INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING

2021, VOL. 34, NO. 9, 933–949

https://doi.org/10.1080/0951192X.2021.1946857

An approach to develop a digital twin for industry 4.0 systems: manufacturing

automation case studies
David Guerra-Zubiagaa, Vladimir Kutsb, Kashif Mahmoodb, Alex Bondara, Navid Nasajpour-Esfahania
and Tauno Ottob
a
Department of Robotics and Mechatronics Engineering, Kennesaw State University, Marietta, USA; bTallinn University of Technology,
Department of Mechanical and Industrial Engineering, Tallinn Estonia

ABSTRACT ARTICLE HISTORY

The new paradigm of digital manufacturing and the concept of Industry 4.0 has led to the Received 28 July 2019
integration of recent manufacturing advances with modern information and communication Accepted 19 May 2021
technologies. Therefore, digital simulation tools fused into production systems can improve time KEYWORDS
and cost-effectiveness and enable faster, more flexible, and more efficient processes to produce Digital Manufacturing;
higher-quality goods. The advancement of digital simulation with sensory data may support the Digital Twin; Simulation;
credibility of production systems and improve the efficiency of production planning and execution Automated Production
processes. In this paper, an approach is proposed to develop a Digital Twin of production systems Systems; Industry 4.0
in order to optimize the planning and commissioning process. The proposed virtual cell interacts
with the physical system with the help of different Digital Manufacturing Tools (DMT), which allows
for the testing of various programs in a different scenario to check for any shortcomings before it is
implemented on the physical system. Case studies from the different production systems are
demonstrated to realize the feasibility of the proposed approach.

1. INTRODUCTION
There have been many challenges in industry 4.0 to can advance within virtual environments. This new
become a competitor for the industrial paradigm trend is referred to as Digital Manufacturing. It is
(Guerra-Zubiaga 2020a). However, in contemporary defined as ‘the use of an integrated, computer-
society, technologies are increasingly shifting to the based system that consists of simulation, 3D visualiza­
virtual world. As humanity moves towards total reu­ tion, analytics and collaboration tools to create pro­
sability and environmental friendliness, vehicles duct and manufacturing process definitions
become electric; rockets are designed to fly dozens simultaneously’ (Siemens PLM (2018).
of times; materials are recycled as much as possible. It In a manufacturing setting, a simulation process is
is, therefore, logical for manufacturing itself to essential because it reduces time and expenditures
become flexible and rapidly reconfigurable. One when implementing a new manufacturing process
does not build a new factory to produce each pro­ (Mahmood 2017). By eliminating the traditional
duct, each being different in shape and size. Instead, method of pen and paper planning, engineers can
new and upcoming products should be fabricated now precisely calculate every movement and make
consistently to align with the production line require­ plans for more efficient manufacturing. In addition to
ments they are assigned for. Therefore, production developing the simulation, the implementation for
lines of the future must not only be easily reconfigur­ creating the interconnectivity between the simulation
able, but they must also have the capabilities to and the controls within the manufacturing setting
update themselves when required. This may be provides significant opportunity towards cost mitiga­
achieved through the essence of digital manufactur­ tion along with a visual representation of the produc­
ing and the deployment of digital simulation tools tion process. In addition to the monetary savings,
into manufacturing systems. All current manufactur­ safety is increased by the adoption of simulation pro­
ing processes and the development of new products cesses. Knowing where all cautions and dangers lie

CONTACT David Guerra-Zubiaga [email protected] Department of Robotics and Mechatronics Engineering, Kennesaw State University,
Marietta, USA
© 2021 Informa UK Limited, trading as Taylor & Francis Group
934 D. GUERRA-ZUBIAGA ET AL.

before commissioning, engineers may now provide the implementation of new automation methods and
a safer work environment to the industry. By digitally digital tools. The research by Jackson et al. explores
designing a manufacturing process, more control is automation methods implementing TIA Portal,
given to the engineer who may then reduce many PLCSIM Advanced, and Tecnomatix, among other
risks in the planning process. Visualization of robotics tools. Compared to the state-of-the-art twenty years
and the human connection can be displayed during ago, product life cycles have decreased dramatically,
the simulation. Using this tool helps guide manufac­ and the consumer market demands new, sophisti­
tures with a safety zone around hazard areas on the cated and high-quality products while maintaining
manufacturing floor. low price and short lead times (Paavel (2017)). This
This paper describes a generic approach to forces manufacturers to discover new and more effi­
develop a Digital Twin for a manufacturing system. cient ways of producing goods and offering services.
Two production systems were selected to demon­ Already today, by implementing newer technologies
strate the relevance of the generic approach accord­ and techniques such as virtual commissioning and
ing to the modern digital manufacturing needs, i.e., digital simulation tools, production can be made fas­
a rapidly reconfigurable manufacturing system and, ter and more efficient, resulting in savings for both
ultimately, a Cyber-Physical Production System the companies and their customers in addition to
(CPPS). The motivation for this research is to show decreasing industrial impacts on the environment
the unique approach of Digital Twin and also show (Nikolakis 2019).
the validation of this approach for other works in the For the future perspective, individual products
future. A literature review was carried out to define should be produced quickly, flexibly and sustainably
and understand the fundamental techniques that (Mahmood 2017). Therefore, production lines of the
help to build an approach of a Digital Twin for pro­ future must not only be easily updatable, but they
duction systems. Modeling techniques to construct should also have the capability to self-update. That is
the virtual environment, advancements of the role of where digital manufacturing, Industrial Internet of
automation in a manufacturing system, the impor­ Things (IIOT), and cloud computing integration into
tance of digital manufacturing, and the concept of traditional and flexible manufacturing systems come
a Digital Twin and the roots of CPPS are highlighted. into play (Mahmoodpour (2018)). The digital manu­
The structure of this paper is organized as follows. facturing paradigm requires that each part of
Section 2 covers the most recent and relevant a production line be interconnected via embedded
research papers that include DT as a literature review. systems, i.e. tiny computers, sensors, and actuators
Section 3 highlights our approaches to create a new built into the machinery that forms a universal inter­
platform in NGAS as a methodology. Section 4 net of things (IoT). These systems help automate the
describes our two case studies that one of them was manufacturing process by making it ‘self-organize’,
done in KSU and the other one in TALTECH university. and mitigate constant human intervention while
At the end, the result section provides our result from keeping the process under human control (Wang
both case studies as a one platform and the future 2018). Such flexibility of a manufacturing company is
research is described in the conclusion section to necessary for it to be competitive in the not-so-distant
show our ideas for future work. future.
System modeling – The modeling of a system
defines the process of establishing steps or workflows
2. LITERATURE REVIEW
for a particular system that is selected for the study. In
Due to uncertainties in the market and incremental every discipline of science and engineering, modeling
demand for customized production with low volume, and simulation are principal activities to analyse
industrial companies are driven to enhance the flex­ a system. With the help of relevant modeling tools,
ibility in their manufacturing lines (Mourtzis (2014)). one can perform the analysis of physical systems and
Jackson (2016) explore some digital tools in flexible have a better understanding of the functioning of
manufacturing and specified that recent advance­ a physical world ((Cellier (2013); (Lin (2009)).
ments in industrial technologies may improve the Machine Automation – Machine automation is the
flexibility of production shop floor operations through use of mostly automatic equipment in
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 935

a manufacturing system or another production pro­ layout, material flow, and throughout before pro­
cess (Groover 2014). As technology has advanced, duction ramp-up.
manufacturers created more complex machines to ● Allows part of manufacturing processes to be
automate repetitive everyday tasks. The invention optimized within a managed environment to
of the computer changed the concept of an auto­ produce flexible work instructions capable of
mated machine and defined the understanding of displaying 2D or 3D part information, along
what automation is in the modern world. For the first with machining and tooling instructions.
time in history, humanity invented such ● Reduces commissioning costs through simula­
a sophisticated and powerful tool capable of per­ tion by validating robotic and automation pro­
forming an unprecedented number of operations grams virtually.
over a short period of time. Over the last few dec­
ades, this technology has advanced even further. With consideration of the above points, the usage of
Today, there are computers that can think, create, digitalization tools in novel manufacturing processes
and make decisions (Noble 2011). Moreover, as is crucial to enable the collection of data from digital
a generalization, they are much better and faster at replicas of physical manufacturing systems. Mainly,
this than humans are, which brings a new under­ application of virtual commissioning tools is needed
standing of automation. to avoid failures and errors while integrating new
The contemporary industrialist is no longer satis­ equipment or processes to the shop level of
fied with machines that solely perform redundant a factory. In addition to reducing commissioning
tasks on an assembly line. Instead, they desire them time and allocated resources, data analysis and opti­
to also identify problems, make decisions, and evalu­ mization performed on systems pre-installation with
ate the best approaches to performing complex tasks DMT mitigates downtime and costs when replacing
in non-static environments. Therefore, it is no longer and rearranging existing physical production layouts.
adequate to simply make a machine perform simple, The Digital Twins concept – The Digital Twin is
repetitive tasks. As researchers Sandeep (2015) point a popular topic in the scope of the Industry 4.0 digi­
out, ‘To automate a machine, a secondary brain talization and simulation pillar for the research world
(another machine) is required to “think” and control where advancements towards accurate plant simula­
it to do tasks as per the convenience of the user’. This tions in addition to the field of usage and smart con­
concept of the Digital Twin is implemented in the trol of original assets are being investigated (Grieves
studied manufacturing systems and is demonstrated (2017); Kuts (2019)). NASA was first to introduce vir­
further in the following methodology and case stu­ tual replicas in 1957. However, the primary contribu­
dies sections. tion to the definition of Digital Twins for the new
Importance of Digital Manufacturing – Digital man­ industrial revolution was made by Grieves (2017),
ufacturing is the primary link between Product who introduced the concept of precise replicas of
Lifecycle Management (PLM) software and the actual various manufacturing assets. In contemporary man­
industrial equipment application. This process ufacturing industries, it is a widespread practice to see
enables the reduction of expensive downstream the harmonic collaboration of different simulation
changes as well as the testing of material costs software working in conjunction with Reconfigurable
incurred in traditional systems. Manufacturing Systems (RMS). Digital Twins synthe­
sized with RMS technologies is seen extensively in
Some of the benefits of Digital Manufacturing are as automotive manufacturing. In recent years, BMW
follows (Siemens PLM 2018): Motoren GmbH has started to use Tecnomatix to
reduce costs and to become more energy efficient.
● Enables product, process, plant, and resource With an ever-evolving floor plan to accommodate
information to be associated, viewed, and taken changes in the automotive industry, tools must be
through change processes, with a consistent and reconfigurable to keep up with engineering advances
comprehensive approach to production design. as well as consumer demands. By using Tecnomatix
● Helps to create factory models faster and to
with RMS, BMW can utilize Digital Twins to remain
ensure that they are operating under an optimal competitive in the automotive world.
936 D. GUERRA-ZUBIAGA ET AL.

Digital Twins, however, are not used exclusively in using industry-standard Ethernet protocols. From
the automotive industry but rather represent a typical the software perspective, these direct process con­
application in the field. For example, Zambrano trollers should interact with each other as remote
Degan et al. 2020 presented a 3D manufacturing inputs and outputs (I/O) with no unique setup
design of drilling process by using Tecnomatix soft­ required to achieve system synchronization (Strang
ware. With the new generations of engineers moving (2016); Pilz Australia Industrial Automation LP 2016).
into the manufacturing world, companies from across It is important to review some of the previous
numerous practices have already started to adopt works specifically about different approaches of digi­
various DMT. In addition, experiments on connectivity tal twin in industry 4.0 to understand the new
between real and virtual worlds have been widely approach of this research paper and propose a novel
introduced and tested by Modoni (2016) and Kuts goal for the future. There are recent developments
et al. 2018 in order to feed Digital Twin replicas with and advancements in Digital Twin conceptual models.
data from the real IoT sensors to improve the accuracy For instance, there is the ‘Digital Twin 8-dimension
of the simulation. Digital Twin technology is one of model’ that focuses on the behavior, capabilities, and
the major pillars for smart manufacturing as it allows environmental dimensions of a Digital Twin (Stark
us to analyze the past and predict the future. Future (2019)). The ‘Five-dimension Digital Twin model’
manufacturing can be influenced by the Digital Twin describes and considers the aspects of physical enti­
concept and addresses the different digital aspects ties in Digital Twin, virtual models in Digital Twin,
such as Digital Twins for manufacturing equipment, Digital Twin data, services in Digital Twin, and con­
Digital Twins of humans, Digital Twins for factories, nections in Digital Twin (Qi 2019). Moreover, there is
and Digital Twins for manufacturing networks (Kuts research on the development of CPPS and smart
2018; Lu 2020). manufacturing technologies driven by big data
Overview of Cyber-Physical Production System (Mahmood 2019; Tao 2018). Malik (2018) presented
(CPPS) – The manufacturing-oriented Cyber-Physical a general framework for digital twin by utilizing com­
System (CPS) is called CPPS, which holds the attri­ puter simulations to develop human-robot collabora­
butes of robustness, autonomy, self-organization, tion in an assembly environment. In this study, they
transparency, predictability, efficiency, interoperabil­ used Universal Robot UR-5 and UR-3, which has 6
ity, global tracking, and tracing (Lu 2020; Monostori degrees of freedom and payload capacity of 5 and
2014). The basis of CPS, highlighted by Monostori 3 kg. CAD software is used to create digital objectives
2014 and Lee (2017, is that the physical system repre­ and then converted to Tecnomatix software. The vir­
sentation may be given in the virtual world and vice tual model includes four main steps which are geo­
versa. The cyber-physical systems are an extension of metry, physics, behaviour, and rule. The simulation
embedded systems where sensors are used to record, estimated the time of task completion, and deter­
save, and evaluate data. They interact with the physi­ mined the possible safety issues and desired loca­
cal and virtual world, communicate with each other tions. In other work, IEC61,499 distributed
through interfaces, and use big data and human- automation was used to control cyber-physical com­
machine interfaces (2017); Strang (2016)). ponents of the system in advanced manufacturing
The components of cyber-physical production sys­ operations (Nikolakis (2018). Havard et al. (2019)
tems consist of information that is located in described the new approach for co-simulation archi­
a structured storage to manage manufacturing infor­ tecture between digital twin and virtual reality. They
mation. The information may be exchanged between used CAD (Computer-Aided Design), CAM (Computer-
system components via communication interfaces of Aided Manufacturing) to compose the workflow. On
the CPS and can be adopted for autonomous deci­ the other side, virtual reality includes interaction and
sion-making throughout the production process. The visualization components. The simulation data was
communication or process control in a CPPS can be used in order to improve the system of virtual reality
executed directly through the components, where environment. In this paper, they widely explained
the onboard control system does not include how different components exchanged the data by
a central or supervisory controller. Instead, each direct using C++ program language. The new architecture
process controller interacts with other controllers was developed to create a safe environment for
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 937

welding collaboration between humans and robots ● Identify Smart Manufacturing Systems (SMS) and
by using IEC 61499 Function Blocks (FBs) methods Cyber Manufacturing Systems (CMS) as possible
and MQTT protocol (Garcia 2020). Gervasi (2020) pro­ subcomponents of NGAS.
posed a new framework which can analyze and eval­ ● Explain how DMT connects physical and virtual
uate the interaction between robots and humans prototypes, including its benefits.
especially focused on manufacturing. However, they ● Define a DT concept and goals at NGAS environ­
explained and compared different aspects of human ment exploring new manufacturing paradigms
robot collaboration (HRC) and showed their HRC eva­ such CMS, SMS, and IoT, among others.
luation in other applications instead of manufacturing ● Implement case studies applying the DT concept
too. However, there is still a lack of studies related to at NGAS environment making possible industry
Digital Twins that address the practical impacts of DT, 4.0 applications.
especially by comparing the results from multiple
case studies. Therefore, in this study, the authors (Mourtzis 2019) presented new technologies such as
propose a comprehensive approach to develop Cyber-Physical Systems (CPS), Cloud Computing,
a Digital Twin and validate a new approach through Internet of Things (IoT), and Augmented and Virtual
a series of case studies to explore the practical Reality, which are also called AR and VR, Design of
impacts and limitations of the DT concept at Next Manufacturing Networks, and Hybrid Simulation and
Generation Automated System (NGAS) environment. Digital Twins. In addition, the concept of Cyber-
Physical System has been investigated to improve
productivity and adaptability of Machine Tools
3. METHODOLOGY (Armendia et al. 2019). This can show us that
Industry 4.0, as an important technology trend,
The methodology of this paper is depicted in Figure 1
seeks to improve manufacturing processes according
and is described according to the following main
to new manufacturing paradigms such as IoT, DT,
steps.
CMS, and SMS, among others. Recently, DMT has
been introduced as a platform for visualization and
● Introduction of NGAS platform at Industry 4.0
communication in NGAS (Guerra-Zubiaga 2020b).
revolution.
A difficult task is not only to understand these

Figure 1. An approach for Digital Twin development.

938 D. GUERRA-ZUBIAGA ET AL.

paradigms but also to connect them to develop the transferred by connected systems to the operations
NGAS. This research presents a new approach allow­ (Lee 2016). SMS is defined as a production integra­
ing not only the understanding of the new paradigms tion manufacturing that includes many aspects such
but also guiding other focus groups to develop the as communication technology, data modeling, sen­
NGAS according to Industry 4.0 trends. sors, computing models, simulations, and other pre­
Today, DMT connects the physical and virtual dictive aspects of engineering (Kusiak 2018). So, the
world to make sense between IoT, DT, CMS, and combination CMS and SMS is considered in this
SMS. Digital manufacturing motivates the exploration work to show the power of NGAS. However, the
of NGAS in Industry 4.0 environments through these interconnectivity in SMS is still a problem, and
approaches. If successful, the new approach should Internet-of-Things hub (IIHub) was used to solve
integrate not only people, instruments, controls, and this issue (Tao 2017). Therefore, in this paper the
robots, but will also include the DMT that can simu­ authors implement factory telemetry efforts that
late a new NGAS platform. While different approaches can connect the virtual and the physical prototype
and concepts of digital manufacturing platforms have in real-time. The physical prototype represents the
been investigated (Gerrikagoitia et al. 2019), and real manufacturing system, and the factory teleme­
there is still a need to create a new NGAS platform. try can be obtained by implementing IoT technolo­
This can be a huge achievement for mass-production gies. The obtained information produces the
systems at manufacturing and labor facilities because concept of the DT of the manufacturing system
it increases efficiency, saves time and energy, and going beyond the simple simulation and visualiza­
maximizes profits. tion in the 3D virtual environment. Figure 1 concep­
One of the many misconceptions about a DT is that tualizes this method and aims to explain the use of
it is mistaken as another form of simulation when, in technologies like IoT and DMT to create a DT of
fact, it is not a simulation at all. A DT is more closely a manufacturing system including cyber and smart
compared to virtualization because it can receive manufacturing aspects to support NGAS to be
results from the production line programming and aligned to Industry 4.0 trends. The case studies
manipulate itself to show progress in the same way explored in Section 4 are theoretically extending
a real production line would. Currently, simulations of the understanding of the implementation of this
production lines are widely used but lack the ability to new proposed approach.
test programming without having the adequate hard­ The Kennesaw State University (KSU) case is imple­
ware. Using a DT, the entire process can be run vir­ mented using the methodology described in Figure 1.
tually as a test, and when it appears to be running KSU’s example implements an advanced manufactur­
how it is supposed to, the program can be transferred ing automated system using Siemens technologies
to the production line for normal use. with digital virtual commissioning aspects.
A case study is used as the primary research According to Figure 1, the authors propose this as
method to describe and realize the DT application to NGAS. Smart manufacturing in automation systems
production systems. The studies connect part of could deal with the utilization of instruments and
a system to a DT to demonstrate its functionality controls to anticipate events. In the KSU case study,
and usefulness in industry. An Open Platform safety instrumentation is highlighted to mitigate
Communications (OPC) server provides the connec­ workplace hazards. For example, a camera is installed
tion details for the implementation of these case to detect the presence of a human and stop the
studies. The Digital Twin conceptual model, also subject system from advancing.
termed as an approach for Digital Twin development, The Tallinn University of Technology (TalTech) case
is illustrated in Figure 1. study 2 is implemented based on the same metho­
In this approach, NGAS embraces concepts such dology as case study 1 shown in Figure 1. A related
as SMS, CMS, and DMT supporting the creation of system is focused on close to real-time online com­
a DT for a manufacturing system. According to munication between the physical system and its DT.
Figure 1, the authors suppose that SMS and CMS The focus is on the exchange of the telemetry
are the two main parts of NGAS. CMS is defined as between the real and virtual systems, which enable
a transformative system in which data can be remote teleoperation and programming of the
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 939

industrial robot through the simulation generating Once the simulation is successfully completed, the
ABB Rapid code in real-time and sending commands animations are linked in a process tree to show com­
to the machine. In the TalTech case, an industrial ponent motions prior to and following system initia­
robot is being used with its inspection camera tool lization. In the case of the KSU study, a box moves
where no additional sensors or safety sensors are along a physical conveyor, and an animation of the
being used. box being placed on the belt must be linked to the
system to recognize when the box is removed on the
far end of the belt. This allows for the payload of the
4. CASE STUDIES
Digital Twin system to be loaded at an initial coordi­
Two production lines are considered for case studies, nate and be removed when it reaches its final desti­
and the proposed concept of the DT is applied. nation, both in the tangible and virtual environments.
A Digital Twin of the first production line is estab­ Objects imported into Process Simulate require
lished at KSU and represents a fully automated pro­ appropriate kinematic parameters to be set to mimic
duction system. The second production cell is situated the real system and are defined as either a robot or
in TalTech, and the DT of an industrial robotic cell is a conveyer. Robots are assigned the same poses as
developed. a real system, and conveyors are initialized with
a desired path line and speed using ratios from the
real equipment. Signals are then created in Process
4.1. Case I
Simulate to trigger animations and are defined as
The pick and place line at KSU consist of two Kawasaki inputs, outputs, or memory addresses. Mapping sig­
robotic arms and four conveyor belts controlled with nals externally to the same tag address as in TIA Portal
Siemens SIMATIC S7-1200 series PLCs, a safety PLC, is crucial to ensure successful linkage between the
and an S7-1500 series PLC that overrides the conveyor ladder logic code and the simulation through
systems for the conduction of this study. PLC can PLCSIM Advanced. Figure 2 depicts the virtual proto­
support operators in program development which type using Tecnomatix Process Simulate as DMT.
includes PID controllers, timers, and many other func­
tions (Guerra-Zubiaga 2020a). Along with the physical 4.1.2. Real manufacturing system I
logic controllers, all of the connections are made The physical prototype of the manufacturing line was
using software developed and distributed by commissioned prior to this study for another project.
Siemens. These are Totally Integrated Automation This study saw the addition of a cabinet containing
(TIA) Portal, Process Simulate (Tecnomatix), and S7- a Siemens S7-1500 series PLC and an HMI with an
PLCSIM Advanced (Guerra-Zubiaga 2020b). E-Stop. This PLC acted as an override to control the
various conveyors on the line and represented the
4.1.1. Digital twin of manufacturing system I first step in understanding how to implement the
In the manufacturing industry, setting up Digital Twin. The S7-1500 series was selected as
a production line requires valuable time and a physical surrogate platform to test the override
resources. In response, many workflows try to stream­ functionality. Once the safety system was vetted, it
line this process by designing the system and testing was substituted with a virtual PLC simulated by
it before any equipment is purchased or commis­ PLCSIM Advanced. Each conveyor was connected to
sioned. One method is to create a virtual simulation a variable frequency drive (VFD), which controls the
of the line that incorporates all of the machinery and rotational speed of the conveyor. To adhere to safety
animates the combined functionality. The animations protocols, the conveyor system may be disabled if
are formed from a kinematic structure base where a breach is detected by one of several protective
a start and endpoint are determined in addition to laser curtains controlled by the safety PLC. Figure 3
the desired motion for each component. This study depicts the physical prototype of the conveyor
uses Process Simulate to import models created in system.
Dassault Systèmes’ SolidWorks and applies a custom TIA Portal by Siemens is the software used
kinematic model to the various sections on a time throughout this study to write and execute the pro­
scale basis. gram for each PLC and to view the execution of that
940 D. GUERRA-ZUBIAGA ET AL.

Figure 2. Tecnomatix PS showing signal connections and animations.

logic in real-time. The program is written in ladder inputs and outputs to the S7-1200. The S7-1200
logic with a tag list created from the stored memory, houses a simple program implementing a signal
input, and output addresses, which are mapped to relay that monitors the other PLC for changes with
the Process Simulate model. The HMI is configured a synchronized clock. It executes the program if
here as well, where each button is mapped, and indi­ a difference is detected. It also relays back sensory
cators are linked to the appropriate memory bits of data from the VFD as inputs. This process was done
the PLC. Another option that is set in TIA Portal is the for all I/O, but once mapping was completed accord­
ability to compile the program for virtual PLC use, ingly, the program on the S7-1200 does not need to
a crucial setting for linking the program to PLCSIM be updated again unless the I/O changes.
Advanced. Figure 4 illustrates the implementation of
ladder logic code for the manufacturing system using
the TIA Portal. All physical connections are configured 4.1.3. Factory telemetry using PLCSIM advanced
here, and TIA Portal recognizes the virtual PLC as PLCSIM Advanced is the bridge in connecting the
indistinguishable from a physical module. This allows physical and virtual prototypes. It is a product made
for the same kind of connection as connecting an by Siemens that creates an OPC server locally for
additional PLC or an I/O device like a VFD. Once the interfacing with a physical PLC and, additionally, cre­
connections are configured, the hardware configura­ ates a virtual PLC to connect to. PLCSIM Advanced
tion must be flashed on each device separately makes the connection by creating a virtual ethernet
accompanied by any programming required. connection where the server relays connections to
The system used in this case study uses two PLCs. the physical and virtual hardware. This study used
The first is an S7-1200 that controls the VFD of the a Siemens S7-1500 series PLC that was connected
Vanderlande conveyor belt. The other is the virtual S7- locally to the computer. After the initial PLC setup,
1500 PLC contains the main program and sends the TIA Portal downloads the ladder logic program to the
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 941

a conveyer. This allows for the object to take in

many outputs to execute a task and allows for the
making of input signals that display a status. The
inputs defined in Process Simulate may interact with
the physical system if a full OPC server is used, but the
inputs were not used for manipulation in this case
study. The scripts involved are simple logic blocks
defined in Process Simulate, which trigger an action
when a signal is at a certain state. For the conveyer,
the script tells the conveyor to spin at a given speed,
and it takes the distance calculated from the VFDs
controlling the belt.
The diagram of the connection is seen in Figure 7; it
demonstrates the flow of information and how every
system connected is dependent on one another. The
mediums of information exchange also vary since all
the software communications via the local area net­
work, and PLCSIM Advanced uses a virtual ethernet
adapter to allow connectivity with PROFINET, con­
necting all the physical systems together. HMIs can
also be used in the connection as seen in Figure 7.
Even though they are connected via PROFINET, they
are linked to the virtual PLC in PLCSIM Advanced so
the memory bits changed by the HMI can manipulate
both systems. TIA Portal allows for virtual HMIs to be
used in the event of no physical HMI, and this can also
be used in the current setup for controlling both
Figure 3. Picture of the real manufacturing system used (Long systems.
Conveyor Belt).
4.1.4. IoT interaction
virtual PLC and sets it to run mode. After the connec­ Because the simulated PLC uses an OPC server to
tion has been made, Tecnomatix requires the PLC to communicate to the computer and physical PLC on
be defined so that each signal key can be linked the line, it is possible to enable further customization
appropriately to the right I/O address. Figure 5 depicts of the PLC provided that the networking ports are
the implementation of PLCSIM Advanced to activate opened. Once opened, it is possible to interact with
the virtual PLC. the Digital Twin in a series of versatile approaches.
At this point, a connection has been established, The IP address can even be linked to a company
and any input from the HMI is shown on the simula­ domain as a portal for remote operation across
tion panel as seen in Figure 6. The output column a closed network. Special permissions can be set to
lights show when a signal is either active or disabled allow different types of logging, with accompanying
where non-Boolean signals are displayed as their cor­ alarms, to alert the user to any issues that may arise.
responding integer or decimal values. However, the Using the OPC server does not only limit operation
virtual objects do not move by their designated ani­ to a computer, but its use also makes it possible to
mations because, despite receiving signals, they are interact with the PLC on any mobile device. It even
not yet programmed to respond to them. supports the use of multiple users if an account is set
Once the signals are defined in Tecnomatix, a script up, allowing for any supervisor or engineer to access
is made to handle how the object behaves when it is important data that would not normally be available
given an output. A script is made for each object in on an isolated system. An example of how an IoT
the simulation regardless of if it is a robot or system may integrate with a system using PLCSIM
942 D. GUERRA-ZUBIAGA ET AL.

Figure 4. TIA Portal showing signals when the line is operating.

Advanced is shown in Figure 7. Total control of the 4.2.1. Digital twin of manufacturing system II
system can be seen, and it can be used like an HMI In order to create a digital representation of the
by controlling and viewing various parts of the industrial robotic cell, a model provided by the
process. vendor and post-processed in a Blender to import
to Unity3D environment is used (see Figure 8). 3D
design tools are implemented to adjust pivot points
4.2. Case II
of each joint of the robot in addition to a built-in
This case study was conducted at TALTECH physical model of the game engine for kinematic
University. The selected manufacturing cell consists model calculation inside the environment. All spe­
of an ABB industrial robot with attached 3D Nikon cific animations are performed in simulations by the
scanner. The operator does not interfere with the movement of the joints in motion. A related
machine; a virtual system is sent commands to the Unity3D application target platform is run on a PC,
physical system to control it in the real-time. and application input is managed through Virtual
A virtual model of the robotic cell can be seen in Reality headset/joysticks or keyboard/mouse on
Figure 8, and its physical model can be seen in a desktop application. Application use can be
Figure 9. While this case study follows the NGAS widened to also include Android and Augmented
approach introduced in the previous chapter, differ­ Reality devices through future development. This
ent DMT tools were used to prove that the concept so-called Digital Twin is connected to its physical
is independent of the chosen software and hard­ source through the virtual server.
ware. Still, system design is similar – the Factory
Telemetry data collection approach is being used 4.2.2. Real manufacturing system II
for synchronization of the Digital Twin and the Physical equipment is used, i.e. the ABB IRB1600-10/
Real Manufacturing System. 1.2 industrial robot with a 3D scanner attached to it
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 943

which is connected to the virtual server for

synchronization.

4.2.3. Real robot remote control from the computer

application
To get remote access to the IRC5 controller, the ABB
free software development kit PC SDK was used. It
provides the libraries needed to set the connection
with the controller over the network and an ABB
Robot Communications Runtime module necessary
to connect the controller during run-time (ABB
Developer Center 2016).
PC SDK uses the Microsoft .NET framework (an
environment for developing software applications)
and Microsoft Visual Studio (an integrated develop­
ment environment for writing applications and web­
sites) and provides support for C# and Visual Basic
languages. However, any .NET programming lan­
guage can be used to write the application. Since
the Unity project scripts are written in C#, it was also
used to write the application for remote ICR5 control­
ler management.
To connect to the IRC5 controller, the computer
must be connected to the controller through the
Ethernet network so that it can be found by the PC
application. To acquire the possibility to access the
controller data, it is required to log in to the controller
system. This can be done with the default user since it
has all the necessary permissions for the project
Figure 5. PLCSIM Advanced screen with active virtual PLC.
(Figure 10).
The process for obtaining the current joint values is
(see Figure 9). It is controlled by the vendor controller, quite straightforward. Controller class has the method

Figure 6. Simulation screen in Tecnomatix.

944 D. GUERRA-ZUBIAGA ET AL.

Figure 7. The mapping of the connections and how they interact.

GetPosition that returns an object of type JointTarget, following solution to control the real robot from the
a structure that contains current values for all robot virtual environment is implemented, inspired by the
axes. By repeatedly calling the GetPosition method method suggested in (Cederberg (2012)). A RAPID
when the robot is moving, the joints’ angle data can module called DigitalTwin is created and uploaded
be obtained in every moment of time and passed to to the controller. This module declares a persistent
the virtual robot model to copy the motion. variable (persistent means that the variable remem­
However, to set the joints’ data format by the bers the last value it was assigned between the pro­
external system similarly is not possible because of gram execution) of the type jointTarget (RAPID
the limitation from the system. Therefore, the language type corresponding to the JointTarget class

Figure 8. Virtual model of a selected manufacturing cell.

 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 945

Figure 9. Physical model of the selected manufacturing work cell.

Figure 10. PC application connected to the controller interface.

Figure 11. Digital Twin module in Robot Studio.

946 D. GUERRA-ZUBIAGA ET AL.

in the PC SDK libraries) called targets and MAIN pro­ which uses the PC SDK libraries to communicate
gram (Figure 11). with the robot, is written. To support the data
MoveAbsJ is the RAPID instruction that moves the exchange between this application and the Unity
robot to the absolute position specified not by the project, a named pipe is implemented.
coordinates of the TCP, but by the joints’ angle values. Named pipe is a method for communication
Originally, the targetPos variable is initialized with between programs called a pipe server that instanti­
zeros, the home position of the robot arm. The mod­ ates the pipe and pipe clients that connect to the
ule is set as the main task T_ROB1, and the program pipe. For this project, a duplex pipe is written since,
run mode is set to continuous so that when started, for the proper implementation of the Digital Twin
the program executes continuously in a loop until concept, both applications should be able to send
manually stopped. and receive the data. The ControllerAPI application
To make the real robot follow the movement of the acts as a server; it sets the pipe called
virtual robot model, first, the T_ROB1 task execution is ‘digitalTwinPipe’, and the Unity project acts as
started. To start the execution, the remote client a client. It is important that the server application is
should request for the mastership. Since several cli­ started before the client process attempts to connect
ents can log on to the controller at the same time, for to the pipe because, otherwise, the client tries to
security reasons and to avoid overwriting of the connect to the pipe that does not exist yet. To coor­
RAPID data or program commands, only one client dinate which side has the control over the other and
can have written access and operate the controller avoid overriding the joint values, the direction of the
resources at a time. When the controller is in the data flow is specified in the virtual environment using
automatic mode, the first client to request the master­ buttons Send and Get from the Digital Twin user inter­
ship is granted write access. If the mastership is taken face. Additionally, virtual joystick capabilities are
by another client, the exception is raised. In the man­ blocked when the virtual model follows the real
ual mode, Flex Pendant has the priority, and the robot motion.
mastership is granted to the remote client only if the ControllerAPI is a multithreading application. Main
operator gives it through the pendant. Therefore, the thread (static void Main function) controls the work­
controller should be in automatic mode when the flow of the whole application. After it establishes
Digital Twin project is running. When the mastership a connection with the controller, it waits for the con­
is not needed, it should be released so that other nection with the client. When the client connects to
clients can access the writing mode. the pipe, main thread constantly reads data from the
When the new joints’ angle values are passed from client. Depending on the message received from the
the virtual robot model, the targetPos variable value is client, two other threads are started and finished, and
updated, making the real robot move to the continu­ the pipe is closed.
ously-updated position. Changing the RAPID variable Unity application sends the message ‘readMode’ if
value also requires the mastership; therefore, before it is ready to receive data from the real robot. Then,
changing the targetPos value, the mastership request the writing thread is started. It receives the joint data
should be sent. When the user disconnects from the from the controller and sends it to the client in the
controller, the targetPos value is reset to zero position form of a string. Unity receives the message and
by the specifically-written application function. parses it to get six float variables. The Unity applica­
tion sends the message ‘writingMode’ if it is ready to
4.2.4. Telemetry exchange between the virtual robot send the data from the virtual robot model to the real
model and the real robot one. Then, the reading thread is started. It receives the
The obvious solution is to include PC SDK libraries in joint data from the client and passes it to the con­
the Unity3D project to work with the defined types troller and, therefore, to the real robot. Unity applica­
and functions directly. However, because the Mono tion sends the message ‘disconnect’ if it is ready to
developer that Unity utilizes does not support mixed disconnect from the pipe and run independently from
dynamic libraries (libraries that contain both native the real robot. The server then closes the pipe, moves
codes and managed code), a different solution is the real robot to the initial position (all joint angles set
found. An external application called ControllerAPI, to zero), and disposes of the mastership and
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 947

Figure 12. Communication process diagram.

controller object. Both client and server applications physical prototype depicted in Figure 4 in real-time.
operate in an asynchronous mode, which means that Based on the results made in the presented case
they can read and write data to the pipe at the same studies, the full connection was achieved using PLCs
time. This is necessary for the proper control of the from Siemens along with TIA Portal, PLCSIM
project. The communication process diagram is pre­ Advanced, and Process Simulate. Results were
sented in Figure 12. achieved using free software tools such as Unity3D
The created communication system is tested in the in combination with ABB Robot Studio, which proved
Flexible Manufacturing Systems and Robotics Demo that the concept is independent of the chosen soft­
Center of TalTech and is proven as a working solution. ware and hardware. Important advances on the inte­
However, while the digital model follows the physical gration between simulated processes and DT on a real
robot motion in a real-time manner, the control of the production line have been shown. These contribu­
real robot through the virtual model does not reflect tions were achieved using Siemens software such as
the same precision. Since the control is executed PLCSIM Advanced. Elevant computer integrated man­
through the RAPID command, which sets not the ufacturing systems on a virtual PLC with the internet
speed of the joints but the speed of the TCP, the behaving like an IoT device were likewise presented.
movement of two models is not entirely synchronous. Also, it is shown that each case has different compo­
The real robot moves slower than the virtual model if nents, but both cases follow the same platform, which
the speed is too low and starts to move abruptly if the describes the combination of SMS and CMS in NGAS
speed is too high. The solution to this problem gives to connect virtual and prototype environments by
a premise for future investigation for this case study. factory telemetry. However, there is still a need to
work on this platform to improve the connection of
each part. In future work, there is a possibility to
Results
create subdivisions for each component in the plat­
A novel approach for developing Digital Twin tech­ form and have a better interconnectivity between
nologies was presented, showing the implementation them to elucidate the NGAS as a main concept that
of two case studies from two different universities. involves SMS, CMS, digital and real manufacturing,
Digital Twins, their functionality, advancements, and and factory telemetry.
use in industry through the literature review was
explained in detail. Then, two case studies provided
Conclusion and further work
evidence on how to implement DTs. It has been
shown that connections between virtual and physical The new approach presented is a useful frame­
prototypes were obtained, implementing the Digital work to create NGAS according to industry 4.0
Twin concept. For example, using PLCSIM advanced, it trends. However, the creation of new methodolo­
is possible to manipulate the virtual prototype cre­ gies supporting NGAS at industry 4.0 environ­
ated in Tecnomatix shown in Figure 3 and the ments is required. It is important to acquire
948 D. GUERRA-ZUBIAGA ET AL.

understanding about technologies such as IoT and Industry 4.0 From Private and Public Perspectives.” Appl. Sci
DMT to create a Digital Twin in a Manufacturing 9(14):2934. doi:10.3390/app9142934. 2019.
System to support NGAS aligned to Industry 4.0 Gervasi, M., Franceschini. 2020. “A Conceptual Framework to
Evaluate Human-robot Collaboration.” The International
trends. This work supports the creation of new
Journal of Advanced Manufacturing Technology (3). doi:
products including Cyber and Smart 10.1007/s00170-020-05363-1.
Manufacturing Systems. However, more research Grieves. 2017. Digital twin: Mitigating unpredictable, undesir­
about how these technologies could be used at able emergent behavior in complex systems.“
NGAS is required. From the development of the In Transdisciplinary perspectives on complex systems, pp.
case studies, it is seen that new approaches and 85-113. Springer, Cham, 2017
Groover. 2014. Automation, Production Systems, and Computer-
technologies can be used to create an SMS com­
Integrated Manufacturing, 4th Edition, USA, Pearson. Essex
pliant system with the inclusion of IoT. The pro­ (UK).
posed methodology makes the DT fully capable of Guerra-Zubiaga. 2020a. Improved Process Control Using
being a part of the NGAS, which helps to reduce Hybrid Industrial Internet of Things Network Topology.“
downtime by fully testing a system virtually before In ASME International Mechanical Engineering Congress
it is run with the added benefit of simultaneously and Exposition, vol. 84492, p. V02BT02A008. American
Society of Mechanical Engineers, 2020.
running both systems together to verify complete
Guerra-Zubiaga. 2020b. “Implementing a Novel Framework to
functionality. Important differentiations on simula­ Create Tacit Knowledge Models to Support Human-Robot
tion, visualization, and virtualization aspects were Interactions.„ In ASME International Mechanical Engineering
clarified through the implementation of a Digital Congress and Exposition, vol. 84492, p. V02BT02A013.
Twin in a manufacturing system. However, the American Society of Mechanical Engineers, 2020.
Digital Twin shall continue to improve with Havard. 2019. “Digital Twin and Virtual Reality: A Co-simulation
Environment for Design and Assessment of Industrial
newer technology and is an excellent example of
Workstations.” Production & Manufacturing Research 7 (1):
the new paradigms of digital manufacturing and 472–489. doi:10.1080/21693277.2019.1660283.
the concepts of industry 4.0. Jackson, E.; Borton. 2016. “Digital Manufacturing and Flexible
Assembly Technologies for Reconfigurable Aerospace
Production Systems.” Procedia CIRP 52: 274–279.
doi:10.1016/j.procir.2016.07.054.
Disclosure statement
Kusiak. 2018. “Smart Manufacturing.” International Journal of
No potential conflict of interest was reported by the author(s). Production Research56, no. 1-2 (2018): 508–517. doi:10.1080/
00207543.2017.1351644.
Kuts. 2018. “Adaptive Industrial Robots Using Machine Vision.”
References in ASME 2018 International Mechanical Engineering Congress
and Exposition, Pittsburgh.
ABB Developer Center. 2016. Pc Sdk. [Online]. Available at: Kuts. 2019. “Digital Twin Based Synchronised Control and
[Accessed 6 May 2019]. http://developercenter.robotstudio. Simulation of the Industrial Robotic Cell Using Virtual
com/pcsdk/ Reality.” Journal of Machine Engineering 19 (1): 128–145.
Armendia. 2019. Mikel, Aitor Alzaga, Flavien Peysson, Tobias doi:10.5604/01.3001.0013.0464.
Fuertjes, Frédéric Cugnon, Erdem Ozturk, and Dominik Flum. Lee, B., Jin. 2016. Introduction to Cyber Manufacturing.
“Machine tool: from the digital twin to the cyber-physical Manufacturing Letters 8 (2016): 11-15. doi: 10.1016/j.
systems.„ In Twin-Control, pp. 3-21. Springer, Cham. mfglet.2016.05.002.
Cederberg, O., and Bolmsjö. 2012. Remote Control of Lee, S. 2017. Introduction to Embedded Systems: A Cyber-
a Standard ABB Robot System in Real Time Using the physical Systems Approach. London: MIT.
Robot Application Protocol (RAP). In: Proceedings of the Lin, F.; Newman. 2009. “Manufacturing Process Analysis with
33rd ISR (International Symposium on Robotics)October, vol. Support of Workflow Modelling and Simulation.”
7, no. 11. 2002. International Journal of Production Research 47 (7):
Cellier, G. 2013. Continuous System Modeling. NY: Springer 1773–1790.
Science & Business Media. Lu. 2020. “Digital Twin-Driven Smart Manufacturing:
Garcia, M.-L., Garcia. 2020. Human-Robot Collaboration Based Connotation, Reference Model, Applications and Research
on Cyber-Physical Production System and MQTT. Procedia Issues.” Robotics and Computer-Integrated Manufacturing 61:
Manufacturing 42 (2020): 315-321. doi: 10.1016/j. 101837. doi:10.1016/j.rcim.2019.101837.
promfg.2020.02.088. Mahmood. 2017. “Performance Analysis of a Flexible
Gerrikagoitia, J. K., G. Unamuno, E. Urkia, and A. Serna, Manufacturing System (FMS).” Procedia CIRP 63: 424–429.
Gerrikagoitia. 2019. “Digital Manufacturing Platforms in the doi:10.1016/j.procir.2017.03.123.
 INTERNATIONAL JOURNAL OF COMPUTER INTEGRATED MANUFACTURING 949

Mahmood. 2019. “Development of Cyber-Physical Production Pilz Australia Industrial Automation LP. 2016. “The Future of
Systems Based on Modelling Technologies.” Proceedings of Control Systems and Industry 4.0.” Available: http://www.
the Estonian Academy of Sciences 68 (4): 348. doi:10.3176/ processonline.com.au/content/industrial-networks-buses
proc.2019.4.02. /article/the-future-of-control-systems-and-industry
Mahmoodpour, M., and Lobov. 2018. “IoT-Based Senses for -4-0-1365495420 .
Virtual Enterprises.” In IECON 2018-44th Annual Conference PLM, S. 2018. “Digital Manufacturing Definition.” Siemens AG,
of the IEEE Industrial Electronics Society, Washington, DC, [Online]. http://www.plm.automation.siemens.com/en_us/
USA. plm/definition/digital-manufacturing.shtml .
Malik and Bilberg. 2018. “Digital Twins of Human Robot Qi. 2019. “Enabling Technologies and Tools for Digital Twin.”
Collaboration in a Production Setting.” Procedia Journal of Manufacturing Systems. doi:10.1016/j.
Manufacturing 17(2018):278–285. doi: 10.1016/j. jmsy.2019.10.001.
promfg.2018.10.047. Sandeep, G., and Naveen. 2015. “Globally Accessible Machine
Modoni, S., Terkaj. 2016. “A Telemetry-driven Approach to Automation Using Raspberry Pi Based on Internet of Things.”
Simulate Data-intensive Manufacturing Processes.” In International Conference on Advances in Computing,
Procedia CIRP. doi: 10.1016/j.procir.2016.11.049. Communications and Informatics (ICACCI),pp. 1144-1147. IEEE,
Monostori. 2014. “Cyber-physical Production Systems: Roots, 2015.
Expectations and R&D Challenges.” Procedia CIRP 17: 9–13. Stark, F.; Lindow. 2019. “Development and Operation of Digital
doi:10.1016/j.procir.2014.03.115. Twins for Technical Systems and Services.” CIRP Annals 68 (1):
129–132. doi:10.1016/j.cirp.2019.04.024.
Mourtzis. 2019. “Simulation in the Design and Operation of
Strang, G., Anderl. 2016. Dynamic, Adaptive Worker Allocation for
Manufacturing Systems: State of the Art and New Trends.”
the Integration of Human Factors in Cyber-Physical Production
International Journal of Production Research 58, (7): 1927-
Systems. In Advances in Ergonomics of Manufacturing:
1949. doi:10.1080/00207543.2019.1636321.
Managing the Enterprise of the Future, pp. 517-529. Springer,
Mourtzis, D.; Bernidaki. 2014. “Simulation in Manufacturing:
Cham, 2016.
Review and Challenges.” Procedia CIRP 25: 213–229.
Tao. 2018. “Data-driven Smart Manufacturing.” Journal of
doi:10.1016/j.procir.2014.10.032.
Manufacturing Systems 48 (Part C): 157–169. doi:10.1016/j.
Nikolakis. 2019. “The Digital Twin Implementation for
jmsy.2018.01.006.
Linking the Virtual Representation of Human-based
Tao, C. Qi. 2017. “IIHub: An Industrial Internet-of-Things Hub
Production Tasks to Their Physical Counterpart in the
toward Smart Manufacturing Based on Cyber-Physical
Factory-floor.” International Journal of Computer
System.” IEEE Transactions on Industrial Informatics.
Integrated Manufacturing 32 (1): 1–12. doi:10.1080/
doi:10.1109/TII.2017.2759178.
0951192X.2018.1529430.
Wang. 2018. “Digital Twin for Rotating Machinery Fault Diagnosis
Nikolakis, S., Makris. 2018. A Cyber-physical Context-aware in Smart Manufacturing.” International Journal of Production
System for Coordinating Human-robot. Procedia CIRP 72 Research 1: 1–15.
(2018): 27-32. doi: 10.1016/j.procir.2018.03.033. Zambrano Degan. 2020. “Bruno, Gustavo Franco Barbosa, David
Noble. 2011. Forces of Production: A Social History of Industrial Guerra-Zubiaga, Navid Nasajpour-Esfahani, and Abdullah Al
Automation. NJ: Transaction. Mamun. 3D Digital Manufacturing Applied on Drilling of
Paavel, K., Majak. 2017. Development of a Product Lifecycle 7050-T651 Aluminum Performed by Robot.“ In ASME
Management Model Based on the Fuzzy Analytic Hierarchy International Mechanical Engineering Congress and
Process. Proceedings of the Estonian Academy of Sciences Exposition, vol. 84492, p. V02BT02A011. American Society of
66 (3): 279–286. doi: 10.3176/proc.2017.3.05. Mechanical Engineers.
Copyright of International Journal of Computer Integrated Manufacturing is the property of
Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted
to a listserv without the copyright holder's express written permission. However, users may
print, download, or email articles for individual use.