Digital Twin in General

Advanced Engineering Informatics 55 (2023) 101876
Contents lists available at ScienceDirect
Advanced Engineering Informatics

journal homepage: www.elsevier.com/locate/aei
A systematic review of digital twin about physical entities, virtual models,

twin data, and applications
Xin Liu a, b, Du Jiang b, c, *, Bo Tao a, c, Feng Xiang c, d, Guozhang Jiang b, Ying Sun c, d,
Jianyi Kong a, b, Gongfa Li a, b, c, *
a
Key Laboratory of Metallurgical Equipment and Control Technology of Ministry of Education, Wuhan University of Science and Technology, Wuhan 430081, China
b
Hubei Key Laboratory of Mechanical Transmission and Manufacturing Engineering, Wuhan University of Science and Technology, Wuhan 430081, China
c
Research Center for Biomimetic Robot and Intelligent Measurement and Control, Wuhan University of Science and Technology, Wuhan 430081, China
d
Precision Manufacturing Research Institute, Wuhan University of Science and Technology, Wuhan 430081, China
A R T I C L E I N F O A B S T R A C T
Keywords: The digital twin is a crucial technology for realizing smart manufacturing and industrial digital transformation,
Digital Twin which has received extensive attention and research from industry and academia. After 20 years of development,
Physical entity the application area of digital twins has been pervasive. Due to the diversity of application areas, various
Virtual model
reference models and research methods have been presented for the components of the digital twin. Therefore,
Twin data
Applications
this paper provides systematic research of current studies on the basic components of the digital twin. This paper
analyzed 117 articles from 2017 to 2022. By clarifying the relationship between the digital twin and the cyber-
physical system, it first clarified the definition, characteristics, and application areas of the digital twin. On this
basis, the research methodology of the core components of the digital twin (physical entities, virtual models, and
twin data) is analyzed. At the same time, the application areas of digital twins are analyzed and delineated, and
the application potential of the digital twin is explored. Finally, the research results and future research rec
ommendations are presented.
1. Introduction called digital twins. In 2010, the “digital twin” was first proposed in
writing by NASA and further developed [6]. In 2012, NASA and the Air
Digital transformation and intelligent upgrading have become vital Force Research Laboratory collaborated to propose digital twin exam
factors in releasing tremendous development momentum and have ples of future aircraft to address the need for future aircraft to be highly
become a consensus among countries regarding future global develop loaded, lightweight, and serve longer in extreme environments [7]. In
ment [1]. Taking the manufacturing industry as an example, the United 2014, Professor Michael Grieves published a white paper on digital
States, Germany, the United Kingdom and other countries have put twins, formally defining the 3 main parts of a digital twin: physical
forward their own national manufacturing development strategies, and entities in real space, virtual models in virtual space, and data that
the most basic technology drivers are digitalization and intelligence [2]. connects physical entities and virtual models together [8]. In order to
As a multidisciplinary technology that fully uses models, data, ma study the development trend of digital twins, papers related to digital
chines, and computers, digital twins can provide real-time, efficient, and twins from 2012 to September 2022 were counted. The search keyword
intelligent services in multiple areas of smart manufacturing and can was digital twins and the database were Scopus, Web of Science, IEEE
also serve as a bridge between the physical world and the information Xplore, and Google Scholar. Finally, an annual graph of the number of
world [3,4]. The concept of a digital twin dates back to a 2002 report by digital twins is obtained, as shown in Fig. 1. It can be seen from the graph
the University of Michigan to industry for the establishment of the that the number of documents has soared every year since 2017, with
Center for Product Lifecycle Management [5]. Limited by the immatu the number of papers exceeding 800 in 2021. From the growth trend in
rity of data acquisition technology, computer performance and algo the figure, it is speculated that the number of papers in 2022 is expected
rithms at the time, the early concepts proposed by Professor Michael to exceed 1000.
Grieves did not receive widespread attention at the time and were not Physical entities, virtual models, and twin data are the three
* Corresponding authors.
E-mail addresses: jiangdu@wust.edu.cn (D. Jiang), ligongfa@wust.edu.cn (G. Li).
https://doi.org/10.1016/j.aei.2023.101876
Received 11 October 2022; Received in revised form 29 November 2022; Accepted 4 January 2023
Available online 28 January 2023
1474-0346/© 2023 Elsevier Ltd. All rights reserved.
X. Liu et al. Advanced Engineering Informatics 55 (2023) 101876
devices to ensure seamless connectivity and data exchange through

direct physical or indirect cloud-based connectivity. Lu et al. [27] pro
posed a reference model for the digital twin, including an information
model for abstract physical object specification, a data communication
mechanism, and a data processing module. Semeraro et al. [28] answer
the question of what a digital twin is, when it can be developed and why
it should be used by analyzing the concept of a digital twin, its life cycle
stages, and its primary functions at different stages. Tao et al. [9] con
ducted a systematic study of the current state of research on digital twin
modeling, giving a comprehensive and insightful analysis of digital twin
models in terms of application areas, levels, disciplines, dimensions,
universality, and functionality.
Several articles have also investigated the application areas of the
digital twin. Zheng et al. [29] conducted a systematic study on the
application of digital twins. They proposed a digital twin application
framework consisting of physical space, virtual space, and information
Fig. 1. Number of digital twin papers published per year. processing layers for product lifecycle management. Tao et al. [30]
analyzed 50 papers and 8 patents. This article provides an overview of
foundational components of the digital twin [9]. Physical entities are the key technologies for data processing, interactive collaboration, and
foundation of the digital twin, a single device, or an extensive system services. The current state of research on digital twins in product design,
like a shop-floor. Physical entities are objectively present and can production, prediction, and health management is discussed. Due to the
perform at least one task independently. Monitoring the working process significant differences in their composition, usage conditions, and
of physical entities through different kinds of sensors to obtain data on application scenarios, the research objects in different industries have
the operating status of physical entities is the focus of physical entities’ different characteristics regarding modeling strategies. Therefore, Fang
research. The virtual model is the key to differentiating digital twin et al. [31] outlined the focus of digital twin modeling in various in
technology from other technologies and is also the basis for the reali dustries and summarized the related supporting technologies and
zation of virtual-real interaction. A virtual model is a copy of a physical methods. There are also scholars who focus on the current application of
entity that can characterize and describe physical entities from multiple digital twins in a certain field. Opoku et al. [25] studied the application
temporal and multi-spatial dimensions [10]. The geometric dimensions, of digital twin technology in the construction industry. The study ana
material properties, and other parameters of the physical entity are lyzes in detail application scenarios such as construction process infor
given to the virtual model, and historical working data drive the simu mation modeling, equipment management and energy simulation.
lation of the virtual model to realize the control, optimization, and Table 1 shows a summary of these review papers. From these review
prediction of the actual working state of the physical entity [11]. Twin papers, it can be seen that there is a lot of research on the concept,
data is the driver of digital twins, which generally includes physical, technology and application field of digital twin. However, the current
virtual, knowledge, and derivative data [12]. Twin data contains study does not emphasize the respective approaches to physical entities,
different types that can better monitor physical entities’ working status virtual models, and twin data. In addition, most of the reviews on digital
and drive virtual models’ operation. twin applications focus on areas such as industry and manufacturing,
Digital twins can be used for the prediction of the remaining lifetime and there is a lack of reviews of promising but less researched areas.
of spacecraft [13], for the digital design of roll-shops [14], and for
healthcare for the elderly [15]. Digital twins have excellent performance 1.2. Purpose of this paper
in design simulation optimization, operation monitoring, predictive
maintenance, etc., and have been widely used in industrial production, The digital twin is an essential technology for smart manufacturing
smart cities, smart medical and other fields [16]. It has gradually and Industry 4.0. Its importance has been recognized by academia and
become an essential driving force for digital transformation and devel industry, so with a detailed and insightful review, this paper aims to:
opment in various areas. Different domains have different research ap
proaches to the perception of the working state of physical entities, the 1) Analyze and compare the similarities and differences between the
construction of virtual models, and the management of twin data. digital twin and cyber-physical system technologies.
Therefore, it is urgent to divide the application fields according to the 2) Provides a comprehensive summary and analysis of physical entities,
characteristics of digital twin models and to summarize the general virtual models, and twin data.
research methods applicable to different areas. 3) Summarize and divide the popular application areas of the digital
In this paper, we will systematically review digital twins’ current twin, and summarize some potential application neighborhoods.
research status and achievements, mainly from the three core compo 4) Propose possible future research directions and research methods.
nents mentioned above and the application areas of digital twins.
The rest of this paper is organized as follows. Section 2 describes the
1.1. Related review works literature review methodology. Section 3 clarifies the differences be
tween the digital twin and cyber-physical system. Section 4 reviews the
Since 2017 digital twins have been developing rapidly. Numerous research methods for physical entities, virtual models, and twin data.
review articles on digital twin definitions [17,18], reference models Section 5 analyzes the application areas of the digital twin. Section 6
[19], key technologies [18,20], and application areas have been pub presents some observations and future developments of digital twins.
lished [16,18,21–25]. In this paper, 8 review articles are selected for Section 7 summarizes the whole text.
brief introduction, and the key points of this research are proposed
through analysis. 2. Research method
Many articles have studied in detail the definition, connotation,
development history, key technologies, and reference models of the 2.1. Classification criteria
digital twin. B. R. Barricelli et al. [26] analyzed 75 articles. The litera
ture concludes that the digital twin must be equipped with network The primary literature classification criteria for this study are based
2
Table 1
Comparison of digital twin review articles.
Ref. Timeframe Number of Object of the overview Key findings
papers
reviewed
[26] 1990–2019.07 75 • To explore the definition of digital twin. • Definition: A computer-based model that simulates, emulates, mirrors,
or “combines” the life of a physical entity.
• Features: Both physical and digital twins must be equipped with
network devices.
[27] 2004–2019.07 85 • To explore the characteristics of the digital twin. • Application scenarios: manufacturing, aviation and healthcare.
• Digital twin reference model: information model for abstracting
physical object specifications, communication mechanism for
transmitting data in both virtual and real directions, and data
processing module.
• Most used scenarios: healthcare, maritime and shipping,
manufacturing, urban management, aerospace
[28] 2002–2020 115 • To explore the application areas of the digital twin. • Research challenge: Need to develop digital twins that can interact with
humans. Need to explore modular approach to build flexible DT
solutions.
• Current research on digital twin modeling rests mainly on model
construction, and modeling integrity is a pressing issue.
• Established a digital twin application framework for product lifecycle
management (consisting of three parts: physical space, virtual space
and information processing layer)
• Key Theories: Modeling, Data Fusion, Interaction and Collaboration,
and Services.
[9] 2014–2021.12.31 160 • Exploring frameworks and development methods • Industrial applications: Design, Production, PHM.
for implementing the digital twin in manufacturing.
[29] 2012–2018 27 • To explore the methods, application areas and • Key implementation technologies for digital twin:
implementation challenges of designing and
implementing the digital twin.
[30] 2003.01–2018.04 50 & 8 patents • To explore the modeling approach of digital twin. • Basic technology: IoT, Sensor, Visualization, Data transmission, Data
management, Data storage.
[31] 2005–2021.10 144 • Research and build the application framework of • Core technology: Data fusion, FEM, Data drive, Physical model, State
digital twin. model, Agent model.
• Definition of the digital twin, application areas, key • Advanced technology: AI, Big data analytics, Mobile internet,
technologies and challenges of implementation. Blockchain, Cloud computing.
• The development of digital twins in the construction industry is still in
its infancy.
• To explore the digital twin application scenarios, • The current focus on the application of digital twins in architecture is
and the modeling approaches in different scenarios. focused on the design and engineering phase, while the demolition and
restoration phases are neglected.
[25] 2010–2020 22 • To examine the application of digital twin in
respective lifecycle phases of a construction project.
Fig. 2. Literature review methodology.
3
on the digital twin research approach. The research methodology of the

digital twin is deconstructed, classified, and summarized in terms of
physical entities, virtual models, and twin data. On this foundation, the
different application areas are dissected and classified based on the
application characteristics of digital twins, and the general research
methodology of digital twins is summarized.
2.2. Literature review methodology
The literature review method is divided into three steps: literature

collection, literature screening, and literature classification, as shown in
Fig. 2.
STEP 1: literature collection. As seen in Fig. 1, the number of papers
on digital twins has proliferated since 2017. Therefore, the time span of
the literature collection is from January 2017 to September 2022. Select
the databases as Scopus, Web of Science, IEEE Xplore, and Google
Scholar. The methods of literature collection are shown in Table 2.
Search for keywords to see if the digital twin is included in the titles and
abstracts of the documents obtained. After screening, 412 articles con
taining keywords were preliminarily retained, of which the number of
articles from 2017 to 2022 was 6, 11, 57, 128, 156 and 54, respectively.
It can be seen that the number of digital twin articles published every Fig. 3. Top five countries in the number of articles.
year is increasing year by year.
STEP 2: literature filtering. The screening method for papers is to
read the abstract, introduction, and conclusion of all documents. If these
three parts of the articles contain one of the digital twin definitions,
digital twin applications, physical entities, virtual models, and data, the
paper is left for review, otherwise deleted. In this way, 117 documents
related to the content of this article were finally screened.
STEP 3: literature classification. Classify the screened literature, read
each type of paper intensively and summarize commonalities and
claims.
2.3. Status analysis of digital twin research
After eliminating the irrelevant literature, 117 papers were finally

obtained. To demonstrate the current status of the development of
digital twins, we counted the countries and published journals of the 412
documents obtained by filtering in STEP 1, as shown in Fig. 3 and Fig. 4,
respectively.
The five countries with the highest number of these papers published
were China, the United States, the United Kingdom, Italy, and Germany.
Among them, China issued the most articles, accounting for 30.83 %.
These countries have a high level of science and technology and a spe Fig. 4. Top five journals in the number of articles.
cific foundation of information technology, which can provide a sup
portive environment for digital twins’ research, development, and research are not only in large quantity but also in high quality [32].
application. Although digital twins originated in the United States,
research on digital twins in other countries shows signs of catching up. 3. Concept clarification
The attention on digital twin by these world powers reflects the huge stir
digital twin has caused in the industry. Modern technologies such as artificial intelligence (AI), cloud
From the analysis of publications, the five journals that published the computing (CC), big data analytics (BDA), cyber-physical systems, and
most digital twins among these papers are shown in Fig. 4. The Journal of the Internet of Things (IoT) have extensively promoted the development
manufacturing systems published the most articles, accounting for 11.65 of intelligent manufacturing. Among these technologies, cyber-physical
%. The analysis of the publications shows that digital twins’ current systems are similar to digital twin concepts and are easily confused [33].
research and application are mainly focused on the manufacturing field. Cyber-physical systems and digital twins have received much
At the same time, these journals are top journals with an international attention from academia, industry, and government [34]. Some scholars
reputation, which reflects that the current papers on digital twin believe that the digital twin is a tool to implement the virtual-real
interaction of cyber-physical system, and study the digital twin as part
Table 2 of cyber-physical systems. Zheng et al. [35] believe that digital twin
Literature collection methods. technology is one of the core technologies for implementing cyber-
Searching Index Corresponding Content physical systems. It is an accurate virtual copy of a physical object or
system, including its properties and environment. Son et al. [36]
Time span January 2017 - September 2022
Database Scopus, Web of Science, IEEE Xplore, Google Scholar
consider that the digital twin is a virtual factory that mimics a
Keyword digital twin manufacturing site, synchronizes its information and operations, and
Article Type Journal papers that have been published promotes the cyber-physical system’s functionality by predicting the
4
manufacturing site’s future state using various analytical results. Park Table 4
et al. [37] suggested a digital twin-based CPPS (cyber-physical pro Cyber-Physical System definition.
duction system) architectural framework that overcomes the perfor Ref. Publication Definitions
mance hurdle. A digital twin-driven design and control cyber-physical time
system was introduced in [38] to lessen the difficulties of RCL (roller [41] 2017 “Cyber-physical system is a new trend in Internet-of-
conveyor line) construction and remove faults between design and Thing-related research works, where physical systems act
control. However, according to the definitions of digital twin and cyber as the sensors to collect real-world information and
physical system, they both have the ability to achieve virtual-real communicate them to the computation modules to the
corresponding physical systems through a feedback loop.”
interaction themselves. Therefore, this section compares the defini [49] 2018.1 “Cyber-physical systems are physical facilities with
tion, mapping mode, and application scenario to distinguish the rela embedded sensors, processors, and actuators controlled or
tionship between the digital twin and cyber-physical system. monitored by computers.”
[33] 2019.8 “Cyber-physical systems are multidimensional and
complex systems that integrate the cyber and dynamic
3.1. Definition physical worlds.”
[36] 2021.5 “A cyber-physical system is an engineering system that
The concept of a digital twin was first proposed by Michael Grieves in seeks to enhance the performance of manufacturing
2003 in a product lifecycle management class at the University of systems, aggregates the information of heterogeneous
manufacturing elements and applications, and uses it to
Michigan [5]. The concept of cyber-physical systems originated in 2006 predict future conditions in manufacturing plants wherein
by Helen Gill of the National Science Foundation to describe increas abnormal scenarios occur.”
ingly complex systems [39]. Although digital twins were proposed
earlier than cyber-physical systems, it was cyber-physical systems that
received attention first, and digital twins have only developed in recent physical systems. The core purpose of the digital twin is to control,
years. Tables 3 and 4 are some definitions of digital twins and cyber- optimize and predict the working state of the physical entity through a
physical systems in recent years. It can be seen from the definitions of virtual model. Digital twins use many functional models to describe
both that although digital twins and cyber-physical systems have the physical entities, such as geometric models, physical models, and so on
function of interacting with the virtual world and the physical world, [43]. Digitalization and visualization of production processes through
they are implemented in different ways. Digital twin optimizes the the combination of different functional models. The working data
working state of physical entities through the simulation of virtual generated by the physical entity and the simulation data obtained from
models [40]. In contrast, cyber-physical systems control the operation of the virtual model together form the twin data [30]. Based on twin data,
physical equipment by analyzing the raw data collected by sensors and virtual models can be used to guide, control, optimize, and predict the
issuing instructions by computers. At the same time, the essential working state of physical entities. Therefore, models and data are core
components of the digital twin are physical entities, virtual models, and elements of a digital twin.
twin data, while the basic elements of cyber-physical systems are the
perception layer, the network layer, and the control layer. The most 3.3. Virtual reality mapping method
significant difference is that the digital twin must build a virtual model,
while the cyber-physical system does not need to. The matching method between physical and virtual is different.
Digital twin emphasizes the one-to-one correspondence between the
3.2. Core elements physical entity and the virtual model. The virtual model must have the
exact structural dimensions, physical properties, assembly relationships,
The core elements of digital twins and cyber-physical systems are etc., as the physical entity. In cyber-physical systems, the computer at
different. Cyber-physical systems enable precise control of physical the network layer controls the physical devices at the sensing layer. The
processes, remote collaboration, and self-management [41]. Data is the data obtained at the sensing layer is sent to the network layer, which
foundation on which cyber-physical systems manage physical processes. analyzes it and sends control commands to the sensing layer. Therefore,
Therefore, data acquisition equipment is extremely important for cyber- the virtual-real mapping method of the information-physical system is
physical systems. Large-scale data acquisition with multiple sensors one-to-many.
distributed across physical devices and environments. The collected data
is analyzed and processed by a computer, and control commands are 3.4. Application scenario
generated according to defined rules [42]. Actuators operate according
to control commands in response to changes in physical processes. Both digital twin and cyber-physical systems can be applied to large
Therefore, sensors and controllers are the core elements of cyber- industrial systems, such as aerospace and shop-floor, and realize virtual-
real interaction through their respective mapping methods. However,
digital twin also has applications in other areas, such as smart healthcare
Table 3
and smart cities. Liu et al. [44] proposed a digital twin-based cloud
Digital Twin definition.
medical system framework that enables monitoring, diagnosis, and
Ref. Publication Definitions
prediction of users’ health status. Youngjib et al. [45] propose a new
time
framework for building digital twin cities that can better understand
[46] 2019.11 “Digital twins can be defined as (physical and/or virtual) spatiotemporal information about physical vulnerabilities for effective
machines or computer-based models simulating,
emulating, mirroring, or “twinning” the life of a physical
risk-informed decision-making. The digital twin generally has more
entity.” application scenarios than cyber-physical systems.
[47] 2020.6 “Digital Twins (DT) are defined as simulation models that Digital twin emphasizes the correspondence between a physical en
are both getting data from the field and triggering actions tity and its virtual model, optimizing the performance of physical en
on the physical equipment.”
tities through virtual and real data interaction. Cyber-physical systems
[15] 2020.8 “A digital twin is a digital representation of a physical
asset reproducing its data model, behavior, and consider biological systems composed of multiple physical devices and
communication with other physical assets.” carry out data acquisition and processing operations from the system.
[48] 2021.7 “A digital twin is a virtual model of a physical entity, with Therefore, digital twins and cyber-physical systems are different tech
dynamic, bi-directional links between the physical entity nologies with different application purposes and implementation
and its corresponding twin in the digital domain.”
methods, and there is no inclusion relationship between them. Digital
5
twins or cyber-physical systems should be selected according to the achieving accurate match modeling between physical entities and vir
specific needs when applying. tual models. Duan et al. [55] divided the physical entities into two parts,
the equipment layer, and the test layer. The components that make up
4. Current development of digital twin components each device can be regarded as the equipment layer, such as blades,
bearings, motors, etc. The test layer includes static tests and dynamic
The three core components of digital twins are the physical entity, tests of the system. Dividing physical entities into multiple modules
virtual model, and twin data [30,50], so this section reviews the current facilitates the modular modeling of virtual models and the accurate
research methods in these three sections. matching of virtual models to physical entities.
The usefulness analysis of the components of the physical entity is
conducive to improving the modeling speed. The complexity of physical
4.1. Physical entities entities varies in different fields, and the elements of physical entities are
analyzed to screen out parts that do not affect the model’s functionality.
Through the intensive reading of 117 documents, we first determine For example, the shop-floor includes many components, not all of which
whether there is research content in the literature on one of the three need to appear in the virtual model. Only a few core components need to
parts of physical entities, virtual models and twin data. Ignoring the be modeled, and the rest can be ignored. In this way, the modeling speed
literature without these three parts, 88 documents were obtained. Sec is accelerated, and the complexity of the model is reduced without
ondly, according to the research focus, the corresponding number of affecting the functionality of the virtual model.
documents studying physical entities, virtual models and twin data is
obtained, as shown in Fig. 5. As can be seen from the figure, the most 4.1.3. Perception of physical information
research on virtual models accounted for 45.45 %, and the slightest The acquisition of physical information is the focus of physical entity
research on physical entities accounted for 21.59 %. This may be due to research, and information collection is inseparable from sensors. The
the fact that physical entities, as objects of study, are machines or de acquisition of biological information importantly includes the sensing of
vices that already exist, which leads to the neglect of the study of environmental information, the sensing of worker operation informa
physical entities. However, physical entities are the basis for virtual tion, and the acquisition of equipment working data, as shown in Fig. 6.
model construction and are the source of twin data. It is essential to Sensors distributed throughout the manufacturing process could enable
conduct research on physical entities. twins to capture operational and environmental data relevant to phys
ical processes in the real world [56].
4.1.1. Definition of physical entities Selection of the sensor type. Different types of sensors are used for
Physical entities refer to the study object ontologies and their various data acquisitions. For example, position/angle encoders for
ancillary resources. Bevilacqua et al. [51] proposed a digital twin obtaining position information, current and voltage sensors for obtain
reference model, and the physical entities consist of physical industry ing current and voltage data, and torque sensors for obtaining force
resources such as products, personnel, equipment, material, process, information. There are also devices for sensing working conditions: vi
environment, and facility. Zhang et al. [52] proposed an implementation bration accelerometers, vibration displacement sensors, thermometers,
framework for the digital twin shop-floor. The physical factory they microphones, acoustic emission sensors, RFID readers for code identi
defined includes material resources, such as equipment resources, in fication, energy sockets, and industrial cameras, etc., which can be used
dustrial robots, AGV cars, intelligent production lines, and information to collect the data people need to build digital twin [57].
infrastructures, such as computer network systems, data centers, and Perception of physical environment information. Choi et al. [58]
information interaction systems. Zhang et al. [53] believe that the used two RGB-D sensors to reconstruct and track the working environ
workshop entity refers to the physical elements that are indispensable ment. A sensor scans an area of the physical environment through 3D
for daily production and processing in the production workshop, point cloud data. The other also scans another area of the environment,
including processing equipment, material transportation systems, op tracking the user’s 3D bone information.
erators, and production environments. Physical entities in a physical Perception of operator information. Nikolakis et al. [59] proposed a
assembly space are defined by Sun et al. [54] as peoples, devices, way to identify the natural actions of operators in manufacturing sys
components, and products. Therefore, physical entities refer not only to tems. The Kinect v2 depth camera and an electronic glove are used to
themselves but also to the ancillary resources such as workers, envi identify the operator’s movements and motion constraints. The corre
ronment, and materials associated with the physical entities. sponding data is generated in the physical environment and then sent to
the virtual model so that the virtual model can simulate the behavior of
4.1.2. Selection of physical entities the actual operator in the virtual environment.
Classifying the components of a physical entity facilitates the con Perception of work data. Aivaliotis et al. [60] used the digital twin to
struction of accurate virtual models. When studying the physical entity calculate the service life of a six-axis robot. When collecting signals, in
ontology, it is reasonable if the products, workers, etc., associated with it addition to the real sensors, 3 virtual sensors were constructed, because
are taken into account and built into virtual models, but this method in the digital twin model simulation process, data that could not be
may increase the difficulty of building virtual models, and at the same reached by real machines was generated. Virtual sensors could monitor
time, the complexity of the model will also increase. Therefore, classi and collect data from various virtual parts of the device. Three virtual
fying and packaging the components of physical entities is conducive to sensors acquire position, velocity, and acceleration signals.
4.2. Virtual model
A virtual model is a digital mirror image of a physical entity and an

essential part of enabling digital twins [22]. Therefore, establishing a
virtual model with a high degree of match with physical entities is the
key to achieving digital twin technology. There is currently a lot of
research on virtual model modeling methods, and many methods for
constructing virtual models have been proposed.
Fig. 5. Distribution of the three parts.
6
Fig. 6. Physical information perception method.
4.2.1. Modeling method paradigm are proposed.

Tao et al. [50] proposed that a complete digital twin should include Wu et al. [61] proposed a multidimensional digital twin conceptual
five parts: the physical entity, the virtual entity, the data, the service, modeling method, which describes the digital twin’s composition,
and the connection. Professor Tao Fei proposed the digital twin five- behavior, and rules in detail. It introduces the TRIZ function model into
dimensional model, which defines the components of the virtual the five-dimensional framework and improves its system construction
model, namely the geometric model, the physical model, the behavior process. Zhang et al. [43] proposed a five-dimensional fusion model,
model, and the rule model, as shown in Fig. 7. This provides a research including a geometric model (GM), a physical model (PM), a capability
paradigm for virtual models. Most of the digital twin virtual models in model (CM), a behavioral model (BM), and a rules model (RM). This
research are built on these four parts. five-dimensional fusion model adds an ability model to the five-
These four parts can describe and characterize physical entities from dimensional model, which describes the capabilities of each entity in
multi-time scales and degree-space scales. Geometric models describe the physical layer, including what it can do, what it can do, what it is
the geometric parameters of physical entities, and standard tools are doing, and what it is doing. The purpose of adding a capability model is
SolidWorks, 3D MAX, CATIA, etc. The physical model adds the physical to understand the capabilities and roles of various manufacturing re
properties of the physical entity and constraints based on the geometric sources in the manufacturing system and infer them into relevant
model, and tools such as ANSYS and ABAQUS are usually available. The knowledge to support outcome prediction, performance evaluation, plan
behavior model describes the evolutionary behavior, real-time response optimization, etc., to achieve automatic reconstruction of the
behavior, performance degradation behavior, etc., of the virtual model. manufacturing system. Liu et al. [62] proposed a biological mimicry-
The behavioral models can usually be constructed using Markov chains, inspired digital twin modeling method. The proposed digital twin
neural networks, and other methods. The rule model describes the laws simulation model simulates the physical process from three aspects:
of change in historical data and the standards and guidelines in related geometry, behavior, and environment. In addition, the digital twin
fields. simulation model has adaptive variation characteristics that enable the
These four models depict physical entities from the perspective of synchronization of workpiece changes during machining. A six-layer
structural dimensions, physical properties, rules, and evolutionary digital twin model is proposed, conceptually classifying digital twin
development, which is very comprehensive. Dividing the different components according to their responsibilities and collaborative ways
properties of physical entities is conducive to accurate modeling of [63]. These six layers are the consumption layer, the service layer, the
virtual models. reasoning layer, the persistence layer, the ingestion layer, and the
physical layer. Data signals are generated by the physical layer, absor
4.2.2. Modeling method variants bed by the ingestion layer, stored in the persistence layer, computed at
On the basis of the geometric model, physical model, behavior the inference layer, provided by the service layer as a data service, and
model, and rule model, some new methods for virtual model modeling consumed at the consumer layer, ultimately providing the user with
insights.
4.2.3. Domain-specific modeling methods

The modeling methods mentioned above refer to the general tech
niques of virtual model modeling and are applicable in different fields.
Many scholars have focused on applying digital twins in a particular
area, proposing virtual model construction methods suitable for that
field.
Modeling methods in human-robot collaborative manufacturing
systems. Fan et al. [64] discussed digital twin modeling of the life cycle
of human-cyber-physical manufacturing systems (HCPS) and proposed a
digital twin concept of “Geometry-History-Object-Snapshot-Topology”
(GHOST). The main elements in GHOST include Geometric information
(GDT), Historical samples (HDT), Object collection (ODT), Snapshot
collection (SDT), and Topological constraint (TDT). GDT refers to the
geometric information of a physical asset. HDT refers to historical
Fig. 7. Virtual model component paradigms. example information about the service processes of a production system.
7
ODT describes the relevant properties and information of the operating approaches in different domains shows that although the modeling ap
objects in the manufacturing system. SDT refers to the collection of proaches in different domains are different, they are all based on the
physical data and virtual data. TDT comprises the system hierarchy and virtual model paradigm, i.e., geometric model, physical model, behav
relation tree of each object in a manufacturing system. This article ioral model, and rules model. Therefore, this paper summarizes the
presents GHOST as a universal reference method for the digital twin general steps when constructing the virtual model, as shown in Fig. 8.
model construction of flexible manufacturing systems. Firstly, the analysis is performed to determine the functions the virtual
Modeling methods in CNC machine tools. Luo et al. [65] proposed a model should have. Secondly, determine whether the virtual model
multi-field unified modeling method for CNC machine tools (CNCMT) paradigm meets the required model functions. If it cannot be satisfied,
based on digital twins, which include a sensing system, digital space, the corresponding functional model is added. If the functionality of some
and a mapping model. In the digital space, the digital twin of the CNC models of the paradigm is useless, the paradigm model is modified, or
machine tool (DTMT) consists of DTMT descriptive model and the DTMT the model is deleted. For example, in the modeling of human–machine
algorithm model. The primary function of the description model is to collaborative manufacturing systems, the function of Topological
describe the geometric, physical, and electrical properties of the constraint (TDT) is added to the model to reflect the system hierarchy
CNCMT. The DTMT algorithm model stores and analyzes the operational and relationships between different objects in the manufacturing
state data and then uses machine learning algorithms to make decisions system.
on the running state adjustment of the CNCMT. Due to the excellent
encapsulation ability of CNCMT objects, the object-oriented modeling 4.2.4. Dynamic update of virtual models
method is adopted for the description and algorithm models. The The working state of physical entities is constantly changing. In order
modeling language used is Modelica, a multi-domain unified modeling to make predictions about the operational state of physical entities, the
language with object-oriented characteristics. This paper establishes a virtual model must have the ability to update the model parameters
multi-domain unified digital twin modeling method and explores the continuously. From the literature, it is found that using Bayesian net
mapping strategy of physical space and digital space. The established works to give virtual models the ability to update dynamically is a more
digital twin model effectively reduces the probability of sudden failure applied method.
of CNCMT and improves stability. Dynamic Bayesian networks have shown good application in fault
Modeling methods in product manufacturing. Zhang et al. [66] first diagnosis and prediction, which can track and correct the evolution of
explored the Product Manufacturing Digital Twin (PMDT), which fo time-varying variables and predict the probability of future failures. Ye
cuses on the production phase of the intelligent shop-floor. The PMDT et al. [13] proposed a digital twin framework for spacecraft structural
consists of five models: Product Definition Model (PDM), Geometry life tracking. The proposed framework consists of two parts: the offline
Model (GSM), Manufacturing Attribute Model (MAM), Behavioral Rule phase and the online phase. Under this framework, the crack propaga
Model (BRM), and Data Fusion Model (DFM). PDM refers to product tion model with uncertain parameters is integrated with the observed
design and manufacturing information such as bill of materials, and crack data using the dynamic Bayesian network. The model parameters
surface roughness. GDM refers to the geometry and shape of intelligent and crack length predictions are updated. Li et al. [67] used the dynamic
workshop components, such as the machine tool’s the length/height/ Bayesian network to establish a probabilistic model for aircraft wing
width. MAM describes the non-geometric properties of smart shop-floor fault diagnosis and realized the accurate diagnosis and prediction of
elements, such as energy consumption, product quality, processing cracks. Yu et al. [68] proposed a digital twin model based on a
methods, etc. BRM describes the behavior and rules of intelligent shop- nonparametric Bayesian network to represent the dynamic degradation
floor elements. Behavior includes the operation activities of workers and process of health conditions and the propagation of cognitive uncer
the work activities of machinery and equipment, and the rules have tainty. A real-time model update technique is provided, which improves
material energy consumption constraints, work process constraints, etc. the model’s flexibility by using enhanced mixed models of Gaussian
Add rule knowledge to modeling to promote the decision analysis ca particles and Dirichlet processes.
pabilities of behavioral models. DFM refers to the data storage and as
sociation method of individual elements of the smart shop-floor.
4.3. Twin data
Compared with the digital twin five-dimensional model, this paper di
vides the composition of the digital twin model more accurately, and the
The sources and categories of twin data are diverse and can generally
proposed PMDT modeling method pays more attention to the descrip
tion of complex element attributes and correlation relationships in the be divided into physical and virtual data [69]. Physical data refers to the
data generated during the work of a physical entity, which is collected
workshop, which is more in line with the needs of each element in the
workshop modeling. by different sensors. Virtual data is generated during the operation of a
virtual model. Because physical and virtual data are different formats, it
The analysis of the above-mentioned virtual model modeling
is essential to study data exchange protocols that link physical and
Fig. 8. General steps of virtual model construction.
8
virtual data. At the same time, the data acquisition method and pro exchange to meet real-time and reliability requirements. The data on the
cessing process are also the difficulties of twin data research. This sec machine dashboard can be stored in the database in real-time, while the
tion provides an overview of data acquisition methods and data data stored in the database can also be accessed and displayed on the
exchange protocols. digital dashboard.
4.3.1. Data acquisition methods 4.3.2. Data exchange protocol

This section summarizes several methods of collecting twin data The operational data of physical entities collected by sensors cannot
obtained through literature reading. be used directly in the virtual environment due to different data formats
Sensors are the most commonly used to collect data. Kong et al. [70] [73,74]. At the same time, the virtual data produced by the virtual
proposed a data construction method that provides stable and efficient model cannot be transferred directly to the physical entity. The data
data support for applying digital twin systems. The shop-floor digital exchange protocol is the key to realizing virtual and real data interac
twin data construction framework consists of four parts: the data rep tion, and a proper data interaction protocol must be chosen [75].
resentation, the data collection, the data organization, and the data Currently, the leading transmission technologies include mainstream
management, as shown in Fig. 9. The data representation module refers industrial bus protocols: Industrial Internet, Profinet, Profibus, Ethernet,
to the hierarchical representation of manufacturing data and the char EtherCAT, etc. In addition, the general wireless transmission method
acter representation of the application. The data organization module can be divided into WIFI, Bluetooth, Zigbee, and so on [76]. Table 5
contains data customized processing and Data pre-processing. The data summarizes the data conversion protocols in these papers, and several
management module consists of a database and the corresponding are analyzed.
storage and retrieval policies. First, the raw data is collected by the TCP/IP protocol is a connection-oriented, reliable, byte-stream-
sensor, DAQ (data acquisition), and RFID (radio frequency identifica based transport layer communication protocol that can accommodate
tion) under the guidance of the data representation module. After that, hierarchical hierarchies that support multi-network applications. Bai
the data is processed in the data management module, and the cleaned et al. [83] proposed a data transmission principle. The sensor signal is
data is stored in the database. Finally, the appropriate information is collected by the data collection card and sent to the computer via the
sent to the top-level application through data customization processing. USB port. The computer/other network devices then send the data to the
Blockchain can be used for digital twin data acquisition. Huang et al. cloud server via the TCP protocol, which saves the data as a data file.
[71] proposed a method for obtaining data for product digital twins Finally, the data are read, updated, analyzed, and calculated in real-time
based on blockchain technology. The product life cycle data is recorded by MATLAB to realize the information interaction between the physical
on the blockchain, and the digital twin data at a specific time can be entity and virtual model, and the principle is shown in Fig. 11.
queried on the blockchain. In addition, a shared network is built where OPC UA is an open standard for horizontal communication from
each participant can send data directly over the network to the machine to machine and vertical communication from device to the
requestor, which improves the efficiency of data sharing. cloud, which provides a beneficial basis for digitalization. Luo et al. [65]
Programmable Logic Control (PLC) could receive and send data from proposed a multi-domain unified modeling method of a digital twin for
the machine. Wang et al. [72] detailed how to obtain the data of con the CNC machine tool, which includes a sensing system, digital space,
ventional devices, as shown in Fig. 10. The operational data is read in and a mapping model. The mapping method was designed according to
real-time by PLC and displayed on the human–machine interface (HMI). the OPC UA standards in this research. The mapping model server
The PLC and the human–machine interface are connected via CN1 cable. consists of four layers, as shown in Fig. 12. The physical interface layer
The digital dashboard receives data from the HMI, and the HMI supports implements the compatibility of different interfaces such as RS485,
Ethernet to exchange data with the digital dashboard. The communi RS232, Wi-Fi, Bluetooth, and Can. The protocol driver layer provides a
cation protocol between the digital dashboard and the human–machine standardized read-and-write technique for various interfaces. The data
interface is Modbus TCP/IP, which uses the Ethernet protocol for data parsing layer is responsible for data processing for various
Fig. 9. Framework of data construction method for workshop digital twin [70].
9
Fig. 10. Data acquisition process.
on the information model created by the agent, the data is interpreted

Table 5
and converted, and the converted information is integrated into the
Partial data exchange protocol.
terminal or database for the customer’s use.
Ref. Protocol name Description
[72] TCP/IP A transmission control/network protocol, 5. Digital twin application scenarios

also known as a network communication
protocol. It is the most basic communication
protocol in the use of the network.
With intelligent manufacturing, all walks of life are developing in the
[77] MQTT (Message Queue A “lightweight” communication protocol direction of intelligence, digitalization, and visualization, and digital
Telemetry Transport) based on the publish/subscribe model, built twin technology is the key technology to achieve this goal. After 20 years
on the TCP/IP protocol. The biggest of development, the concept and technology of digital twins have
advantage of MQTT is that it can provide
gradually improved, and the application fields have extended from
real-time and reliable messaging services for
connecting remote devices with very little manufacturing to the service industry, agriculture, commerce, etc. In
code and limited bandwidth. order to analyze the application of digital twins in various fields, 412
[78] AMQP (Advanced Message An application-tier standard advanced articles were first screened according to whether the research content
Queuing Protocol) message queuing protocol that provides was related to digital twin applications, and 193 articles were obtained.
Unified Messaging services, is an open
standard for application-layer protocols
Secondly, since there are only 1 or 2 documents in some fields, 193
designed for message-oriented middleware. documents are divided with 3 documents as the threshold, and 16
[65] OPC UA (OPC Unified OPC UA is an OPC unified architecture that application areas are finally obtained. It is shop-floor, city, product
Architecture) can better realize factory-level data design, predictive operations, robotics, monitoring, fault diagnosis,
acquisition and management, historical data
manufacturing systems, logistics transportation, medical, assembly
access, and interactive communication of
complex data. management, aerospace, power plant, scheduling systems, CNC ma
[79] oneM2M A set of specifications for machines-to- chine tools, and quality control. The number of articles in these 16 areas
machines in the Internet of Things that is shown in Fig. 14.
enable all devices to communicate with
M2M application servers by creating a
common M2M service layer.
5.1. Application classification of digital twin
[80] XMPP (Extensible An xml-based instant messaging transport
Messaging and Present protocol
Protocol) As can be seen from Fig. 14, the number of articles published in
[81] MTConnect A data transmission protocol for CNC workshop, city, and product design is relatively large, with 25, 22, and
machine tools 20 articles, respectively, which is a hot area of digital twin research at
[82] Automation ML An XML-based automated description
(Automation Markup language designed to interconnect various
present. In order to find the characteristics of the digital twin in different
Language) engineering tools application areas and to obtain general rules, this section analyzes and
compares the application purposes and model characteristics in different
areas. And finally divides digital twin application into three phases, i.e.,
communication protocols. The information model layer uses a data the design phase, the operation process phase, and the dynamic inter
mapping dictionary to turn various types of data into information. The action phase, as shown in Fig. 15. Firstly, the digital twin is used in the
data mapping dictionary contains precise information regarding CNCMT design phase to simulate and verify the design results. Secondly, real-
data, such as standard data element definitions, meanings, and allow time optimization and interaction are performed during operation,
able values. such as fault diagnosis and energy consumption optimization. Finally,
MTConnect is an interconnection communication protocol between visualization and dynamic interaction of the working results are realized
CNC equipment and is specifically used for the interconnection of ma to facilitate people’s work.
chine tools. Tong et al. [84] proposed a framework diagram of the data
transformation based on the MTConnect protocol for the CNC system. 5.1.1. Design phase
The framework is shown in Fig. 13. After retrieving the data, the data is Application purpose: The design phase refers to using digital twin
cleaned and pre-processed. The adapter converts the data into a stan virtual models to simulate and verify product design results or to rede
dard format and sends it to the MTConnect agent. the MTConnect sign and optimize the structure of critical components of physical en
adapter can be designed separately in the sensor or CNC system, tities to improve efficiency of physical entities. The main application
providing different sets of interfaces for different data transfers. Based areas are product design [55,85–88], structural design [89–92].
10
Fig. 11. Cloud platform data transmission principle.
The model features: The digital twin model constructed during the
design phase should be able to describe the physical entity or system
accurately and in detail. At the same time, the structural dimensions
need to be identical between the virtual model and the physical entity in
real-time. Therefore, the digital twin model should have the character
istics of accuracy and dynamic consistency. Accurate models can reduce
the accumulation of model errors, thus effectively avoiding the severe
problems caused by iterative amplification of model errors in structural
refinement design.
5.1.2. Operation process phase

Application purpose: The operation process phase refers to the use of
the iterative evolution ability of digital twin models to predict the future
working state of physical entities. The representative application areas
are fault diagnosis [93–98], predictive operations [99–105], medical
[48,106–110], quality control [111–113], and other fields.
The model features: The digital twin model in the operational pro
cess phase should be able to be updated iteratively based on physical
Fig. 12. DTMT mapping strategy. data to improve the decision-making and evaluation capabilities of the
model continuously. Therefore, the digital twin model in the behavioral
domain should have evolvable and reconfigurable features.
Fig. 13. Data transformation under the MTConnect protocol [84].
11
does not mean that a particular field belongs to only a specific stage,
which may involve two phases in a certain area, or all three phases are
included. Therefore, the division of application phases can help to un
derstand the characteristics of the research field and thus determine the
features that the digital twin should have to build the digital twin model
according to the steps in Fig. 8.
5.2. The emerging applications of digital twin
In the course of the literature survey, in addition to the application

areas shown in Fig. 14, some application areas are currently less studied
but are very promising. This section analyzes the current state of
research in these promising application areas and summarizes the gen
eral characteristics of digital twins in these fields.
5.2.1. Digital twin in food

Food process modeling is becoming more mature with the develop
ment of multi-scale, multi-stage, multi-physical methods. The digital
twin is used to increase productivity, reduce waste and improve trace
ability when answering questions related to food security, changing and
diverse needs, climate change, sustainability, and consumer demand.
Digital twin enables monitoring, control, and prediction of food
quality. The freshness of food is the key to food quality, which is related
to the environment in which it is located. Food preserved in low-
Fig. 14. The number of articles published in different areas.
temperature environments has better freshness and less energy loss.
Environmental factors that affect the freshness of food are temperature,
5.1.3. Dynamic interaction phase humidity, light, etc., which can be monitored by wireless sensors and
Application purpose: The dynamic interaction phase refers to the synchronized with the obtained data in the virtual environment so that
digital twin virtual model to visualize the running state of physical en the virtual model of food is in the same environment as real food. In
tities or to optimize and interact in real-time during operation. The order to grasp the decay law of food and predict the decay of food, in
typical areas are workshop [114–121], city [122–127], monitoring addition to environmental impact factors, it is more important to study
[68,128–130], manufacturing systems [131–133], power plants the biochemical degradation reaction inside the food so that the virtual
[134–137], etc. model of food can simulate the biochemical degradation reaction.
The model features: Digital twin models in the dynamic interaction Defraeye et al. [138] developed a digital twin of fruit based on me
phase focus on data interaction and storage methods between virtual chanical modeling. Based on the measured ambient temperature, a
models and physical entities. At the same time, the operation and digital twin model of the food was constructed, simulating the thermal
interaction process of the virtual model should be observed by the user behavior of the mango. At the same time, a temperature-sensing device
to facilitate the user’s in-depth operation of the model. Therefore, the corresponding to the real mango was developed for model verification of
digital twin model in the dynamic interaction phase should be guaran pulp temperature. The thermal behavior of the artificial fruit sensor
teed to have the characteristics of interoperability and visualization. The device is very similar to that of natural fruit, and its temperature is
interactive and visualized digital twin model facilitates the visual distributed within the data range of real fruit. Digital twin mangoes also
monitoring of the working conditions of physical entities and the effi can iteratively evolve, providing predictions for enzyme-driven, tem
cient collection and transmission of virtual and real data. perature-dependent biochemical degradation reactions.
The division of these three application phases provides a reference. It Digital twin facilitates the control of food processing lines. In the case
Fig. 15. Application segmentation and model features.
12
of changing environmental conditions, the digital twin can predict methods to improve oil and gas production efficiency.
whether the quality loss occurs during food processing and where it Digital twin plays an essential role in improving oil and gas pro
occurs. In this case, the digital twin can help reduce food losses, making duction efficiency. The oil and gas production process is multi-tasking
the food processing process greener. Verboven et al. [139] proposed a and complex, involving many disciplines. Establish a digital twin vir
food processing process control method based on digital twins. The tual model of the oilfield to participate in all aspects of the entire life
method includes three main elements: food process operation, virtual cycle, such as early design guidance, monitoring and optimization of the
replica of food process operation, and IoT platform. IoT platforms pro operation process, and fault diagnosis and prediction. Shen et al. [142]
vide the necessary framework and tools for integrated sensor commu established a digital twin model for oil and gas production. The oil and
nication. The data collected by the sensors during food processing is gas production digital twin model includes six parts: physical entity,
used to drive the operation of the virtual model of the food process virtual model, data acquisition, intelligent algorithm, service, and
operation. Running the virtual model of the food process operation and interactive control, and each element is connected and driven by twin
the actual process provides real-time output for process control related data. Case studies show that the efficiency of the oil and gas production
to the production process and operation. model system based on the digital twin has increased by 3 %.
The digital twin can help understand the storage, flow, and unique
5.2.2. Digital twin in agriculture transport mechanisms of unconventional reservoirs. Zhang et al. [143]
Agriculture is a complex and uncertain field. The production process designed a digital twin model of an unconventional reservoir. Numerical
is dynamic because crops are affected by natural conditions such as simulation and analysis of representative mechanisms affecting the
weather, soil, etc. Farmers must always pay attention to the growth of production of unconventional reservoirs such as capillary, dynamic
crops and intervene in emergencies. In addition, farmers need to be adsorption and salinity injection are used by virtual models, and the
responsible for the management and operation of the farm, constantly influence of these mechanisms on the flow phenomenon of oil reservoirs
evaluating production strategies to adapt to changing market is simulated and illustrated by combined with multi-scale algorithms.
conditions.
Digital twin facilitates process management on farms. The use of
5.3. Other digital twin applications
digital twin to plan, monitor, control, and optimize farm management is
currently-one of the directions in which digital twins are presently
In addition to applications in food, agriculture, and oil and gas,
applied in the agricultural field. Using digital twins as a means of farm
digital twin technology is also used in some areas, as shown in Fig. 16.
management, farmers can remotely manage operations based on digital
Although digital twin research in these areas is not much at present, it is
information without relying on direct observation and on-site manual
fascinating and contributes to the diversification of digital twin appli
tasks. This allows farmers to take immediate action when deviations
cation areas.
occur and simulate the effects of interventions using virtual farm models
The digital twin can be used for indoor safety management. Liu et al.
based on real-world data. Verdouw et al. [140] proposed a conceptual
[144] applied digital twins to indoor safety management construction
framework for designing and implementing digital twin agriculture,
and proposed an indoor safety management system framework. The
including a control model based on a common systems approach and an
framework leverages Building Information Modeling (BIM), the Internet
IoT-based implementation model characterized by timeliness, fidelity,
of Things (IoT), and the Internet to build a building interior security
integration, intelligence, and complexity. This allows farmers to monitor
digital twin. BIM provides safety-related building information and 3D
and simulate the effects of interventions and to control objects remotely
geometric models. Indoor operation information is collected in real-time
via actuators.
through IoT sensors, and data is stored on the Internet. SVM automati
Digital twins can be used for the growth process management of
cally classifies and grades indoor hazards through the mining of safety
crops. In agriculture, the digital twin can be applied not only to farm
data. The framework provides a good way for indoor safety management
management but also to control crops’ growth process. Establish digital
to improve the level of intelligence in indoor safety management.
twin crop models, grasp crops’ real-time growth, determine whether
The digital twin can be used to help understand a person’s person
they are malnourished, and provide guidance on crop harvesting.
ality. Sun et al. [145]focus on the digital twin issue to predict users’
Kampker et al. [141] implement the digital twin on potato harvesting.
personalities. In this study, a person’s personality is understood through
When an operator harvests potatoes using a harvester, the operating
the representation of a virtual person, which refers to the content and
parameters of the machine are set based on the previous year’s data and
liking behavior of users on social networks. A method for predicting
each operator’s “intuition”. In this case, the productivity of each device
is highly dependent on the operator’s experience, there is no standard
ization at all, and it is easy to damage the potatoes., Kampker et al. built
a digital twin model of potatoes, a plastic product that weighs and
measures the same as real potatoes and is equipped with sensors to
detect collisions and rotations to reduce the amount of damage to po
tatoes at harvest time. The data is analyzed in real-time on the harvester
at harvest time and presented to the machine user. Operators adjust the
operating machine in real-time based on the data received to increase
the harvester’s productivity and ensure the potatoes’ integrity.
Digital twin agriculture is still in its infancy, with many difficulties.
The environment in which crops are located is dynamic, and how to
ensure the integrity of data is crucial to building digital twin farms. At
the same time, to achieve digital twin agriculture, good network signal
and comprehensive network coverage are necessary, but network
coverage in rural areas is limited, so how to achieve real-time syn
chronization is a problem that needs to be solved.
5.2.3. Digital twin in oil

With the development of digital technology, the oil and gas industry
has also combined a series of digital technologies with traditional Fig. 16. Other application areas for digital twins.
13
individual personality traits based on deep learning is proposed. Predict 6.2. Virtual Model: The core of modeling
the user’s big five personalities by the state content and liking behavior
posted by the user. This work integrates different types of data through Virtual models are the most studied, followed by data and the
digital twins, providing a reference for identifying human personality slightest physical entities. The construction of virtual models is the core
traits. of the digital twin, and the research on virtual models mainly focuses on
The digital twin can be used for the supply chain management. the division of its components and the definition of the functions of each
Saucedo et al. [146] applied digital twins in the supply chain. This study part.
describes a supply chain decision-making tool based on the digital twin The composition of virtual model is generally composed of func
concept. The goal is to share information among supply chain stake tional models that reflect the relevant properties of physical entities and
holders to increase visibility into products and processes. The tool in the functions that the author expects the virtual model to achieve.
tegrates facility siting models, linear mixed integer optimization models, Standard functional models are geometric models, physical models,
and dynamic simulation techniques for hypothetical multi-scenario behavioral models, and rule models. Geometry models describe the
analysis. geometric dimensions of physical entities, and the definition of the
The digital twin can be used for computational efficiency offloading. virtual model composition includes geometric models in most articles.
Dai et al. [147] applied digital twins to the Industrial Internet of Things Physical models refer to the physical properties of physical entities, and
(IIoT), proposing a digital twin network that leverages the real-time some articles combine physical models with geometric models into one
status of monitoring devices and base stations to create a virtual model, while others do not. Behavioral models generally embody the
model of the IIoT. A random computational offloading and resource predictive capabilities of virtual models. The rule model represents some
allocation problem with the optimization goal of long-term energy ef evolutionary laws, expert knowledge, etc. These two models do not have
ficiency minimization is established through a virtual model, which a high appearance rate of geometric and physical models, and re
effectively improves the data processing efficiency of IIoT devices. searchers will modify them according to the application needs of their
The digital twin can be used for LED design. Schans et al. [148] field. For example, change the rule model to Historical samples [64], or
applied the digital twin technique to the LED process and constructed an discard the behavioral model and the rule model and add an algorithmic
LED digital twin model. The results show that using digital twins can model [65].
significantly reduce the design time while maintaining the required In addition to considering virtual models and data, physical entities
accuracy. are also areas that should be focused on in digital twin modeling. Should
The digital twin can be used for mineral resource acquisition. Mak there be a criterion for the selection of physical entities? Can solid
sim et al. [149] proposed a digital twin-based approach for phosphorus physical components that are not related to the purpose of the study be
production and processing, using the digital twin’s virtual simulation discarded? These questions can provide a direction for future research
function to optimize resource consumption during phosphorus produc on physical entities.
tion and processing.
It can be seen from the areas mentioned above that digital twins have 6.3. Digital twin applications: In the conceptual application stage
powerful application capabilities. We can infer from this that digital
twins may be applied to various industries in the future. The reason is The application of digital twins in some traditional intelligent
that in the digital era, the concept and capability of virtual-real inter manufacturing fields has been very mature, and there is a lot of related
action are needed in any industry. research. The emergence of the digital twin has injected vitality into
some common industrial areas and promoted the intelligent trans
6. Observations and recommendations formation of the manufacturing industry.
Due to the excellent decision analysis capabilities, the digital twin is
By analyzing the current research status of the definition, modeling also gradually applied to some new fields, such as the food industry,
methods, and applications of the digital twin, some observations were agriculture, and other non-traditional intelligent manufacturing fields.
obtained, and some suggestions were made. The continuous expansion of digital twin applications reflects the vi
tality of digital twins. However, although the application field of the
6.1. Digital twin Concept: Gradual diversification digital twin is very extensive, most of the application areas are only the
application of digital twin theoretical concepts, and there is still a
The digital twin and the cyber-physical system are different tech certain distance to achieve the guidance, optimization, and prediction of
nologies. With the development of the digital twin, the difference be the working state of physical entities through virtual models. At the
tween the digital twin and the cyber-physical system is becoming more same time, there is a lack of criteria for judging whether a particular
and more evident. At the same time, with the increase of digital twin field is suitable for digital twin applications. Due to the advanced nature
application fields, digital twin definitions have shown diversified of the digital twin concept, researchers in various fields want to apply
development. For example, Hubert et al. [150] proposed a Digital geo digital twin technology to their own areas, creating the illusion that
Twin - a virtual, semantic 3D replica of all elements and objects of the digital twin technology can be used in any field, but this is incorrect. The
city. The authors chose to add the prefix geo to the digital twins to digital twin is a multidisciplinary technology that integrates machinery,
emphasize the author’s concerns about geodetic and geometric aspects computers, etc. When applied to specific applications, it is necessary to
when creating semantic geographic objects for the digital twins. The consider the necessity and feasibility of implementing digital twins
concept of digital twins is an aspect that deserves to be explored in within a particular field.
depth. Physical entities, virtual models, and data are fundamental The future application areas of digital twins will likely be more
components of the digital twin. On this basis, researchers in different extensive after the full implementation of 5G technology. At the same
industries have given a new definition of the digital twin according to time, the advantages of digital twins will be more obvious.
the needs of industrial applications, that is, a variant of the digital twin.
While this is conducive to increasing the diversity of digital twin con 6.4. Other recommendations
cepts, most of the proposed definitions of digital twin are not universal.
Therefore, in future research, whether it is possible to obtain a more 6.4.1. Accelerate modeling and reduce model weight
detailed and general digital twin definition according to the classifica The current research on digital twin modeling only focuses on the
tion of digital twin application scenarios to guide the construction of establishment of the digital twin model, and the model’s construction
digital twin models is a direction worth thinking. speed, the model’s lightweight, and the model’s accuracy are not taken
14
into account. Speeding up modeling and reducing the model’s weight is is explored through some emerging application areas, which boost
conducive to saving human and material resources. Consequently, in the promotion and application of digital twin.
future research, how to speed up modeling and improve the model’s
lightweight degree without affecting the model’s function is a reference Overall, despite the proliferation of digital twins, research on digital
direction for digital twin modeling. twins is still in the early stages of rapid development. From the literature
review, it is clear that due to the diversity of application domains, re
6.4.2. Combination of digital twin and ML/DL searchers have proposed a wide variety of digital twin reference
Artificial intelligence harnesses the power of computers and ma frameworks and developed different digital twin components with
chines that mimic human thinking to solve problems and make de various implementation tools. This indicates that the current research on
cisions. Machine learning (ML) and deep learning (DL) are two enabling digital twins lacks industry consensus and is challenging to conduct
technologies for artificial intelligence. Alexopoulos et al. [151] provided systematic research. Therefore, it is urgent to establish a common soft
a systematic methodology for establishing digital twin-driven ML-based ware platform for multi-domain digital twins that integrates model
AI applications in manufacturing. This method replaces the work of real- construction, testing, and implementation.
world dataset generation by generating a virtual dataset in a virtual
model in digital twins. Virtually created datasets enable efficient ML Declaration of Competing Interest
model development that can further enrich real-world datasets through
DT-CPS communication channels. Creating appropriate training data The authors declare that they have no known competing financial
sets through digital twin virtual models and simulating toolchains to interests or personal relationships that could have appeared to influence
automatically tag datasets to accelerate the training phase in ML/DL and the work reported in this paper.
thus reduce user engagement during training is worth exploring direc
tion. At the same time, virtual models must use a large amount of data to Data availability
achieve iterative evolution, and ML/DL is also based on data-training
network models. Therefore, it is possible to use ML/DL technology to No data was used for the research described in the article.
accomplish the iterative evolution of virtual models.
Acknowledgment
6.4.3. 3D point clouds facilitate virtual model building
The construction of digital twin virtual models currently adopts the This research was supported by the National Natural Science Foun
method of 3D modeling. The structural shape of the virtual model is first dation of China (Nos. 51975431, 51575407).
built on 3D modeling software, such as SolidWorks, 3D MAX, etc. The
virtual model created by this method is highly accurate and matches the References
physical entity well. However, when the physical entity structure
changes, the constructed virtual model cannot dynamically update its [1] F. Tao, Q. Qi, Make more digital twins, Nature 573 (2019) 490–491, https://doi.
org/10.1038/d41586-019-02849-1.
structural dimensions as the physical entity changes. The virtual model [2] C. Gehrmann, M. Gunnarsson, A digital twin based industrial automation and
can only be modified manually, which is time-consuming. 3D point control system security architecture, IEEE Trans. Ind. Inform. 16 (2019) 669–680,
clouds are one of the most important representations of 3D data, and https://doi.org/10.1109/TII.2019.2938885.
[3] L. Zhang, Y. Guo, W. Qian, W. Wang, D. Liu, S. Liu, Modelling and online training
point clouds can retain the original geometric information in 3D space. It method for digital twin workshop, Int. J. Prod. Res. (2022) 1–20, https://doi.org/
has been used in three-bit reconstruction, instance segmentation, and 10.1080/00207543.2022.2051088.
three-dimensional object detection. Therefore, whether it is possible to [4] C. Cimino, E. Negri, L. Fumagalli, Review of digital twin applications in
manufacturing, Comput. Ind. 113 (2019), 103130, https://doi.org/10.1016/j.
obtain the 3D point cloud data of physical entities through lidar, cam
compind.2019.103130.
eras, and other equipment to achieve the construction of digital twin [5] M. Grieves, J. Vickers, Digital twin: mitigating unpredictable, undesirable
virtual models is also a direction worth exploring. emergent behavior in complex systems, in: F.-J. Kahlen, S. Flumerfelt, A. Alves
(Eds.), Transdiscipl. Perspect. Complex Syst., Springer International Publishing,
Cham, 2017, pp. 85–113, doi: 10.1007/978-3-319-38756-7_4.
7. Conclusion [6] M. Shafto, M. Conroy, R. Doyle, et al., Modeling, Simulation, Information
Technology and Processing Roadmap, Technol. Area. 11 (2010).
[7] E. Glaessgen, D. Stargel, The digital twin paradigm for future NASA and U.S. Air
The digital twin is in a rapid development phase, and various
Force Vehicles, in: 53rd AIAAASMEASCEAHSASC Struct. Struct. Dyn. Mater.
research types are arising. At the same time, the digital twin study is Conf. AIAAASMEAHS Adapt. Struct. Conf. AIAA, American Institute of
becoming more profound and specific. This article reviews 117 papers Aeronautics and Astronautics, Honolulu, Hawaii, 2012, doi: 10.2514/6.2012-
related to digital twin concepts, component research methods, and 1818.
[8] M. Grieves, Digital Twin: manufacturing excellence through virtual factory
application areas from 2017 to 2022. And some observations and rec replication, White Paper 1 (2014) 1–7.
ommendations are made. The main contributions of this work are as [9] F. Tao, B. Xiao, Q. Qi, J. Cheng, P. Ji, Digital Twin modeling, J. Manuf. Syst. 64
follows: (2022) 372–389, https://doi.org/10.1016/j.jmsy.2022.06.015.
[10] X. Liu, D. Jiang, B. Tao, G. Jiang, Y. Sun, J. Kong, X. Tong, G. Zhao, B. Chen,
Genetic algorithm-based trajectory optimization for digital twin robots, Front.
1) It clarifies the digital twin’s definition, characteristics, and applica Bioeng. Biotechnol. 9 (2022), 793782, https://doi.org/10.3389/
tion scenarios by analyzing the difference between the digital twin fbioe.2021.793782.
[11] B. Tipary, G. Erdős, Generic development methodology for flexible robotic pick-
and the cyber-physical system. and-place workcells based on Digital Twin, Robot. Comput.-Integr. Manuf. 71
2) It reviewed the research methods of physical entities, virtual models, (2021), 102140, https://doi.org/10.1016/j.rcim.2021.102140.
and twin data. The types and methods of physical information [12] X. Sun, R. Zhang, S. Liu, Q. Lv, J. Bao, J. Li, A digital twin-driven human–robot
collaborative assembly-commissioning method for complex products, Int. J. Adv.
acquisition are summarized. The general steps of virtual model Manuf. Technol. 118 (2022) 3389–3402, https://doi.org/10.1007/s00170-021-
construction based on the virtual model paradigm are presented. The 08211-y.
methods of twin data acquisition and data exchange protocols are [13] Y. Ye, Q. Yang, F. Yang, Y. Huo, S. Meng, Digital twin for the structural health
management of reusable spacecraft: a case study, Eng. Fract. Mech. 234 (2020),
summarized.
107076, https://doi.org/10.1016/j.engfracmech.2020.107076.
3) It summarizes the popular application areas of the digital twin. Ac [14] W. Terkaj, P. Gaboardi, C. Trevisan, T. Tolio, M. Urgo, A digital factory platform
cording to the characteristics of digital twins in different fields, the for the design of roll shop plants, CIRP J. Manuf. Sci. Technol. 26 (2019) 88–93,
application areas of digital twins are roughly divided into three https://doi.org/10.1016/j.cirpj.2019.04.007.
[15] A. Croatti, M. Gabellini, S. Montagna, A. Ricci, On the integration of agents and
phases: the design phase, the operation process phase, and the dy Digital Twins in healthcare, J. Med. Syst. 44 (2020) 161, https://doi.org/
namic interaction phase. The application potential of the digital twin 10.1007/s10916-020-01623-5.
15
[16] M.R. Enders, N. HoÃŁbach, Dimensions of Digital Twin applications - a literature Adv. Manuf. Technol. 97 (2018) 1209–1221, https://doi.org/10.1007/s00170-
review, Comput. Ind. 123 (2020), 103316, https://doi.org/10.1016/j. 018-2001-2.
compind.2020.103316. [43] C. Zhang, W. Xu, J. Liu, Z. Liu, Z. Zhou, D.T. Pham, Digital twin-enabled
[17] D. Jones, C. Snider, A. Nassehi, J. Yon, B. Hicks, Characterising the Digital Twin: reconfigurable modeling for smart manufacturing systems, Int. J. Comput. Integr.
a systematic literature review, CIRP J. Manuf. Sci. Technol. 29 (2020) 36–52, Manuf. 34 (2019) 709–733, https://doi.org/10.1080/0951192X.2019.1699256.
https://doi.org/10.1016/j.cirpj.2020.02.002. [44] Y. Liu, L. Zhang, Y. Yang, L. Zhou, L. Ren, F. Wang, R. Liu, Z. Pang, M.J. Deen,
[18] M. Liu, S. Fang, H. Dong, C. Xu, Review of digital twin about concepts, A novel cloud-based framework for the elderly healthcare services using Digital
technologies, and industrial applications, J. Manuf. Syst. 58 (2021) 346–361, Twin, IEEE Access 7 (2019) 49088–49101, https://doi.org/10.1109/
https://doi.org/10.1016/j.jmsy.2020.06.017. ACCESS.2019.2909828.
[19] T.Y. Melesse, V.D. Pasquale, S. Riemma, Digital Twin models in industrial [45] Y. Ham, J. Kim, Participatory sensing and Digital Twin city: updating virtual city
operations: a systematic literature review, Proc. Manuf. 42 (2020) 267–272, models for enhanced risk-informed decision-making, J. Manag. Eng. 36 (2020)
https://doi.org/10.1016/j.promfg.2020.02.084. 04020005, https://doi.org/10.1061/(ASCE)ME.1943-5479.0000748.
[20] A. Fuller, Z. Fan, C. Day, C. Barlow, Digital Twin: enabling technologies, [46] F. Tao, M. Zhang, Y. Liu, A.Y.C. Nee, Digital twin driven prognostics and health
challenges and open research, IEEE Access 8 (2020) 108952–108971, https://doi. management for complex equipment, CIRP Ann. 67 (2018) 169–172, https://doi.
org/10.1109/ACCESS.2020.2998358. org/10.1016/j.cirp.2018.04.055.
[21] W. Kritzinger, M. Karner, G. Traar, J. Henjes, W. Sihn, Digital Twin in [47] E. Negri, S. Berardi, L. Fumagalli, M. Macchi, MES-integrated digital twin
manufacturing: a categorical literature review and classification, IFAC-Pap. 51 frameworks, J. Manuf. Syst. 56 (2020) 58–71, https://doi.org/10.1016/j.
(2018) 1016–1022, https://doi.org/10.1016/j.ifacol.2018.08.474. jmsy.2020.05.007.
[22] D.J. Wagg, K. Worden, R.J. Barthorpe, P. Gardner, Digital Twins: State-of-the-art [48] M.N. Kamel Boulos, P. Zhang, Digital Twins: from personalised medicine to
and future directions for modeling and simulation in engineering dynamics precision public health, J. Pers. Med. 11 (2021) 745, https://doi.org/10.3390/
applications, ASCE-ASME J. Risk Uncert. Engrg. Sys. Part B Mech. Engrg. 6 jpm11080745.
(2020), 030901, https://doi.org/10.1115/1.4046739. [49] L.D. Xu, L. Duan, Big data for cyber physical systems in industry 4.0: a survey,
[23] I. Errandonea, S. Beltrán, S. Arrizabalaga, Digital Twin for maintenance: a Enterp. Inf. Syst. 13 (2019) 148–169, https://doi.org/10.1080/
literature review, Comput. Ind. 123 (2020), 103316, https://doi.org/10.1016/j. 17517575.2018.1442934.
compind.2020.103316. [50] F. Tao, M. Zhang, Digital Twin shop-floor: a new shop-floor paradigm towards
[24] B. He, K.-J. Bai, Digital twin-based sustainable intelligent manufacturing: a smart manufacturing, IEEE Access 5 (2017) 20418–20427, https://doi.org/
review, Adv. Manuf. 9 (2021) 1–21, https://doi.org/10.1007/s40436-020-00302- 10.1109/ACCESS.2017.2756069.
5. [51] M. Bevilacqua, E. Bottani, F.E. Ciarapica, F. Costantino, L. Di Donato, A. Ferraro,
[25] D.-G.-J. Opoku, S. Perera, R. Osei-Kyei, M. Rashidi, Digital twin application in the G. Mazzuto, A. Monteriù, G. Nardini, M. Ortenzi, M. Paroncini, M. Pirozzi, M.
construction industry: a literature review, J. Build. Eng. 40 (2021), 102726, Prist, E. Quatrini, M. Tronci, G. Vignali, Digital Twin reference model
https://doi.org/10.1016/j.jobe.2021.102726. development to prevent operators’ risk in process plants, Sustainability. 12
[26] B.R. Barricelli, E. Casiraghi, D. Fogli, A survey on Digital Twin: definitions, (2020) 1088, doi: 10.3390/su12031088.
characteristics, applications, and design implications, IEEE Access 7 (2019) [52] Z. Zhang, Z. Guan, Y. Gong, D. Luo, L. Yue, Improved multi-fidelity simulation-
167653–167671, https://doi.org/10.1109/ACCESS.2019.2953499. based optimisation: application in a digital twin shop floor, Int. J. Prod. Res. 60
[27] Y. Lu, C. Liu, K.-I.-K. Wang, H. Huang, X. Xu, Digital Twin-driven smart (2022) 1016–1035, https://doi.org/10.1080/00207543.2020.1849846.
manufacturing: connotation, reference model, applications and research issues, [53] Z. Zhang, Z. Zhu, J. Zhang, J. Wang, Construction of intelligent integrated model
Robot. Comput.-Integr. Manuf. 61 (2020), 101837, https://doi.org/10.1016/j. framework for the workshop manufacturing system via digital twin, Int. J. Adv.
rcim.2019.101837. Manuf. Technol. 118 (2022) 3119–3132, https://doi.org/10.1007/s00170-021-
[28] C. Semeraro, M. Lezoche, H. Panetto, M. Dassisti, Digital twin paradigm: a 08171-3.
systematic literature review, Comput. Ind. 130 (2021), 103469, https://doi.org/ [54] X. Sun, J. Bao, J. Li, Y. Zhang, S. Liu, B. Zhou, A digital twin-driven approach for
10.1016/j.compind.2021.103469. the assembly-commissioning of high precision products, Robot. Comput.-Integr.
[29] Y. Zheng, S. Yang, H. Cheng, An application framework of digital twin and its case Manuf. 61 (2020), 101839, https://doi.org/10.1016/j.rcim.2019.101839.
study, J. Ambient Intell. Hum. Comput. 10 (2019) 1141–1153, https://doi.org/ [55] J.-G. Duan, T.-Y. Ma, Q.-L. Zhang, Z. Liu, J.-Y. Qin, Design and application of
10.1007/s12652-018-0911-3. digital twin system for the blade-rotor test rig, J. Intell. Manuf. (2021), https://
[30] F. Tao, H. Zhang, A. Liu, A.Y.C. Nee, Digital Twin in industry: state-of-the-art, doi.org/10.1007/s10845-021-01824-w.
IEEE Trans. Ind. Inform. 15 (2018) 2405–2415, https://doi.org/10.1109/ [56] P. Evangeline, Anandhakumar, Digital twin technology for “smart
TII.2018.2873186. manufacturing,” in: Adv. Comput., Elsevier, 2020, pp. 35–49, doi: 10.1016/bs.
[31] X. Fang, H. Wang, G. Liu, X. Tian, G. Ding, H. Zhang, Industry application of adcom.2019.10.009.
digital twin: from concept to implementation, Int. J. Adv. Manuf. Technol. 121 [57] R. Ala-Laurinaho, J. Autiosalo, A. Nikander, J. Mattila, K. Tammi, Data link for
(2022) 4289–4312, https://doi.org/10.1007/s00170-022-09632-z. the creation of Digital Twins, IEEE Access 8 (2020) 228675–228684.
[32] F. Tao, H. Zhang, Q. Qi, et al., Ten questions towards digital twin: analysis and [58] S.H. Choi, K.-B. Park, D.H. Roh, J.Y. Lee, M. Mohammed, Y. Ghasemi, H. Jeong,
thinking, Comput. Int. Manuf. Syst. 26 (2020) 1–17, https://doi.org/10.13196/j. An integrated mixed reality system for safety-aware human-robot collaboration
cims.2020.01.001. using deep learning and digital twin generation, Robot. Comput.-Integr. Manuf.
[33] F. Tao, Q. Qi, L. Wang, A.Y.C. Nee, Digital Twins and cyber-physical systems 73 (2022), 102258, https://doi.org/10.1016/j.rcim.2021.102258.
toward smart manufacturing and industry 4.0: correlation and comparison, [59] N. Nikolakis, K. Alexopoulos, E. Xanthakis, G. Chryssolouris, The digital twin
Engineering 5 (2019) 653–661, https://doi.org/10.1016/j.eng.2019.01.014. implementation for linking the virtual representation of human-based production
[34] S. Dai, G. Zhao, Y. Yu, P. Zheng, Q. Bao, W. Wang, Ontology-based information tasks to their physical counterpart in the factory-floor, Int. J. Comput. Integr.
modeling method for digital twin creation of as-fabricated machining parts, Manuf. 32 (2019) 1–12, https://doi.org/10.1080/0951192X.2018.1529430.
Robot. Comput.-Integr. Manuf. 72 (2021), 102173, https://doi.org/10.1016/j. [60] P. Aivaliotis, K. Georgoulias, G. Chryssolouris, The use of Digital Twin for
rcim.2021.102173. predictive maintenance in manufacturing, Int. J. Comput. Integr. Manuf. 32
[35] P. Zheng, A.S. Sivabalan, A generic tri-model-based approach for product-level (2019) 1067–1080, https://doi.org/10.1080/0951192X.2019.1686173.
digital twin development in a smart manufacturing environment, Robot. [61] C. Wu, Y. Zhou, M.V. Pereia Pessôa, Q. Peng, R. Tan, Conceptual digital twin
Comput.-Integr. Manuf. 64 (2020), 101958, https://doi.org/10.1016/j. modeling based on an integrated five-dimensional framework and TRIZ function
rcim.2020.101958. model, J. Manuf. Syst. 58 (2021) 79–93, https://doi.org/10.1016/j.
[36] Y.H. Son, K.T. Park, D. Lee, S.W. Jeon, S. Do Noh, Digital twin–based cyber- jmsy.2020.07.006.
physical system for automotive body production lines, Int. J. Adv. Manuf. [62] S. Liu, J. Bao, Y. Lu, J. Li, S. Lu, X. Sun, Digital twin modeling method based on
Technol. 115 (2021) 291–310, https://doi.org/10.1007/s00170-021-07183-3. biomimicry for machining aerospace components, J. Manuf. Syst. 58 (2021)
[37] K.T. Park, J. Lee, H.-J. Kim, S.D. Noh, Digital twin-based cyber physical 180–195, https://doi.org/10.1016/j.jmsy.2020.04.014.
production system architectural framework for personalized production, Int. J. [63] F. Mostafa, L. Tao, W. Yu, An effective architecture of digital twin system to
Adv. Manuf. Technol. 106 (2020) 1787–1810, https://doi.org/10.1007/s00170- support human decision making and AI-driven autonomy, Concurr. Comput.
019-04653-7. Pract. Exp. 33 (2020), https://doi.org/10.1002/cpe.6111.
[38] P. Wang, W. Liu, N. Liu, Y. You, Digital twin-driven system for roller conveyor [64] Y. Fan, J. Yang, J. Chen, P. Hu, X. Wang, J. Xu, B. Zhou, A digital-twin visualized
line: design and control, J. Ambient Intell. Hum. Comput. 11 (2020) 5419–5431, architecture for Flexible Manufacturing System, J. Manuf. Syst. 60 (2021)
https://doi.org/10.1007/s12652-020-01898-z. 176–201, https://doi.org/10.1016/j.jmsy.2021.05.010.
[39] H. Gill, NSF perspective and status on cyber–physical systems, in: NSF Workshop [65] W. Luo, T. Hu, C. Zhang, Y. Wei, Digital twin for CNC machine tool: modeling and
on Cyber–physical Systems, National Science Foundation, Alexandria, USA, 2006, using strategy, J. Ambient Intell. Hum. Comput. 10 (2019) 1129–1140, https://
pp. 16–17. doi.org/10.1007/s12652-018-0946-5.
[40] L. Wright, S. Davidson, How to tell the difference between a model and a digital [66] H. Zhang, G. Zhang, Q. Yan, Digital twin-driven cyber-physical production system
twin, Adv. Model. Simul. Eng. Sci. 7 (2020) 13, https://doi.org/10.1186/s40323- towards smart shop-floor, J. Ambient Intell. Hum. Comput. 10 (2019)
020-00147-4. 4439–4453, https://doi.org/10.1007/s12652-018-1125-4.
[41] K.M. Alam, A. El Saddik, C2PS: a Digital Twin architecture reference model for [67] C. Li, S. Mahadevan, Y. Ling, S. Choze, L. Wang, Dynamic Bayesian network for
the cloud-based cyber-physical systems, IEEE Access 5 (2017) 2050–2062, aircraft wing health monitoring Digital Twin, AIAA J. 55 (2017) 930–941,
https://doi.org/10.1109/ACCESS.2017.2657006. https://doi.org/10.2514/1.J055201.
[42] Y. Cheng, Y. Zhang, P. Ji, W. Xu, Z. Zhou, F. Tao, Cyber-physical integration for [68] J. Yu, Y. Song, D. Tang, J. Dai, A Digital Twin approach based on nonparametric
moving digital factories forward towards smart manufacturing: a survey, Int. J. Bayesian network for complex system health monitoring, J. Manuf. Syst. 58
(2021) 293–304, https://doi.org/10.1016/j.jmsy.2020.07.005.
16
[69] G. Yu, S. Zhang, M. Hu, Y.K. Wang, Prediction of highway tunnel pavement [95] S. Hu, S. Wang, N. Su, X. Li, Q. Zhang, Digital twin based reference architecture
performance based on Digital Twin and multiple time series stacking, Adv. Civ. for petrochemical monitoring and fault diagnosis, Oil Gas Sci. Technol. – Rev.
Eng. 2020 (2020) 1–21, https://doi.org/10.1155/2020/8824135. D’IFP Energ. Nouv. 76 (2021) 9, https://doi.org/10.2516/ogst/2020095.
[70] T. Kong, T. Hu, T. Zhou, Y. Ye, Data construction method for the applications of [96] Y. Wang, F. Tao, M. Zhang, L. Wang, Y. Zuo, Digital twin enhanced fault
workshop Digital Twin system, J. Manuf. Syst. 58 (2021) 323–328, https://doi. prediction for the autoclave with insufficient data, J. Manuf. Syst. 60 (2021)
org/10.1016/j.jmsy.2020.02.003. 350–359, https://doi.org/10.1016/j.jmsy.2021.05.015.
[71] S. Huang, G. Wang, Y. Yan, X. Fang, Blockchain-based data management for [97] T.G. Ritto, F.A. Rochinha, Digital twin, physics-based model, and machine
digital twin of product, J. Manuf. Syst. 54 (2020) 361–371, https://doi.org/ learning applied to damage detection in structures, Mech. Syst. Sig. Process. 155
10.1016/j.jmsy.2020.01.009. (2021), 107614, https://doi.org/10.1016/j.ymssp.2021.107614.
[72] K.-J. Wang, Y.-H. Lee, S. Angelica, Digital twin design for real-time monitoring – a [98] B.D. Deebak, F. Al-Turjman, Digital-twin assisted: Fault diagnosis using deep
case study of die cutting machine, Int. J. Prod. Res. 59 (2021) 6471–6485, transfer learning for machining tool condition, Int. J. Intell. Syst. (2021)
https://doi.org/10.1080/00207543.2020.1817999. int.22493, https://doi.org/10.1002/int.22493.
[73] W. Luo, T. Hu, Y. Ye, C. Zhang, Y. Wei, A hybrid predictive maintenance approach [99] S. Meraghni, L.S. Terrissa, M. Yue, J. Ma, S. Jemei, N. Zerhouni, A data-driven
for CNC machine tool driven by Digital Twin, Robot. Comput.-Integr. Manuf. 65 digital-twin prognostics method for proton exchange membrane fuel cell
(2020), 101974, https://doi.org/10.1016/j.rcim.2020.101974. remaining useful life prediction, Int. J. Hydrog. Energy. 46 (2021) 2555–2564,
[74] A. Murphy, C. Taylor, C. Acheson, J. Butterfield, Y. Jin, P. Higgins, R. Collins, https://doi.org/10.1016/j.ijhydene.2020.10.108.
C. Higgins, Representing financial data streams in digital simulations to support [100] G. Yu, Y. Wang, Z. Mao, M. Hu, V. Sugumaran, Y.K. Wang, A digital twin-based
data flow design for a future Digital Twin, Robot. Comput.-Integr. Manuf. 61 decision analysis framework for operation and maintenance of tunnels, Tunn.
(2020), 101853, https://doi.org/10.1016/j.rcim.2019.101853. Undergr. Space Technol. 116 (2021), 104125, https://doi.org/10.1016/j.
[75] C. Liu, P. Jiang, W. Jiang, Web-based digital twin modeling and remote control of tust.2021.104125.
cyber-physical production systems, Robot. Comput.-Integr. Manuf. 64 (2020), [101] Y. Wang, W. Ren, Y. Li, C. Zhang, Complex product manufacturing and operation
101956, https://doi.org/10.1016/j.rcim.2020.101956. and maintenance integration based on digital twin, Int. J. Adv. Manuf. Technol.
[76] J. Yan, Z. Liu, C. Zhang, T. Zhang, Y. Zhang, C. Yang, Research on flexible job 117 (2021) 361–381, https://doi.org/10.1007/s00170-021-07350-6.
shop scheduling under finite transportation conditions for digital twin workshop, [102] G. Wang, G. Zhang, X. Guo, Y. Zhang, Digital twin-driven service model and
Robot. Comput.-Integr. Manuf. 72 (2021), 102198, https://doi.org/10.1016/j. optimal allocation of manufacturing resources in shared manufacturing, J. Manuf.
rcim.2021.102198. Syst. 59 (2021) 165–179, https://doi.org/10.1016/j.jmsy.2021.02.008.
[77] P. Wang, M. Yang, Y. Peng, J. Zhu, R. Ju, Q. Yin, Sensor control in anti-submarine [103] L. Liu, X. Zhang, X. Wan, S. Zhou, Z. Gao, Digital twin-driven surface roughness
warfare—a Digital Twin and random finite sets based approach, Entropy 21 prediction and process parameter adaptive optimization, Adv. Eng. Inf. 51 (2022),
(2019) 767, https://doi.org/10.3390/e21080767. 101470, https://doi.org/10.1016/j.aei.2021.101470.
[78] S.H. Khajavi, N.H. Motlagh, A. Jaribion, L.C. Werner, J. Holmstrom, Digital Twin: [104] X. Fang, H. Wang, W. Li, G. Liu, B. Cai, Fatigue crack growth prediction method
vision, benefits, boundaries, and creation for buildings, IEEE Access 7 (2019) for offshore platform based on digital twin, Ocean Eng. 244 (2022), 110320,
147406–147419, https://doi.org/10.1109/ACCESS.2019.2946515. https://doi.org/10.1016/j.oceaneng.2021.110320.
[79] V. Damjanovic-Behrendt, W. Behrendt, An open source approach to the design [105] S. Mi, Y. Feng, H. Zheng, Y. Wang, Y. Gao, J. Tan, Prediction maintenance
and implementation of Digital Twins for Smart Manufacturing, Int. J. Comput. integrated decision-making approach supported by digital twin-driven
Integr. Manuf. 32 (2019) 366–384, https://doi.org/10.1080/ cooperative awareness and interconnection framework, J. Manuf. Syst. 58 (2021)
0951192X.2019.1599436. 329–345, https://doi.org/10.1016/j.jmsy.2020.08.001.
[80] J. Guo, N. Zhao, L. Sun, S. Zhang, Modular based flexible digital twin for factory [106] A. Lal, G. Li, E. Cubro, S. Chalmers, H. Li, V. Herasevich, Y. Dong, B.W. Pickering,
design, J. Ambient Intell. Hum. Comput. 10 (2019) 1189–1200, https://doi.org/ O. Kilickaya, O. Gajic, Development and verification of a Digital Twin patient
10.1007/s12652-018-0953-6. model to predict specific treatment response during the first 24 hours of sepsis,
[81] P.D. Urbina Coronado, R. Lynn, W. Louhichi, M. Parto, E. Wescoat, T. Kurfess, Crit. Care Explor. 2 (2020) e0249.
Part data integration in the Shop Floor Digital Twin: mobile and cloud [107] S. Sinisi, V. Alimguzhin, T. Mancini, E. Tronci, F. Mari, B. Leeners, Optimal
technologies to enable a manufacturing execution system, J. Manuf. Syst. 48 personalised treatment computation through in silico clinical trials on patient
(2018) 25–33, https://doi.org/10.1016/j.jmsy.2018.02.002. Digital Twins, Fundam. Informaticae. 174 (2020) 283–310, https://doi.org/
[82] J. Bao, D. Guo, J. Li, J. Zhang, The modelling and operations for the digital twin 10.3233/FI-2020-1943.
in the context of manufacturing, Enterp. Inf. Syst. 13 (2019) 534–556, https:// [108] J. Corral-Acero, F. Margara, M. Marciniak, et al., The ‘Digital Twin’ to enable the
doi.org/10.1080/17517575.2018.1526324. vision of precision cardiology, Eur. Heart J. 41 (2020) 4556–4564, https://doi.
[83] L. Bai, Y. Zhang, H. Wei, J. Dong, W. Tian, Digital Twin modeling of a solar car org/10.1093/eurheartj/ehaa159.
based on the hybrid model method with data-driven and mechanistic, Appl. Sci. [109] N.K. Chakshu, I. Sazonov, P. Nithiarasu, Towards enabling a cardiovascular
11 (2021) 6399, https://doi.org/10.3390/app11146399. digital twin for human systemic circulation using inverse analysis, Biomech.
[84] X. Tong, Q. Liu, S. Pi, Y. Xiao, Real-time machining data application and service Model. Mechanobiol. 20 (2021) 449–465, https://doi.org/10.1007/s10237-020-
based on IMT digital twin, J. Intell. Manuf. 31 (2020) 1113–1132, https://doi. 01393-6.
org/10.1007/s10845-019-01500-0. [110] S. Sengan, K. Kumar, V. Subramaniyaswamy, L. Ravi, Cost-effective and efficient
[85] B. Schleich, N. Anwer, L. Mathieu, S. Wartzack, Shaping the digital twin for 3D human model creation and re-identification application for human digital
design and production engineering, CIRP Ann. 66 (2017) 141–144, https://doi. twins, Multimed. Tools Appl. 81 (2021) 26839–26859, https://doi.org/10.1007/
org/10.1016/j.cirp.2017.04.040. s11042-021-10842-y.
[86] K. Kannan, N. Arunachalam, A Digital Twin for grinding wheel: an information [111] R. Söderberg, K. Wärmefjord, J. Madrid, et al., An information and simulation
sharing platform for sustainable grinding process, J. Manuf. Sci. Eng. 141 (2019), framework for increased quality in welded components, CIRP Ann. 67 (2018)
021015, https://doi.org/10.1115/1.4042076. 165–168, https://doi.org/10.1016/j.cirp.2018.04.118.
[87] F. Tao, F. Sui, A. Liu, Q. Qi, M. Zhang, B. Song, Z. Guo, S.-C.-Y. Lu, A.Y.C. Nee, [112] D.-J. Cheng, J. Zhang, Z.-T. Hu, S.-H. Xu, X.-F. Fang, A Digital Twin-driven
Digital twin-driven product design framework, Int. J. Prod. Res. 57 (2019) approach for on-line controlling quality of marine diesel engine critical parts, Int.
3935–3953, https://doi.org/10.1080/00207543.2018.1443229. J. Precis. Eng. Manuf. 21 (2020) 1821–1841, https://doi.org/10.1007/s12541-
[88] R. Cupek, M. Drewniak, A. Ziebinski, M. Fojcik, “Digital Twins” for highly 020-00403-y.
customized electronic devices – case study on a rework operation, IEEE Access 7 [113] S. Zhang, C. Kang, Z. Liu, J. Wu, C. Ma, A product quality monitor model with the
(2019) 164127–164143, https://doi.org/10.1109/ACCESS.2019.2950955. Digital twin model and the stacked auto encoder, IEEE Access 8 (2020)
[89] H. Zhang, Q. Liu, X. Chen, D. Zhang, J. Leng, A Digital Twin-based approach for 113826–113836, https://doi.org/10.1109/ACCESS.2020.3003723.
designing and multi-objective optimization of hollow glass production line, IEEE [114] Y. Wang, Z. Wu, Digital twin-based production scheduling system for heavy truck
Access 5 (2017) 26901–26911, https://doi.org/10.1109/ACCESS.2017.2766453. frame shop, Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 236 (2022)
[90] B. Yu, H. Xie, L. Chen, W. Zhao, Z. He, Exploration of Digital Twin design 1931–1942, https://doi.org/10.1177/0954406220913306.
mechanism of the deep in situ rock insulation coring device, Geofluids 2020 [115] K. Xia, C. Sacco, M. Kirkpatrick, C. Saidy, L. Nguyen, A. Kircaliali, R. Harik,
(2020) 1–12, https://doi.org/10.1155/2020/8835085. A digital twin to train deep reinforcement learning agent for smart manufacturing
[91] R. Bambura, M. Šolc, M. Dado, L. Kotek, Implementation of Digital Twin for plants: environment, interfaces and intelligence, J. Manuf. Syst. 58 (2021)
engine block manufacturing processes, Appl. Sci. 10 (2020) 6578, https://doi. 210–230, https://doi.org/10.1016/j.jmsy.2020.06.012.
org/10.3390/app10186578. [116] M. She, Deep reinforcement learning-based smart manufacturing plants with a
[92] Q. Wu, Y. Mao, J. Chen, C. Wang, Application research of Digital Twin-driven novel Digital Twin training model, Wirel. Pers. Commun. (2021), https://doi.org/
ship intelligent manufacturing system: pipe machining production line, J. Mar. 10.1007/s11277-021-09072-0.
Sci. Eng. 9 (2021) 338, https://doi.org/10.3390/jmse9030338. [117] E. Yildiz, C. Møller, A. Bilberg, Demonstration and evaluation of a digital twin-
[93] P.M. Karve, Y. Guo, B. Kapusuzoglu, S. Mahadevan, M.A. Haile, Digital twin based virtual factory, Int. J. Adv. Manuf. Technol. 114 (2021) 185–203, https://
approach for damage-tolerant mission planning under uncertainty, Eng. Fract. doi.org/10.1007/s00170-021-06825-w.
Mech. 225 (2020), 106766, https://doi.org/10.1016/j. [118] K.T. Park, Y.H. Son, S.W. Ko, S.D. Noh, Digital Twin and reinforcement learning-
engfracmech.2019.106766. based resilient production control for micro smart factory, Appl. Sci. 11 (2021)
[94] M.S. Ibrahim, J. Fan, W.K.C. Yung, A. Prisacaru, W. Driel, X. Fan, G. Zhang, 2977, https://doi.org/10.3390/app11072977.
Machine learning and Digital Twin driven diagnostics and prognostics of light- [119] X. Li, L. Wang, C. Zhu, Z. Liu, Framework for manufacturing-tasks semantic
emitting diodes, Laser Photonics Rev. 14 (2020) 2000254, https://doi.org/ modelling and manufacturing-resource recommendation for digital twin shop-
10.1002/lpor.202000254. floor, J. Manuf. Syst. 58 (2021) 281–292, https://doi.org/10.1016/j.
jmsy.2020.08.003.
17
[120] Z. Liu, W. Chen, C. Zhang, C. Yang, Q. Cheng, Intelligent scheduling of a feature- [136] Z. Jiang, H. Lv, Y. Li, Y. Guo, A novel application architecture of digital twin in
process-machine tool super network based on digital twin workshop, J. Manuf. smart grid, J. Ambient Intell. Hum. Comput. (2021), https://doi.org/10.1007/
Syst. 58 (2021) 157–167, https://doi.org/10.1016/j.jmsy.2020.07.016. s12652-021-03329-z.
[121] Q. Bao, G. Zhao, Y. Yu, S. Dai, W. Wang, The ontology-based modeling and [137] C.A. Alves de Araujo Junior, J.M. Mauricio Villanueva, R.J.S. de Almeida, I.E.
evolution of digital twin for assembly workshop, Int. J. Adv. Manuf. Technol. 117 Azevedo de Medeiros, Digital Twins of the water cooling system in a power plant
(2021) 395–411, https://doi.org/10.1007/s00170-021-07773-1. based on fuzzy logic, Sensors 21 (2021) 6737, doi: 10.3390/s21206737.
[122] F. Xue, W. Lu, Z. Chen, C.J. Webster, From LiDAR point cloud towards digital twin [138] T. Defraeye, G. Tagliavini, W. Wu, K. Prawiranto, S. Schudel, M. Assefa Kerisima,
city: clustering city objects based on Gestalt principles, ISPRS J. Photogramm. P. Verboven, A. Bühlmann, Digital twins probe into food cooling and biochemical
Remote Sens. 167 (2020) 418–431, https://doi.org/10.1016/j. quality changes for reducing losses in refrigerated supply chains, Resour. Conserv.
isprsjprs.2020.07.020. Recycl. 149 (2019) 778–794, https://doi.org/10.1016/j.resconrec.2019.06.002.
[123] J. Döllner, Geospatial artificial intelligence: potentials of machine learning for 3D [139] P. Verboven, T. Defraeye, A.K. Datta, B. Nicolai, Digital twins of food process
point clouds and geospatial Digital Twins, PFG –, J. Photogramm. Remote Sens. operations: the next step for food process models? Curr. Opin. Food Sci. 35 (2020)
Geoinformation Sci. 88 (2020) 15–24, https://doi.org/10.1007/s41064-020- 79–87, https://doi.org/10.1016/j.cofs.2020.03.002.
00102-3. [140] C. Verdouw, B. Tekinerdogan, A. Beulens, S. Wolfert, Digital twins in smart
[124] T. Greif, N. Stein, C.M. Flath, Peeking into the void: Digital twins for construction farming, Agr. Syst. 189 (2021), 103046, https://doi.org/10.1016/j.
site logistics, Comput. Ind. 121 (2020), 103264, https://doi.org/10.1016/j. agsy.2020.103046.
compind.2020.103264. [141] A. Kampker, V. Stich, P. Jussen, B. Moser, J. Kuntz, Business models for industrial
[125] D.N. Ford, C.M. Wolf, Smart cities with Digital Twin systems for disaster smart services – the example of a Digital Twin for a product-service-system for
management, J. Manag. Eng. 36 (2020) 04020027, https://doi.org/10.1061/ potato harvesting, Proc. CIRP 83 (2019) 534–540, https://doi.org/10.1016/j.
(ASCE)ME.1943-5479.0000779. procir.2019.04.114.
[126] A. Francisco, N. Mohammadi, J.E. Taylor, Smart city Digital Twin-enabled energy [142] F. Shen, S.S. Ren, X.Y. Zhang, H.W. Luo, C.M. Feng, A Digital Twin-based
management: toward real-time urban building energy benchmarking, J. Manag. approach for optimization and prediction of oil and gas production, Math. Probl.
Eng. 36 (2020) 04019045, https://doi.org/10.1061/(ASCE)ME.1943- Eng. 2021 (2021) 1–8, https://doi.org/10.1155/2021/3062841.
5479.0000741. [143] T. Zhang, Y. Li, J. Cai, Q. Meng, S. Sun, C. Li, A Digital Twin for unconventional
[127] C. Fan, Y. Jiang, A. Mostafavi, Social sensing in disaster city Digital Twin: reservoirs: a multiscale modeling and algorithm to investigate complex
integrated textual–visual–geo framework for situational awareness during built mechanisms, Geofluids 2020 (2020) 1–12, https://doi.org/10.1155/2020/
environment disruptions, J. Manag. Eng. 36 (2020) 04020002, https://doi.org/ 8876153.
10.1061/(ASCE)ME.1943-5479.0000745. [144] Z. Liu, A. Zhang, W. Wang, A framework for an indoor safety management system
[128] F. Piltan, J.-M. Kim, Bearing anomaly recognition using an intelligent Digital based on Digital Twin, Sensors 20 (2020) 5771, https://doi.org/10.3390/
Twin integrated with machine learning, Appl. Sci. 11 (2021) 4602, https://doi. s20205771.
org/10.3390/app11104602. [145] J. Sun, Z. Tian, Y. Fu, J. Geng, C. Liu, Digital twins in human understanding: a
[129] J.-S. Kang, K. Chung, E.J. Hong, Multimedia knowledge-based bridge health deep learning-based method to recognize personality traits, Int. J. Comput. Integr.
monitoring using digital twin, Multimed. Tools Appl. 80 (2021) 34609–34624, Manuf. 34 (2021) 860–873, https://doi.org/10.1080/0951192X.2020.1757155.
https://doi.org/10.1007/s11042-021-10649-x. [146] J.A. Marmolejo-Saucedo, Design and development of Digital Twins: a case study
[130] F.K. Moghadam, A.R. Nejad, Online condition monitoring of floating wind in supply chains, Mob. Netw. Appl. 25 (2020) 2141–2160, https://doi.org/
turbines drivetrain by means of digital twin, Mech. Syst. Sig. Process. 162 (2022), 10.1007/s11036-020-01557-9.
108087, https://doi.org/10.1016/j.ymssp.2021.108087. [147] Y. Dai, K. Zhang, S. Maharjan, Y. Zhang, Deep reinforcement learning for
[131] M. Resman, J. Protner, M. Simic, N. Herakovic, A five-step approach to planning stochastic computation offloading in Digital Twin networks, IEEE Trans. Ind.
data-driven Digital Twins for discrete manufacturing systems, Appl. Sci. 11 Inform. 17 (2020) 4968–4977, https://doi.org/10.1109/TII.2020.3016320.
(2021) 3639, https://doi.org/10.3390/app11083639. [148] M. van der Schans, J. Yu, G. Martin, Digital luminaire design using LED Digital
[132] Q. Liu, J. Leng, D. Yan, D. Zhang, L. Wei, A. Yu, R. Zhao, H. Zhang, X. Chen, Twins—accuracy and reduced computation time: a Delphi4LED methodology,
Digital twin-based designing of the configuration, motion, control, and Energies 13 (2020) 4979, https://doi.org/10.3390/en13184979.
optimization model of a flow-type smart manufacturing system, J. Manuf. Syst. 58 [149] M. Dli, A. Puchkov, V. Meshalkin, I. Abdeev, R. Saitov, R. Abdeev, Energy and
(2021) 52–64, https://doi.org/10.1016/j.jmsy.2020.04.012. resource efficiency in apatite-nepheline ore waste processing using the Digital
[133] P. Stavropoulos, A. Papacharalampopoulos, C.K. Michail, G. Chryssolouris, Twin approach, Energies 13 (2020) 5829, https://doi.org/10.3390/en13215829.
Robust additive manufacturing performance through a control oriented Digital [150] H. Lehner, L. Dorffner, Digital geoTwin Vienna: towards a Digital Twin City as
Twin, Metals 11 (2021) 708, https://doi.org/10.3390/met11050708. Geodata Hub, PFG –, J. Photogramm. Remote Sens. Geoinformation Sci. 88
[134] C. Altun, B. Tavli, H. Yanikomeroglu, Liberalization of Digital Twins of IoT- (2020) 63–75, https://doi.org/10.1007/s41064-020-00101-4.
enabled home appliances via blockchains and absolute ownership rights, IEEE [151] K. Alexopoulos, N. Nikolakis, G. Chryssolouris, Digital twin-driven supervised
Commun. Mag. 57 (2019) 65–71, https://doi.org/10.1109/MCOM.001.1900072. machine learning for the development of artificial intelligence applications in
[135] H.-A. Park, G. Byeon, W. Son, H.-C. Jo, J. Kim, S. Kim, Digital Twin for operation manufacturing, Int. J. Comput. Integr. Manuf. 33 (2020) 429–439, https://doi.
of microgrid: optimal scheduling in virtual space of Digital Twin, Energies 13 org/10.1080/0951192X.2020.1747642.
(2020) 5504, https://doi.org/10.3390/en13205504.
18

Digital Twin in General

Uploaded by

Copyright:

Available Formats

Digital Twin in General

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Digital Twin in General

Uploaded by

Copyright:

Available Formats

Advanced Engineering Informatics 55 (2023) 101876

Contents lists available at ScienceDirect

Advanced Engineering Informatics

A systematic review of digital twin about physical entities, virtual models,

devices to ensure seamless connectivity and data exchange through

Fig. 2. Literature review methodology.

on the digital twin research approach. The research methodology of the

2.2. Literature review methodology

The literature review method is divided into three steps: literature

2.3. Status analysis of digital twin research

After eliminating the irrelevant literature, 117 papers were finally

4.2. Virtual model

A virtual model is a digital mirror image of a physical entity and an

Fig. 5. Distribution of the three parts.

Fig. 6. Physical information perception method.

4.2.1. Modeling method paradigm are proposed.

4.2.3. Domain-specific modeling methods

Fig. 8. General steps of virtual model construction.

4.3.1. Data acquisition methods 4.3.2. Data exchange protocol

Fig. 10. Data acquisition process.

on the information model created by the agent, the data is interpreted

[72] TCP/IP A transmission control/network protocol, 5. Digital twin application scenarios

Fig. 11. Cloud platform data transmission principle.

5.1.2. Operation process phase

Fig. 13. Data transformation under the MTConnect protocol [84].

5.2. The emerging applications of digital twin

In the course of the literature survey, in addition to the application

5.2.1. Digital twin in food

Fig. 15. Application segmentation and model features.

5.2.3. Digital twin in oil

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.