Digital Twin Paradigm Shift: The Journey of the Digital Twin

Definition

Martin Tomczyk

and Hendrik van der Valk

Chair for Industrial Information Management, TU Dortmund University, Joseph-von-Frauenhofer-Straße 2-4, Dortmund,

Germany

Keywords: Digital Twin, Literature Review, Concept Matrix.

Abstract: This paper examines the paradigm shift in the definition of the digital twin in recent years and describes how

the definitions differ from each other. After an extensive literature review and the development of a concept

matrix, it became apparent that a paradigm shift is taking place from the classic three-dimensional definition

– physical and virtual space with a bidirectional connection – to an expanded five-dimensional definition –

data and services as extended dimensions. In particular, the focus and developments in Information and

Communication Technologies lead to the recognition of the dimensions of data and services as an independent

dimension. In addition, further descriptions of the concept of the digital twin were assigned to the known

dimensions.

1 INTRODUCTION

By combining real and virtual space, new

applications are emerging that lead to higher and

more uniform production and quality (Grieves 2014).

Due to the many possible applications, the concept of

Digital Twins (DT) receives a lot of attention in many

different fields in research and practice (Zhao et al.

2019a). Since its introduction in 2003, the concept

has steadily gained interest, and, especially since

2017, the number of annual publications has

increased exponentially (van der Valk et al. 2020). In

the future, the DT will gain an important role in the

industry. Especially through the advances in

information technology (new IT), such as Cloud

Computing, Internet of Things (IoT), Big Data, and

Artificial Intelligence (AI), the merging of the digital

and real-world is gaining influence (Tao et al. 2019).

The theoretical foundations of a DT steam from

different disciplines, such as information science,

production engineering, data science, and computer

science (Tao et al. 2018b). Thus, the term DT is used

in many areas of scientific literature, for instance, in

information science, production engineering, data

science, and computer science, as well as in industry.

However, this usage is often divergent (Dahmen and

https://orcid.org/0000-0002-0767-3246

https://orcid.org/0000-0001-6329-792X

Rossmann 2021; Kritzinger et al. 2018). The

literature has also not yet provided a clear definition

of the construct (Negri et al. 2017; Dahmen and

Rossmann 2021). Most definitions mention

comparable characteristics, but no unified definition

is available. To this end, most definitions show great

differences (van der Valk et al. 2020), which in turn

can be attributed to the variation in the fields of

application (Dahmen and Rossmann 2021; Schleich

et al. 2017). It can therefore be concluded that there

is currently no uniform understanding of the concept

of digital twins (Cimino et al. 2019).

In general, it is very interesting to gain a deeper

insight in how digital twins are defined and which

definition is used to be more common. Hence, in this

paper, we aim to answer the following research

questions (RQ):

RQ1: Which definitions of Digital Twins are the

most commonly used?

RQ2: How do the definitions of Digital Twins

differ?

In the following, chapter 2 describes the

methodology and the research design. In chapter 3,

we provide the results. In section 4, we discuss the

Tomczyk, M. and van der Valk, H.

Digital Twin Paradigm Shift: The Journey of the Digital Twin Deﬁnition.

DOI: 10.5220/0010997600003179

In Proceedings of the 24th International Conference on Enterprise Information Systems (ICEIS 2022) - Volume 2, pages 90-97

ISBN: 978-989-758-569-2; ISSN: 2184-4992

development and results of the concept matrix.

Whereas in chapter 5, an outlook shows further

implications.

2 RESEARCH DESIGN

We use the methodological guidelines of Vom

Brocke et al. (2009) and Webster and Watson (2002)

to conduct a literature review. According to Vom

Brocke et al. (2009), a literature review consists of

five steps – publication specification, topic

conceptualization, literature search, literature

analysis, and revision. First, we specified the

publications, i.e., they have to be published in a

journal, as conference proceedings, as a book chapter,

or as a report. In terms of the topic conceptualization,

the publications should contain the concept of the DT

as a central element to be included in the study.

Furthermore, we used the databases Science Direct,

Scopus, IEEE Xplore, and Digital Library with the

keywords "digital twin", "digital twin application",

and "digital twin service". To keep the quality as high

as possible, we examined the selected articles for their

relevance. We also applied a backward and a forward

search. Since the DT covers an interdisciplinary field

(Schleich et al. 2017; Tao et al. 2018b), one has to

investigate adjacent topics (Webster and Watson

2002). Thus, we conducted a deeper review in

information science, production engineering, data

science, and computer science.

The review resulted in the realization that the

basic definitions of the DT used in the examined

literature reviews show substantial differences. To

generate a conceptual analysis of the DT, we create a

concept matrix according to Webster and Watson

(2002).

At last, it was important to find out which topics

need further investigation and which areas should be

investigated in more detail in the future. All steps

were run through twice. Thus, the literature base

consists of two iterations and reached theoretical

saturation.

Due to space limitations, we have not included the

conceptual matrix with all 54 references in this paper,

but in the sense of transparency and knowledge

accumulation, we provide them and our coding

results in the form of a concept matrix

https://bit.ly/3H1dJiM

3 THE DIGITAL TWIN

CONCEPT

3.1 Classic Approach (Three

Dimensions)

As the father of the DT, Michael Grieves (2014)

introduced an early approach to the DT in 2003,

thereby shaping an initial – classic – view of this

concept. At that time, digitization was limited, so

most product information was collected manually and

was often handwritten on paper. More detailed

execution of the Digital Twin, favored by advancing

digitization, was described nine years later. In a white

paper, Grieves (2014) specifies his former concept

and describes the DT as a virtual representation of a

physical product that contains information about that

product. Shafto et al. (2010) extended the description

of a classical DT in a NASA roadmap paper. Based

on this, Glaessgen and Stargel (2012, p. 7) define the

DT in detail: "A Digital Twin is an integrated

multiphysics, multiscale, probabilistic simulation of

an as-built vehicle or system that uses the best

available physical models, sensor updates, fleet

history, etc., to mirror the life of its corresponding

flying twin. The Digital Twin is ultra-realistic [...]

integrates sensor data [...] maintenance history and all

available historical and fleet data obtained“.

According to this, a DT consists of three

dimensions: (a) the physical product in real space, (b)

the virtual product in virtual space, (c) the

bidirectional data link.

The bidirectional data link connects the virtual

and physical world and delivers data from the

physical to the virtual representation and at the same

time transports information and management

instructions from the virtual representation to the

physical (Grieves 2014; Grieves and Vickers 2017).

Driven by technological developments in recent

years, the virtual space transformed from an initially

simple representation of the real object, mostly

consisting of visual attributes (such as dimensions of

the product) as well as the technical data, to a

multitude of virtual spaces with additional subspaces,

each containing or performing specific operations

such as modeling, testing or optimization (Jones et al.

2020; Grieves 2014). The real space transmits the

collected data to the virtual space, while at the same

time, information and processes are transmitted from

the virtual space to the real space. Whereby this

information can be used profitably to the respective

Digital Twin Paradigm Shift: The Journey of the Digital Twin Deﬁnition

application. This was initially described as a human-

related activity, which is often supported by further

process control systems (Grieves 2005). Taking this

further, new technological developments make an

autonomous or automated process control system

conceivable (Grieves 2014). The bidirectional

connection, especially in the case of real-time

synchronization, leads to the fact that the right

decisions can be made based on the latest information

(Negri et al. 2017). Finally, the virtual space contains

the properties, state, and behavior of the real object

and consists of models and data that can predict the

actual behavior in the operational environment (Haag

and Anderl 2018). Figure 1 illustrates the three

dimensions of the DT and shows the interaction of the

respective dimensions.

Figure 1: Classic Approach (Three-Dimensional) of a

Digital Twin (Grieves 2014).

3.2 Extended Approach (Five

Dimensions)

Tao et al. (2019) extend the classic DT by two

dimensions and introduce a five-dimensional

extended approach (see Figure 2).

The description of the physical entity (1) is

comparable to the physical product (a) in real space,

as well as the virtual entity (2) is comparable to the

virtual product (b) in virtual space. The data collected

in the first two dimensions (3) can provide much

additional information. All data form the core of the

DT (Qi et al. 2018; Tao et al. 2019). This core

provides four functions:

First, it reflects the real-world characteristics,

behavior, and rules of the physical counterpart,

creating an exact duplicate that records all changes.

Second, by extrapolation, an ideal action/handling of

events can be inferred. Third, by prediction, the cause

of problems can be eliminated before they occur, and

fourth, an assessment of performance can be achieved

beforehand (Tao et al. 2019).

Further, a seamless connection of the individual

dimensions (4) represents a critical point because it is

indispensable that all information is transmitted. In

this context, the individual connections can be

optimized iteratively. Here, the various connections

can be accessed independently and thus can be

controlled, edited, or modified as desired. These

connections are comparable to the bidirectional data

connection (c) of the classical view of the DT but

cannot be equated due to a large number of individual

connections (Tao et al. 2019; Tao et al. 2018a; Tao et

al. 2018c). The fifth dimension is defined as the

resulting services. More precisely, this refers to the

assessments, optimizations, predictions, and

validations via a user-defined interface. This also

allows authorized people, or employees, to take

advantage of a DT who have little to no knowledge of

how it works and what mechanisms are hidden behind

the interface (Tao et al. 2019). Through the service

dimension, each additional dimension of the DT can

achieve a higher value, as this allows each dimension

to be used more efficiently. More specifically, Tao et

al. (2019, p. 203) describe it as follows: "Combined

with the services, the DT will be easier for usage and

will generate more acceptable analysis and evaluation

results on product design, manufacturing, usage,

prognostics and health management [PHM], and

other processes.".

Figure 2: Extended Approach (Five-Dimensional) of a

Digital Twin (Tao et al. 2019).

Here, these services do not have to be created or

used. Furthermore, services can be divided into

resource service, data service, and model service, or

according to the product lifecycle, physical unit

service, virtual unit service, and data service (Tao et

al. 2019, 203 & 208-211).

In general, according to Tao et al. (2018a; 2018c),

the DT consists of five equally important dimensions.

It is a highly dynamic concept whose complexity can

increase during its lifecycle.

ICEIS 2022 - 24th International Conference on Enterprise Information Systems

3.3 Investigation of Real Applications

With a focus on production systems, Negri et al.

(2017, p. 946) define the DT as follows: "The [Digital

Twin] consists of a virtual representation of a

production system that is able to run on different

simulation disciplines that are characterized by the

synchronization between the virtual and real system,

thanks to sensed data and connected smart devices,

mathematical models and real-time data elaboration."

In most cases, the DT is a model that represents a

system while incorporating various levels of

simulations (Negri et al. 2017). A DT can be

categorized by its area of focus and the technologies

used. A digital counterpart to the physical object is

identified as the salient commonality. Besides, the

integration of data can lead to a different dimension.

Here, a distinction is made between the data exchange

between the physical and digital counterparts. Some

concepts exchange data automatically in real-time,

whereas others require manual transfers of data

(Kritzinger et al. 2018).

Kritzinger et al. (2018) describe the origin of DT

in the reflection of a product or object in virtual space.

Furthermore, the main aspect is the ability to provide

different information in a uniform format. Here, it

was investigated to which extent data is exchanged

and how a delimitation can be made. Kritzinger et al.

(2018) distinguish between digital model (fully

manual data exchange), digital shadow (one data flow

automated, one manual), and DT (fully automated

data flows).

3.4 Variation or Linking Approach

In 2020, van der Valk et al. (2020) assigned properties

to dimensions of the DT. After a literature review of

233 papers, a total of 18 characteristics were derived,

which can be divided into eight dimensions. Further,

some characteristics may exist in parallel, and some

may not ("exclusivity").

The taxonomy developed by van der Valk et al.

(2020) was enhanced after a thorough interview series

with experts from the industry (van der Valk et al.

2021) and should be seen as a configuration tool for

DT. It provides balanced dimensions with associated

characteristics of the DT construct. The individual

configuration of a DT is highly dependent on the use-

case in mind while designing a DT. Therefore, it aims

to be a non-specific collection of properties.

Nevertheless, certain patterns of commonly used DT

configurations are visible, which leads to the

conclusion that a very common DT contains a

bidirectional data flow, processes data in an identical

model of the physical system. It contains human-

machine interfaces and machine-to-machine

interfaces, receives permanent updates from multiple

data sources (van der Valk et al. 2020).

In order to identify the most frequently mentioned

characteristics of a DT, Jones et al. (2020) conducted

a literature analysis. This revealed a total of 12

characteristics, which describe the DT at its base

(Table 1 and Table 2 are provided in footnote 1

above).

4 DISCUSSION

4.1 Comparison of the Definitions

The literature reviews presented in the foregoing

illustrate how different the views and understandings

of the concepts of DT are. The concepts are used in

information science, production engineering, data

science, and computer science (Tao et al. 2018b) and

are repeatedly described with comparable

characteristics and dimensions. The definition and

concept of DT is ultimately influenced depending on

the target (van der Valk et al. 2020). Further, the

concepts are versatile due to the multitude of research

and application fields. The overlaps and definitions

listed above can generally be compared with each

other, which is why comparable statements of

individual definitions can be divided into five

superordinate areas (see Table 3).

The physical product in real space (a), the

physical entity, and the combination of the physical

entity (I) and the physical environment (III) can be

equated. All three descriptions are to be understood

as a physical part, object, or product, which is

tangible, i.e., physically existent (Grieves 2005, 2014;

Tao et al. 2018b; Jones et al. 2020).

The same approach can be applied to the second

dimension. This describes the virtual product in

virtual space (b), the virtual entity (2), as well as the

combination of the virtual entity (II) and the virtual

environment (IV), summarizing as the counterpart of

the physical ones. Through certain computer systems,

models and simulations can be created in a controlled

manner in a wide variety of virtual spaces. In these

spaces, different scenarios can be extrapolated, which

gain added value, especially by using the virtual

environment for optimization, controlling, or control

(Grieves 2005, 2014; Jones et al. 2020). Additionally,

it should be mentioned that the designation of the

virtual environment is determined by the underlying

technology, such as database, data warehouse, cloud

platform, server, or application programming

Digital Twin Paradigm Shift: The Journey of the Digital Twin Deﬁnition

interface (API) (Jones et al. 2020). In this context, van

der Valk et al. (2020) formulated the quantity of

model accuracy (D4). This can be classified as either

identical or partially identical and represents how

accurately the mapping of the real entity with its

environment is to be classified as a virtual likeness. A

partially identical mapping may result in a smaller

amount of data being processed and may therefore be

easier to implement but contain the risk of not

achieving the desired results. Kritzinger et al. (2018)

also introduced a granularity, which focused on the

flow of information between the physical and virtual

entity (Kritzinger et al. 2018).

The third dimension is the link between the first

two dimensions. This is referred to as (c) the

bidirectional data link or (4) a seamless link (Grieves

2005, 2014; Tao et al. 2018a; Tao et al. 2018b). Jones

et al. (Jones et al. 2020) consolidated separate

characteristics for this. These are composed of the

(VIII) physical-virtual connection, the (IX) virtual-

physical connection, and the (X) twinning, or

twinning rate. Comparable characteristics were

elaborated by van der Valk et al. (2020). Here, Data

Link (D1), Conceptual Elements (D3), and (D6)

Synchronization are equivalent to the description of

data links earlier (and by Grieves 2005, 2014 and Tao

et al. 2018b; Tao et al. 2018a). If either VIII or IX is

present, the DT has a one-way connection between

the first and second dimensions. After Kritzinger et

al. (2017), it can be considered as a Digital Shadow.

However, if both connections (VIII and IX) are

present, then it is a bidirectional connection and is

examined as DT. Here, it is possible to infer from one

explanation - of VIII and IX - to the other -

unidirectional or bidirectional data connection - and

vice versa. If there is a connection, then the

conceptual elements will be physically linked to each

other. If, again, there is no connection, there is

independence between the physical and virtual

entities. Similarly, a twinning process (associated

twinning rate) exists only if a (VIII) physical-to-

virtual connection and a virtual-to-physical

connection exist (Jones et al. 2020; Tao et al. 2018b;

Tao et al. 2019, p. 4).

Furthermore, this is the basis of synchronization,

which cannot take place if there are no connections

between the units to be synchronized. In addition,

Synchronization (D6) addresses whether continuous

synchronization takes place over the course of

(work/process) time, i.e., over the course of the

product life cycle (van der Valk et al. 2020). This is

the essential aspect already addressed by Grieves

(2005) and presented as a benefit for PLM. Tao et al.

(2018a) also describe accurate synchronization and

fidelity to perform real-time analyses that lead to the

best possible results. Similarly, Jones et al. (2020)

describe (VI) fidelity as an important characteristic of

the data link, which translates to the number of

parameters, their accuracy, as well as the degree of

abstraction that occurs between the virtual and

physical twin, or environment. Together, the

aforementioned dimensions/characteristics describe

the data link between the real space in the real

environment and the virtual space in the virtual

environment.

Grieves (2014) describes the Digital Twin as a

three-dimensional construct that includes the

Table 3: Comparison of the presented breakdowns.

three-dimensional construct

(Grieves & Vickers)

Extended Approach

(Five-dimensions)

(Tao et al.)

twelve Characteristics

(Jones et al.)

Taxonomy

(van der Valk et al.)

(a) the physical product

in real space

(1) the physical entity (I) physical entity

(III) physical

environment

(D4) model accuracy

(b) the virtual product in

virtual space

(2) the virtual entity (II) virtual entity

(IV) virtual environment

link

(4) a seamless link (VIII) physical-virtual

connection

(IX) virtual-physical

connection

(X) twinning, or twinning

rate

(

)

fidelit

(D1) Data Link

(D3) Conceptual Elements

(D6) Synchronization

(3) collected data (V) parameters

(VII) state

(D7) data input

(5) resulting services (XI) physical process

(

XII

)

virtual

rocess

(D2) purpose

(

)

Interface

ICEIS 2022 - 24th International Conference on Enterprise Information Systems

elements explained above. Unlike Tao et al. (2018a),

he did not refer to the data collected, processed, or

generated by the DT as a separate dimension.

However, there is a fluid boundary or transition. Tao

et al. (2018a) describe that the (3) collected data leads

to a continuous improvement in self-development, as

the data is exchanged between the first two

dimensions automatically and in real-time (here, an

ideal case is assumed, which define the dimensions

described in advance). Furthermore, data is presented

as the central element of the DT. They built the core

and have data connections to each associated part

(Tao et al. 2018b). The data lying in it is referred to

as (V) parameters by Jones et al. (2020), which

characterizes the nature of the data, its information,

and the processes that occur between the real space in

the real environment and the virtual space in the

virtual environment. In addition, the (VII) state

addresses the current values of the measured

parameters (Jones et al. 2020).

According to van der Valk et al. (2020), the data

input (D7) can come in as raw data, e.g., directly from

sensors or as already processed data. In the big

picture, these are different ways to characterize the

data within the DT. Here, the processed data can

come from a physical (XI) or virtual process (XII).

This is defined by the foreseen purpose (D2) of the

DT. The purpose can ultimately include various

services that can improve the convenience, reliability,

and productivity of a technical system, among other

things (Tao et al. 2018b). Such a service can be

transmitted via IoT, from machine-to-machine

interfaces (M2M), or be carried via human-to-

machine interfaces (HMI), which can perform further

work steps (van der Valk et al. 2020). The interface

(D5), in turn, displays the state of the data or

parameters and can provide a service by highlighting

changes in the data. Overall, the services of the DT

are very versatile and can therefore not be easily

concretized, which is why the two dimensions around

the data and services are characterized by a smooth

transition.

4.2 Concept Matrix

In order to identify which definition is most

frequently used, a concept matrix was created. Due to

the limited space, the concept matrix is provided

through the above link.

The DT is a complex model that is a digital copy

of various entities and can be extended arbitrarily and

independently of other entities. Additionally, it is

automated and globally available in real-time

(Aivaliotis et al. 2019). Among other things, this can

create a high level of transparency (Kampker et al.

2019). This can lead to significant advantages in the

development, production, maintenance, and servicing

of products, especially in the case of complex

products such as aircraft or electrical systems,

manufacturing, and supply chain management (Tao et

al. 2018a). Fittingly, Enders and Hoßbach (2019, p.

7) define the DT as follows: "A Digital Twin is a

virtual representation of a physical object called a

Physical Twin. The Physical and the Digital Twin

may be connected to each other. A Digital Twin can

provide more information about its Physical Twin

than the Physical Twin itself can provide." In the

literature review, the two authors referred to the

classic approach of the DT. The result of the

investigations of the concept matrix shows that this is

the most widespread and thus most common view of

the DT. Here, the definition according to Grieves

(2014) describes a total of three dimensions that make

up a Digital Twin. At 37.04%, more than one-third of

the articles studied use this definition as the

foundational knowledge for the DT. Almost a similar

number of articles (35.19%) use the definition of

Glaessgen and Stargel (2012). This definition does

not use specific dimensions but describes a DT with

some characteristics that are to be valued in general

and used especially in the context of spaceflight.

Since the Digital Twin has its origin in NASA's

Apollo missions, this is a popular reference point for

many authors. The third most referenced definition,

at 33.33%, is that of Tao et al. (2018a, p. 3566). This

describes a total of five dimensions and was

explained as the extended approach. In this view of

the DT, the focus is on Data and is further expanded

by the services.

In the previously described determination of the

most common definitions, the frequency of citations

was implied as to the basis. If the data are now

adjusted with respect to publication years, the result

is in a slightly different order. The reason for

adjusting the order is as follows. An article published

in 2018 cannot have been referenced as basic

knowledge in a paper from 2015, so it may not be

relevant for the percentage calculation. In the

adjusted order, a trend towards the extended approach

of Tao et al. (2018a) of the DT can be identified.

Although it can be criticized that a lower total number

of contributions in the calculation leads to a higher,

percentage share, over the years the interest in the

construct of the DT increased more and more.

Furthermore, constructs such as for data generation,

Big Data, data mining, etc., are becoming more and

more in focus (Tao et al. 2019, p. 19; Qi et al. 2018).

Therefore, an overall trend can be identified, which

Digital Twin Paradigm Shift: The Journey of the Digital Twin Deﬁnition

puts data at the core. At 66.67%, more than half of the

published contributions as of 2018, take their

foundational knowledge from Tao et al. (2018a).

Furthermore, 40.82% also use the classic approach

according to Grieves (2014), and 38.00% the

definition is referring to Glaessgen and Stargel

(2012).

Thus, it is evident that there is an evolution in the

construct of the DT. From the classic three-

dimensional approach according to Grieves (2014) to

an extended five-dimensional view according to Tao

et al. (2018a). Furthermore, it can be noticed that

technological developments allow a DT to take on

different sizes. On the one hand, it can be a single

small component, and on the other hand, it can

represent an entire production line, with different

sizes and elements, even a complete supply chain.

This leads to the dissolution of previous limitations

and expands the possible applications (Qi et al. 2018,

p. 238). This brings the general definition, according

to Boschert and Rosen (2016), back to the forefront,

allowing the construction of the DT to be applied in

many areas. The authors refer to the product lifecycle

(PLC) and define: "The vision of the Digital Twin

itself refers to a comprehensive physical and

functional description of a component, product or

system, which includes more or less all information

which could be useful in all - the current and

subsequent - lifecycle phases." (Boschert and Rosen

2016, p. 59). In each phase of the PLC, more data is

generated that represents one's own state as well as

the state of the environment (Boschert and Rosen

2016; Haag and Anderl 2018). The underlying idea of

the DT is that it should act as an information construct

via the PLC (Grieves and Vickers 2017).

Finally, by shifting the perception of the DT from

the classic approach of Grieves (2014) to the

extended approach of Tao et al. (2018a; 2018b), the

focus is primarily on data. This concept can ensure

information availability and traceability across

lifecycle phases (Ströer et al. 2018). By having data

at the core (Qi et al. 2018), increased services can be

provided throughout PLM (Tao et al. 2019, p. 11).

Through intelligent data integration and data

processing across all product lifecycles, the DT can

lead to increased performance of the product, or

processes, in the physical space (Qi and Tao 2018).

5 CONCLUSIONS

This paper reviews different definitions and influence

streams of digital twins in the literature. From this

database, mainly influenced by information science,

production engineering, data science, and computer

science, we have shown that there are three streams

of literature that coin the different types of definitions

(RQ1):

- the classic approach – following Grieves

(2014) and Glaessgen and Stargel (2012) –

with physical and virtual products, as well as

data flows

- the extended approach – following Tao et al.

(2019) – who extends the classic approach by

services and collected data

- the applicational approach, which forms its

definitions by the portrayed use cases

To summarize, the most used definitions are in

descending order: Grieves (2014), Glaessgen and

Stargel (2012), and Tao et al. (2019).

To answer RQ2, a concept matrix was developed.

It provides an overview of the differentiation between

the definitions. Certainly, the development of recent

works shows a divergence towards more complex and

more detailed definitions and supports the extended

approach. Additionally, the definition of DT develops

likewise technological improvements are made and

adapt to the requirements of the user.

Our work is subject to certain limitations. As the

analysis of the literature base is influenced by

subjective decisions, other researchers might define

different scopes and, hence, might gain other results,

especially with regards to other domains.

This paper provides several contributions. As

scientific contributions, this paper analyzes

definitions and streams which influence them through

a thorough review. It provides a deep insight into the

explanation of digital twins and contributes to the

knowledge about digital twins. This brings the

opportunity for researchers and furthermore

managerial contributions for practitioners to fully

understand the field of digital twins.

As possible further research, the more and more

common real-world applications are worthy for a

deeper analysis. The comparison between theoretical

works and operational facts might provide interesting

insights on the implementation of digital twins.

REFERENCES

Aivaliotis, P., Georgoulias, K., Alexopoulos, K. (2019).

Using digital twin for maintenance applications in

manufacturing: State of the Art and Gap analysis. In

IEEE International Conference on Engineering,

Technology and Innovation (ICE/ITMC). p. 1–5.

Boschert, S., Rosen, R. (2016). Digital Twin—The

Simulation Aspect. In Peter Hehenberger und David

ICEIS 2022 - 24th International Conference on Enterprise Information Systems

Bradley (Hg.): Mechatronic Futures. Cham: Springer

International Publishing, p. 59–74.

Cimino, C., Negri, E., Fumagalli, L. (2019). Review of

digital twin applications in manufacturing. In

Computers in Industry 113. p. 103–130. DOI:

10.1016/j.compind.2019.103130.

Dahmen, U., Rossmann, J. (2021). What is a Digital Twin

– A Mediation Approach. In 2021 IEEE International

Conference on Electro Information Technology (EIT),

2021. pp. 165–172, DOI:

10.1109/EIT51626.2021.9491883.

Enders, M.R., Hoßbach, N. (2019). Dimensions of Digital

Twin Applications - A Literature Review. In Twenty-

fifth Americas Conference on Information Systems,

Cancun, 2019.

Glaessgen, E., Stargel, D.S. (2012). Digital Twin paradigm

for future NASA and US Air Force vehicles. In Paper

for the 53rd Structures, Structural Dynamics, and

Materials Conference: Special Session on the Digital

Twin. p. 1–14.

Grieves, M. (2005). Product lifecycle management: the new

paradigm for enterprises. In International Journal of

Product Development (IJPD) 2005 (Vol. 2, No. 1/2), p.

71–84.

Grieves, M. (2014). Digital Twin: Manufacturing

Excellence through Virtual Factory Replication. In

Whitepaper.

Grieves, M., Vickers, J. (2017). Digital Twin: Mitigating

Unpredictable, Undesirable Emergent Behavior in

Complex Systems. In Franz-Josef Kahlen, Shannon

Flumerfelt und Anabela Alves (Hg.): Transdisciplinary

Perspectives on Complex Systems. Cham: Springer

International Publishing, p. 85–113.

Haag, S., Anderl, R. (2018). Digital twin – Proof of concept.

In Manufacturing Letters 15, p. 64–66. DOI:

10.1016/j.mfglet.2018.02.006.

Jones, D., Snider, C., Nassehi, A., Yon, J., Hicks, B. (2020).

Characterising the Digital Twin: A systematic literature

review. In CIRP Journal of Manufacturing Science and

Technology 29, p. 36–52. DOI:

10.1016/j.cirpj.2020.02.002.

Kampker, A., Stich, V., Jussen, P., Moser, B., Kuntz, J.

(2019). Business models for industrial smart services–

the example of a digital twin for a product-service-

system for potato harvesting. In Procedia Cirp, 83, p.

534-540.

Kritzinger, W., Karner, M., Traar, G., Henjes, J., Sihn, W.

(2018). Digital Twin in manufacturing: A categorical

literature review and classification. In IFAC-

PapersOnLine, 51(11), p. 1016-1022.

Negri, E., Fumagalli, L., Macchi, M. (2017). A Review of

the Roles of Digital Twin in CPS-based Production

Systems. In Procedia Manufacturing 11, p. 939–948.

DOI: 10.1016/j.promfg.2017.07.198.

Qi, Q., Tao, F. (2018). Digital Twin and Big Data Towards

Smart Manufacturing and Industry 4.0: 360 Degree

Comparison. In IEEE Access 6, p. 3585–3593. DOI:

10.1109/ACCESS.2018.2793265.

Qi, Q., Tao, F., Zuo, Y., Zhao, D. (2018). Digital Twin

Service towards Smart Manufacturing. In Procedia

CIRP 72, p. 237–242. DOI:

10.1016/j.procir.2018.03.103.

Schleich, B., Anwer, N, Mathieu, L., Wartzack, S. (2017).

Shaping the digital twin for design and production

engineering. In CIRP Annals 66 (1), p. 141–149. DOI:

10.1016/j.cirp.2017.04.040.

Shafto, M., Conroy, M., Doyle, R., Glaessgen, E., Kemp,

C., LeMoigne, J., Wang, L. (2010). Modeling,

Simulation, Information Technology & Processing. In

National Aeronautics and Space Administration, p. 1–

27.

Ströer, F., Sivasothy, P., Faißt, K-G., Apostolov, H.,

Eickhoff, T., Bechev, D. (2018). Combined

development and test of product-service systems in

early product development stages for customized,

availability-oriented business models in the capital

goods industry. In Procedia CIRP 72, p. 714–719. DOI:

10.1016/j.procir.2018.03.246.

Tao, F., Cheng, J., Qi, Q., Zhang, M., Zhang, H., Sui, F.

(2018a). Digital twin-driven product design,

manufacturing and service with big data. In Int J Adv

Manuf Technol 94 (9-12), p. 3563–3576. DOI:

10.1007/s00170-017-0233-1.

Tao, F., Zhang, H., Liu, A., Nee, A.Y.C. (2018b). Digital

Twin in Industry: State-of-the-Art. In IEEE Trans. Ind.

Inf. 15 (4), p. 2405–2415. DOI:

10.1109/TII.2018.2873186.

Tao, F., Zhang, M., Liu, Y., Nee, A.Y.C. (2018c). Digital

twin driven prognostics and health management for

complex equipment. In CIRP Annals 67 (1), p. 169–

172. DOI: 10.1016/j.cirp.2018.04.055.

van der Valk, H., Haße, H., Möller, F., Arbter, M., Henning,

J-L., Otto, B. (2020). A Taxonomy of Digital Twins. In

AMCIS 2018 Template, p. 1–10.

van der Valk, H., Haße, H., Möller, F., Otto, B. (2021).

Archetypes of Digital Twins. In Business &

Information Systems Engineering. DOI:

10.1007/s12599-021-00727-7

vom Brocke, J., Simons, A., Niehaves, B., Reimer, K.,

Plattfaut, R., Cleven, A. (2009). Reconstructing the

Giant: On the Importance of Rigour in Documenting

the Literature Search Process. In Proceedings of the

17th European Conference on Information Systems,

Verona: Italy, p. 2206–2217.

Webster, J., Watson, R.T. (2002). Analyzing the Past to

Prepare for the Future: Writing a Literature Review. In

MIS Quarterly (Vol. 26, No. 2), p. xiii–xxiii.

Zhao, G., Cao, X., Xiao, W., Zhu, Y., Cheng, K. (2019a).

Digital Twin for NC Machining Using Complete

Process Information Expressed by STEP-NC Standard.

In Fumin Zhang (Hg.): Proceedings of the 2019 4th

International Conference on Automation, Control and

Robotics Engineering - CACRE2019. the 2019 4th

International Conference. Shenzhen, China,

19.07.2019 - 21.07.2019. New York, New York, USA:

ACM Press, p. 1–6.

Digital Twin Paradigm Shift: The Journey of the Digital Twin Deﬁnition