This shows that using estimated values in the HVAC
velopment were used to measure the effectiveness of our new HVAC control mechanism.
control algorithm does not significantly change the energy consumption or comfort level.
A detailed examination
Manifestly, of energy
the transformation of simulation results hasproposed
control approaches revealed hasthatgreat
the scenarios
potentialrep-for
resenting our
energy savings. AI-based occupant-centric control approach (S3 and S4) save a minimum of
10% energy consumption as compared to the traditional sensor-based approach (S2) and
Sustainability 2022, 14, 16107 26 of 29

A few limitations should be noted. First, the proposed control algorithms (Algorithms 1 and 2)
were not designed with any complexity; the study design was selected with relatively
simple algorithms to show the savings to be made in a simple way. In cases where
occupancy tends to decrease slightly for long periods and the outdoor temperature is
low, for example, the cooling may remain off for a long time, a situation that was not
represented here. In such cases, the occupancy not being very low could cause the interior
temperature to rise (i.e., even though the air temperature is low). To avoid such a situation,
the algorithm can easily be made more complex with the addition of further parameters,
such as occupancy limit and cooling-off time limit.
Second, even though day-ahead weather forecasts mostly make perfect predictions
for the following day, some days might fall outside the acceptable margin of error. Such a
situation could cause a decrease in the comfort level or inefficiency in the energy consump-
tion, albeit only for very limited periods (or very few days). However, and similarly not
considered in this study, existing sensors might be used as an aid tool to measure the real
situation and included in the algorithm (as stated) to prevent both these shortcomings.
As a major condition of the experimental design and thus a third limitation, only the
cooling function of the HVAC was investigated. Regarding further research, therefore, a
control algorithm can also be developed for heating. Then, the method for HVAC control in-
troduced in this study may be applied to the shopping mall by real experimental setup and
the results observed in reality. Furthermore (as indicated), more complex control algorithms
can be developed according to the specific occupancy pattern of the building studied.
Finally, this study differs from others in considering prediction occupancy numbers
with ANN as the main focus in order that significant energy savings can be achieved with
a simple control algorithm. For this reason, the study can be a pioneer in terms of a new
HVAC system with low installation cost and high energy efficiency. This research can
play a major role in guiding the AI-based occupant-centric control tool for sustainable
development, which can be used as a standalone control mechanism as it improves.

Author Contributions: Conceptualization, A.Y. and O.B.T.; methodology, A.Y., K.S.Ś., Y.A., M.K.
and B.K.; software, A.Y. and K.S.Ś.; validation, M.K. and B.K.; formal analysis, Y.A.; investigation,
A.Y.; resources, B.K., Y.E.A. and O.B.T.; data curation, A.Y.; writing—original draft preparation, A.Y.;
writing—review and editing, Y.A.; visualization, A.Y.; supervision, O.B.T.; project administration,
O.B.T. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Some or all data, models, or codes that support the findings of this
study are available from the corresponding author upon reasonable request.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. IEA (International Energy Agency). Energy Efficiency: Buildings. 2019. Available online: https://www.iea.org/topics/
energyefficiency/buildings/# (accessed on 11 October 2022).
2. Yang, L.; Yan, H.; Lam, J.C. Thermal Comfort and Building Energy Consumption Implications—A Review. Appl. Energy 2014, 115,
164–173. [CrossRef]
3. Alcalá, R.; Benítez, J.M.; Casillas, J.; Cordón, O.; Pérez, R. Fuzzy Control of HVAC Systems Optimized by Genetic Algorithms.
Appl. Intell. 2003, 18, 155–177. [CrossRef]
4. Mirinejad, H.; Sadati, S.H.; Ghasemian, M.; Torab, H. Control Techniques in Heating, Ventilating and Air Conditioning Systems.
J. Comput. Sci. 2008, 4, 777–783. [CrossRef]
5. Gholamzadehmir, M.; del Pero, C.; Buffa, S.; Fedrizzi, R.; Aste, N. Adaptive-Predictive Control Strategy for HVAC Systems in
Smart Buildings—A Review. Sustain. Cities Soc. 2020, 63, 102480. [CrossRef]
6. Mizumoto, M. Realization of PID Controls by Fuzzy Control Methods. Fuzzy Sets Syst. 1995, 70, 171–182. [CrossRef]
7. Soyguder, S.; Karakose, M.; Alli, H. Design and Simulation of Self-Tuning PID-Type Fuzzy Adaptive Control for an Expert HVAC
System. Expert Syst. Appl. 2009, 36 Pt 1, 4566–4573. [CrossRef]
Sustainability 2022, 14, 16107 27 of 29

8. Chiou, Y.C.; Lan, L.W. Genetic Fuzzy Logic Controller: An Iterative Evolution Algorithm with New Encoding Method. Fuzzy Sets
Syst. 2005, 152, 617–635. [CrossRef]
9. Mirinejad, H.; Welch, K.C.; Spicer, L. A Review of Intelligent Control Techniques in HVAC Systems. In Proceedings of the 2012
IEEE Energytech, Cleveland, OH, USA, 29–31 May 2012. [CrossRef]
10. Egilegor, B.; Uribe, J.P.; Arregi, G.; Pradilla, E.; Susperregi, L. A Fuzzy Control Adapted by a Neural Network to Maintain a
Dwelling within Thermal Comfort. Proc. Build. Simul. 1997, 97, 87–94.
11. Kruse, R.; Klawonn, F.; Nauck, D. Learning from Fuzzy Rules. Inform. Forsch. Und Entwickl. 1997, 12, 2–6. [CrossRef]
12. Wu, Y.; Zhang, B.; Lu, J.; Du, K.-L. Fuzzy Logic and Neuro-Fuzzy Systems: A Systematic Introduction. Int. J. Artif. Intell. Expert
Syst. 2011, 2, 47–80.
13. Malki, H.A.; Li, H.; Chen, G. New Design and Stability Analysis of Fuzzy Proportional-Derivative Control Systems. IEEE Trans.
Fuzzy Syst. 1994, 2, 245–254. [CrossRef]
14. Ying, H. Practical Design of Non-linear Fuzzy Controllers with Stability Analysis for Regulating Processes with Unknown
Mathematical Models. Automatica 1994, 30, 1185–1195. [CrossRef]
15. Wu, Z.Q.; Mizumoto, M. PID Type Fuzzy Controller and Parameters Adaptive Method. Fuzzy Sets Syst. 1996, 78, 23–35. [CrossRef]
16. Patel, A.V.; Mohan, B.M. Analytical Structures and Analysis of the Simplest Fuzzy PI Controllers. Automatica 2002, 38, 981–993.
[CrossRef]
17. Li, H.X.; Zhang, L.; Cai, K.Y.; Chen, G. An Improved Robust Fuzzy-PID Controller with Optimal Fuzzy Reasoning. IEEE Trans.
Syst. Man Cybern. Part B: Cybern. 2005, 35, 1283–1294. [CrossRef]
18. Ole Fanger, P. Thermal Comfort: Analysis and Applications in Environmental Engineering. Appl. Ergon. 1972, 3, 181. [CrossRef]
19. Liang, J.; Du, R. Design of Intelligent Comfort Control System with Human Learning and Minimum Power Control Strategies.
Energy Convers. Manag. 2008, 49, 517–528. [CrossRef]
20. Gacto, M.J.; Alcalá, R.; Herrera, F. Evolutionary Multi-Objective Algorithm to Effectively Improve the Performance of the Classic
Tuning of Fuzzy Logic Controllers for a Heating, Ventilating and Air Conditioning System. In Proceedings of the IEEE SSCI
2011: Symposium Series on Computational Intelligence—GEFS 2011: 2011 IEEE 5th International Workshop on Genetic and
Evolutionary Fuzzy Systems, Paris, France, 11–15 April 2011; pp. 73–80. [CrossRef]
21. Nowak, M.; Urbaniak, A. Utilization of Intelligent Control Algorithms for Thermal Comfort Optimization and Energy Sav-
ing. In Proceedings of the 2011 12th International Carpathian Control Conference, ICCC, Velke Karlovice, Czech Republic,
25–28 May 2011; pp. 270–274. [CrossRef]
22. Wei, T.; Wang, Y.; Zhu, Q. Deep Reinforcement Learning for Building HVAC Control. In Proceedings of the 54th Annual Design
Automation Conference, Austin, TX, USA, 18–22 June 2017; Volume 12828. [CrossRef]
23. Du, Y.; Zandi, H.; Kotevska, O.; Kurte, K.; Munk, J.; Amasyali, K.; Mckee, E.; Li, F. Intelligent Multi-Zone Residential HVAC
Control Strategy Based on Deep Reinforcement Learning. Appl. Energy 2021, 281, 116117. [CrossRef]
24. Pasgianos, G.D.; Arvanitis, K.G.; Polycarpou, P.; Sigrimis, N. A Non-linear Feedback Technique for Greenhouse Environmental
Control. Comput. Electron. Agric. 2003, 40, 153–177. [CrossRef]
25. Moradi, H.; Saffar-Avval, M.; Bakhtiari-Nejad, F. Non-linear Multivariable Control and Performance Analysis of an Air-Handling
Unit. Energy Build. 2011, 43, 805–813. [CrossRef]
26. Al-Assadi, S.A.K.; Patel, R.V.; Zaheer-Uddin, M.; Verma, M.S.; Breitinger, J. Robust Decentralized Control of HVAC Systems
Using H ∞-Performance Measures. J. Frankl. Inst. 2004, 341, 543–567. [CrossRef]
27. Anderson, M.; Buehner, M.; Young, P.; Hittle, D.; Anderson, C.; Tu, J.; Hodgson, D. MIMO Robust Control for HVAC Systems.
IEEE Trans. Control Syst. Technol. 2008, 16, 475–483. [CrossRef]
28. Dong, B. Non-Linear Optimal Controller Design for Building HVAC Systems. In Proceedings of the IEEE International Conference
on Control Applications, Yokohama, Japan, 8–10 September 2010; pp. 210–215. [CrossRef]
29. Mossolly, M.; Ghali, K.; Ghaddar, N. Optimal Control Strategy for a Multi-Zone Air Conditioning System Using a Genetic
Algorithm. Energy 2009, 34, 58–66. [CrossRef]
30. Yan, Y.; Zhou, J.; Lin, Y.; Yang, W.; Wang, P.; Zhang, G. Adaptive Optimal Control Model for Building Cooling and Heating
Sources. Energy Build. 2008, 40, 1394–1401. [CrossRef]
31. Serale, G.; Fiorentini, M.; Capozzoli, A.; Bernardini, D.; Bemporad, A. Model Predictive Control (MPC) for Enhancing Building
and HVAC System Energy Efficiency: Problem Formulation, Applications and Opportunities. Energies 2018, 11, 631. [CrossRef]
32. Kusiak, A.; Tang, F.; Xu, G. Multi-Objective Optimization of HVAC System with an Evolutionary Computation Algorithm. Energy
2011, 36, 2440–2449. [CrossRef]
33. Kusiak, A.; Xu, G.; Zhang, Z. Minimization of Energy Consumption in HVAC Systems with Data-Driven Models and an
Interior-Point Method. Energy Convers. Manag. 2014, 85, 146–153. [CrossRef]
34. Wei, X.; Kusiak, A.; Li, M.; Tang, F.; Zeng, Y. Multi-Objective Optimization of the HVAC (Heating, Ventilation, and Air
Conditioning) System Performance. Energy 2015, 83, 294–306. [CrossRef]
35. Biyik, E.; Brooks, J.D.; Sehgal, H.; Shah, J.; Gency, S. Cloud-Based Model Predictive Building Thermostatic Controls of Commercial
Buildings: Algorithm and Implementation. In Proceedings of the American Control Conference, Chicago, IL, USA, 1–3 July 2015;
pp. 1683–1688. [CrossRef]
36. Kelman, A.; Ma, Y.; Borrelli, F. Analysis of Local Optima in Predictive Control for Energy Efficient Buildings. J. Build. Perform.
Simul. 2013, 6, 236–255. [CrossRef]
Sustainability 2022, 14, 16107 28 of 29

37. Huang, H.; Chen, L.; Hu, E. A New Model Predictive Control Scheme for Energy and Cost Savings in Commercial Buildings:
An Airport Terminal Building Case Study. Build. Environ. 2015, 89, 203–216. [CrossRef]
38. Garnier, A.; Eynard, J.; Caussanel, M.; Grieu, S. Predictive Control of Multizone Heating, Ventilation and Air-Conditioning
Systems in Non-Residential Buildings. Appl. Soft Comput. J. 2015, 37, 847–862. [CrossRef]
39. Barzin, R.; Chen, J.J.J.; Young, B.R.; Farid, M.M. Application of Weather Forecast in Conjunction with Price-Based Method for
PCM Solar Passive Buildings—An Experimental Study. Appl. Energy 2016, 163, 9–18. [CrossRef]
40. Alibabaei, N.; Fung, A.S.; Raahemifar, K. Development of Matlab-TRNSYS Co-Simulator for Applying Predictive Strategy
Planning Models on Residential House HVAC System. Energy Build. 2016, 128, 81–98. [CrossRef]
41. Afram, A.; Janabi-Sharifi, F. Theory and Applications of HVAC Control Systems—A Review of Model Predictive Control (MPC).
Build. Environ. 2014, 72, 343–355. [CrossRef]
42. Afram, A.; Janabi-Sharifi, F.; Fung, A.S.; Raahemifar, K. Artificial Neural Network (ANN) Based Model Predictive Control (MPC)
and Optimization of HVAC Systems: A State-of-the-Art Review and Case Study of a Residential HVAC System. Energy Build.
2017, 141, 96–113. [CrossRef]
43. Trčka, M.; Hensen, J.L.M. Overview of HVAC System Simulation. Autom. Constr. 2010, 19, 93–99. [CrossRef]
44. Afroz, Z.; Shafiullah, G.M.; Urmee, T.; Higgins, G. Modeling Techniques Used in Building HVAC Control Systems: A Review.
Renew. Sustain. Energy Rev. 2018, 83, 64–84. [CrossRef]
45. Huang, H.; Chen, L.; Hu, E. A Neural Network-Based Multi-Zone Modelling Approach for Predictive Control System Design in
Commercial Buildings. Energy Build. 2015, 97, 86–97. [CrossRef]
46. Javed, A.; Larijani, H.; Ahmadinia, A.; Emmanuel, R.; Mannion, M.; Gibson, D. Design and Implementation of a Cloud Enabled
Random Neural Network-Based Decentralized Smart Controller with Intelligent Sensor Nodes for HVAC. IEEE Internet Things J.
2017, 4, 393–403. [CrossRef]
47. Sala-Cardoso, E.; Delgado-Prieto, M.; Kampouropoulos, K.; Romeral, L. Activity-Aware HVAC Power Demand Forecasting.
Energy Build. 2018, 170, 15–24. [CrossRef]
48. Yang, S.; Wan, M.P.; Chen, W.; Ng, B.F.; Zhai, D. An Adaptive Robust Model Predictive Control for Indoor Climate Optimization
and Uncertainties Handling in Buildings. Build. Environ. 2019, 163, 106326. [CrossRef]
49. Zhou, B.; Chikkala, J.; Schmitt, R. A Load-Adaptive and Predictive Control of Energy-Efficient Building Automation in Production
Environment. Procedia CIRP 2019, 79, 245–250. [CrossRef]
50. Erickson, V.L.; Lin, Y.; Kamthe, A.; Brahme, R.; Surana, A.; Cerpa, A.E.; Sohn, M.D.; Narayanan, S. Energy Efficient Building
Environment Control Strategies Using Real-Time Occupancy Measurements. In Proceedings of the BUILDSYS 2009—1st ACM
Workshop on Embedded Sensing Systems for Energy-Efficiency in Buildings, Berkeley CA, USA, 3 November 2009; pp. 19–24.
[CrossRef]
51. Erickson, V.L.; Cerpa, A.E. Occupancy Based Demand Response HVAC Control Strategy. In Proceedings of the BuildSys’10—2nd
ACM Workshop on Embedded Sensing Systems for Energy-Efficiency in Buildings, Zurich, Switzerland, 2 November 2010;
pp. 7–12. [CrossRef]
52. Oldewurtel, F.; Sturzenegger, D.; Morari, M. Importance of Occupancy Information for Building Climate Control. Appl. Energy
2013, 101, 521–532. [CrossRef]
53. Li, N.; Calis, G.; Becerik-Gerber, B. Measuring and Monitoring Occupancy with an RFID Based System for Demand-Driven HVAC
Operations. Autom. Constr. 2012, 24, 89–99. [CrossRef]
54. Yang, Z.; Ghahramani, A.; Becerik-Gerber, B. Building Occupancy Diversity and HVAC (Heating, Ventilation, and Air Condition-
ing) System Energy Efficiency. Energy 2016, 109, 641–649. [CrossRef]
55. Capozzoli, A.; Piscitelli, M.S.; Gorrino, A.; Ballarini, I.; Corrado, V. Data Analytics for Occupancy Pattern Learning to Reduce the
Energy Consumption of HVAC Systems in Office Buildings. Sustain. Cities Soc. 2017, 35, 191–208. [CrossRef]
56. Aftab, M.; Chen, C.; Chau, C.K.; Rahwan, T. Automatic HVAC Control with Real-Time Occupancy Recognition and Simulation-
Guided Model Predictive Control in Low-Cost Embedded System. Energy Build. 2017, 154, 141–156. [CrossRef]
57. Shi, J.; Yu, N.; Yao, W. Energy Efficient Building HVAC Control Algorithm with Real-Time Occupancy Prediction. Energy Procedia
2017, 111, 267–276. [CrossRef]
58. Peng, Y.; Rysanek, A.; Nagy, Z.; Schlüter, A. Using Machine Learning Techniques for Occupancy-Prediction-Based Cooling
Control in Office Buildings. Appl. Energy 2018, 211, 1343–1358. [CrossRef]
59. Nikdel, L.; Janoyan, K.; Bird, S.D.; Powers, S.E. Multiple Perspectives of the Value of Occupancy-Based HVAC Control Systems.
Build. Environ. 2018, 129, 15–25. [CrossRef]
60. Ahmadi-Karvigh, S.; Becerik-Gerber, B.; Soibelman, L. Intelligent Adaptive Automation: A Framework for an Activity-Driven
and User-Centered Building Automation. Energy Build. 2019, 188–189, 184–199. [CrossRef]
61. Pang, Z.; Chen, Y.; Zhang, J.; O’Neill, Z.; Cheng, H.; Dong, B. Nationwide HVAC Energy-Saving Potential Quantification for
Office Buildings with Occupant-Centric Controls in Various Climates. Appl. Energy 2020, 279, 115727. [CrossRef]
62. Azuatalam, D.; Lee, W.-L.; de Nijs, F.; Liebman, A. Reinforcement Learning for Whole-Building HVAC Control and Demand
Response. Energy AI 2020, 2, 100020. [CrossRef]
63. Deng, Z.; Chen, Q. Development and Validation of a Smart HVAC Control System for Multi-Occupant Offices by Using Occupants’
Physiological Signals from Wristband. Energy Build. 2020, 214, 109872. [CrossRef]
Sustainability 2022, 14, 16107 29 of 29

64. Jung, W.; Jazizadeh, F. Human-in-the-Loop HVAC Operations: A Quantitative Review on Occupancy, Comfort, and Energy-
Efficiency Dimensions. Appl. Energy 2019, 239, 1471–1508. [CrossRef]
65. Jazaeri, J.; Gordon, R.L.; Alpcan, T. Influence of Building Envelopes, Climates, and Occupancy Patterns on Residential HVAC
Demand. J. Build. Eng. 2019, 22, 33–47. [CrossRef]
66. Ryan, E.M.; Sanquist, T.F. Validation of Building Energy Modeling Tools under Idealized and Realistic Conditions. Energy Build.
2012, 47, 375–382. [CrossRef]
67. Crawley, D.B.; Hand, J.W.; Kummert, M.; Griffith, B.T. Contrasting the Capabilities of Building Energy Performance Simulation
Programs. Build. Environ. 2008, 43, 661–673. [CrossRef]
68. Bring, A.; Sahlin, P.; Vuolle, M. Models for Building Indoor Climate and Energy Simulation, A Report of Task 22 Building Energy
Analysis Tools. Report of IEA SHC Task. 1999. Available online: https://www.equa.se/dncenter/T22Brep.pdf (accessed on
10 October 2022).
69. Achermann, M.; Zweifel, G. RADTEST—Radiant Heating and Cooling Test Cases. 2003. Available online: http://www.
equaonline.com/iceuser/validation/old_stuff/RADTEST_final.pdf (accessed on 10 October 2022).
70. ISO (International Organization for Standardization). ISO 15099:2003. Thermal Performance of Windows, Doors and
Shading Devices—Detailed Calculations. 2003. Available online: https://www.iso.org/standard/26425.html (accessed on
10 October 2022).
71. Karlsson, F.; Rohdin, P.; Persson, M.L. Measured and Predicted Energy Demand of a Low Energy Building: Important Aspects
When Using Building Energy Simulation. Build. Serv. Eng. Res. Technol. 2007, 28, 223–235. [CrossRef]
72. Loutzenhiser, P.G.; Manz, H.; Moosberger, S.; Maxwell, G.M. An Empirical Validation of Window Solar Gain Models and the
Associated Interactions. Int. J. Therm. Sci. 2009, 48, 85–95. [CrossRef]
73. Hilliaho, K.; Lahdensivu, J.; Vinha, J. Glazed Space Thermal Simulation with IDA-ICE 4.61 Software—Suitability Analysis with
Case Study. Energy Build. 2015, 89, 132–141. [CrossRef]
74. Salvalai, G. Implementation and Validation of Simplified Heat Pump Model in IDA-ICE Energy Simulation Environment. Energy
Build. 2012, 49, 132–141. [CrossRef]
75. Mazzeo, D.; Matera, N.; Cornaro, C.; Oliveti, G.; Romagnoni, P.; de Santoli, L. EnergyPlus, IDA ICE and TRNSYS Predictive
Simulation Accuracy for Building Thermal Behaviour Evaluation by Using an Experimental Campaign in Solar Test Boxes with
and without a PCM Module. Energy Build. 2020, 212, 109812. [CrossRef]
76. Milić, V.; Ekelöw, K.; Moshfegh, B. On the Performance of LCC Optimization Software OPERA-MILP by Comparison with
Building Energy Simulation Software IDA ICE. In Build. Environ.; 2018; Volume 128, pp. 305–319. [CrossRef]
77. Design Science Research in Information Systems; Vaishnavi, V.; Kuechler, W.; Petter, S. (Eds.) Association for Information Systems:
Atlanta, GA, USA, 2019.
78. Wieringa, R.J. Design Science Methodology: For Information Systems and Software Engineering; Springer: Berlin/Heidelberg, Ger-
many, 2014. [CrossRef]
79. Markus, M.L.; Majchrzak, A.; Gasser, L. A design theory for systems that support emergent knowledge processes. 2002, MIS
Quarterly 26, 179–212. MIS Q. 2002, 26, 179–212.
80. Peffers, K.; Tuunanen, T.; Gengler, C.; Rossi, M.; Hui, W.; Wirtanen, V.; Bragge, J. The design science research process: A model for
producing and presenting information systems research. In Proceedings of the First International Conference on Design Science
Research in Information Systems and Technology DESRIST, Claremont, CA, USA, 24–25 February 2006.
81. The Association of Real Estate and Real Estate Investment Companies of Turkey (GYODER). 2019. Available online: https:
//www.gyoder.org.tr/yayinlar/sektorel-yayinlar (accessed on 1 May 2021).
82. STATISTA 2021 (Number of Shopping Centers in Europe 2017, by Country). Available online: https://www.statista.com/
statistics/912126/shopping-center-numbers-by-country-europe/ (accessed on 10 October 2022).
83. Wijayasekara, D.; Manic, M.; Sabharwall, P.; Utgikar, V. Optimal Artificial Neural Network Architecture Selection for Performance
Prediction of Compact Heat Exchanger with the EBaLM-OTR Technique. Nucl. Eng. Des. 2011, 241, 2549–2557. [CrossRef]
84. Hagan, M.T.; Demuth, H.B.; Beale, M.H.; de Jesus, O. Neural Network Design. In Neural Networks in a Softcomputing Framework,
2nd ed.; Springer Science & Business Media: Berlin/Heidelberg, Germany, 2014.
85. Cartwright, H. (Ed.) Artificial Neural Networks; Springer: New York, NY, USA, 2021.