178 - Digital Twins-Based Smart Manufacturing System Design in Industry 4.0A Review

Journal of Manufacturing Systems 60 (2021) 119–137
Contents lists available at ScienceDirect
Journal of Manufacturing Systems

journal homepage: www.elsevier.com/locate/jmansys
Digital twins-based smart manufacturing system design in Industry 4.0:

A review
Jiewu Leng a, b, c, Dewen Wang a, Weiming Shen b, Xinyu Li b, Qiang Liu a, *, Xin Chen a
a
Guangdong Provincial Key Laboratory of Computer Integrated Manufacturing System, State Key Laboratory of Precision Electronic Manufacturing Technology and
Equipment, Guangdong University of Technology, Guangzhou, 510006, China
b
State Key Laboratory of Digital Manufacturing Equipment and Technology, Huazhong University of Science and Technology, Wuhan, 430074, China
c
Department of Information Systems, City University of Hong Kong, Hong Kong, 999077, China
A R T I C L E I N F O A B S T R A C T
Keywords: A smart manufacturing system (SMS) is a multi-field physical system with complex couplings among various
Digital twins components. Usually, designers in various fields can only design subsystems of an SMS based on the limited
Manufacturing system design cognition of dynamics. Conducting SMS designs concurrently and developing a unified model to effectively
Cyber-physical systems
imitate every interaction and behavior of manufacturing processes are challenging. As an emerging technology,
Smart manufacturing
Function-structure-behavior-control-
digital twins can achieve semi-physical simulations to reduce the vast time and cost of physical commissioning/
intelligence-performance reconfiguration by the early detection of design errors/flaws of the SMS. However, the development of the digital
twins concept in the SMS design remains vague. An innovative Function-Structure-Behavior-Control-Intelligence-
Performance (FSBCIP) framework is proposed to review how digital twins technologies are integrated into and
promote the SMS design based on a literature search in the Web of Science database. The definitions, frame
works, major design steps, new blueprint models, key enabling technologies, design cases, and research di
rections of digital twins-based SMS design are presented in this survey. It is expected that this survey will shed
new light on urgent industrial concerns in developing new SMSs in the Industry 4.0 era.
1. Introduction providers (e.g., PTC, Siemens, ANSYS, Dassault) proposed the concept of
virtual prototype, digital prototype, active prototype, and others. A
The general birth process of a manufacturing system can be digital prototype is a kind of rehearsal in which the information model
described as follows: a designer generates an idea in his mind, then replaces physics. Among various digital prototype concepts, the Digital
draws it in drawings (forms a model) and tries to make some physical Mock-Up [2] concept is widely used, which emphasizes complex map
prototypes for testing, and finally assembles the system. Table 1 pro pings and contextual relationships among 3D model simulations. The
vides an overview of the manufacturing system design in different stages existence of digital prototypes greatly reduces the failure of physical
of industry development from Industry 1.0–4.0. prototypes (sometimes they simply directly replace real prototypes).
In Industry 1.0 and 2.0, the information in the R&D stage exists in the Based on a virtual simulation of the physical manufacturing process, the
drawings; and in the physical prototype stage, the information uses the SMS performance can be evaluated early to guide cutting down the
physical system as the carrier. The conventional manufacturing system reconfiguration costs/losses in establishing physical SMS prototypes (e.
design method requires physical prototypes to be assembled to check g., Plant Simulation). Using virtual reality could greatly enhance the
and verify whether the virtual space model is accurately matched, convenience and efficiency of SMS design (e.g., FlexSim). Due to the
including structure, ergonomics, and performance indexes. However, consistency of information transmission in digital prototypes, the diffi
using physical prototypes to verify the design is very costly. To avoid culty of the manufacturing system design is greatly reduced [3].
unnecessary risks on physical prototypes, simulations of the perfor In Industry 4.0, smart manufacturing has become a development
mance and manufacturability of a model in the virtual space are critical direction of the world’s manufacturing industry. New national advanced
[1]. manufacturing strategies around the world resulted in the increasing
In Industry 3.0, many Computer-Aided Design (CAD) software need of designing new smart manufacturing systems. A Smart
* Corresponding author.
E-mail address: [email protected] (Q. Liu).
https://doi.org/10.1016/j.jmsy.2021.05.011
Received 20 March 2021; Received in revised form 16 May 2021; Accepted 17 May 2021
Available online 25 May 2021
0278-6125/© 2021 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.
J. Leng et al. Journal of Manufacturing Systems 60 (2021) 119–137
Manufacturing System (SMS) is a multi-field physical system composed The definitions, frameworks, major design steps, new blueprint
of intelligent machines, materials, products, and complex couplings models, key enabling technologies, design cases, and research directions
among various elements. In the digital design process, an SMS can be of digital twins-based SMS design in industry 4.0 will be presented in
broken down into digital models of various granularities in digital space, this survey. The organization of this survey is outlined in Fig. 2.
while the physical products and manufacturing processes exist in
another physical space. In the SMS design process, high-fidelity cyber 2. A framework for reviewing the digital twins-based SMSD
models mapping the real worlds of SMSs are critical to fulfilling the gap
between its design domain and operation domain [5]. 2.1. Definition of smart manufacturing system design
Digital twins (DT) technology is a solution. Nowadays, the digital
twins trend is gaining momentum because of rapidly evolving simula Manufacturing System Design (MSD) conventionally includes the
tion and modeling capabilities, better interoperability and IoT sensors, modeling, analyzing, and optimizing of system layout, production ca
and more availability of tools and computing infrastructure (www2. pacity, production flow, material handling system, manufacturing
deloitte.com/). Recent MarketsandMarkets research suggests that the methods, system flexibility, and operation strategies [7]. Decisions on
global digital twin market size was valued at USD 3.1 billion in 2020 and the selection and configuration of resources should be optimized in the
is projected to reach USD 48.2 billion by 2026 (www.marketsandma design process. The goal of the MSD is to find the design solution that
rkets.com). A digital twin could realize the feedback of the digital yields the optimal performance measures. In terms of the modeling of
model of cyberspace to the real physical system. In this way, it is possible the couplings among the system elements, the mathematical
to ensure the coordination of the digital and physical spaces within the programming-based quantitative approaches cannot model and analyze
scope of the entire life cycle. Zhang et al. [5] firstly applied the digital the stochastic disturbances (e.g., machine tool breakdowns, resource
twins-based approach in the manufacturing system design sector. After a operation conflicts) that greatly affect the system performance. On the
review of existing SMS development models, Mahmoud and Grace [4] other hand, simulation models, due to the ability to account for the
suggested a digital twins-based simulation approach for SMS configu time-variant dynamic behaviors, have been utilized to optimize the
ration. Because of the supremacy of digital continuity, realistic system performance.
modeling, and interaction between disciplines against conventional However, MSD is facing new challenges under the smart
simulation, the creation of a digital twin in the SMS designing phase will manufacturing blueprint. Smart Manufacturing System Design (SMSD)
be highly valuable during all the lifecycle of the SMS. refers to the design of an SMS. SMS coordinates various manufacturing
However, the application of the digital twins concept in the SMS elements (e.g., machine tools, material, human, and environment) and
design remains vague [6]. Therefore, it is necessary to explore and sort holistically optimizes the operations based on a unified environment of
out the research on the rational integration of digital twins technology Cyber-Physical Systems (CPS). CPS for manufacturing enables optimi
into the SMS design. This article attempts to show how digital twins zation of product development and total control of production system
technology is integrated into the SMS design to truly promote the through real-time exchange of all information required for production
development of smart manufacturing. A literature search on digital based on Internet-of-Things (IoT) [8]. SMSD is a complicated process
twins-based SMS design is conducted in the Web of Science database. composed of modeling, analyzing, mining, and learning data from many
The collected literature is further refined by following three phases. The sources in SMS. Besides the uncertainties of the input demands and the
first phase is to screen related papers via keyword retrieving. Two process disturbances, the complex interactions, couplings, conflicts
combinations of keywords, namely, {“digital twin” and “manufacturing among the design variables and goals make the SMSD a highly iterative
system design”} and {“digital twin” and “manufacturing system plan and cost/time-consuming task. Since the major difference between an
ning”}, are used for retrieving papers. The retrieving yields 202 papers SMS and a conventional manufacturing system is industrial intelligence,
and 54 papers, respectively (up to 31 December 2020). After the dele the SMSD is more challenging in terms of the design of industrial arti
tion of duplicated papers, this retrieving yield 220 papers for further ficial intelligence models compared to the conventional MSD.
screening. The second phase is a theoretical screening process to identify
high-quality studies related to the concept, key enabling techniques,
models, systems, frameworks, and case studies on SMS design. Research 2.2. The difference between digital twins method and simulation methods
not related to the digital twins or SMS design domain and short papers
less than four pages are excluded. In this phase, 159 papers are finally 2.2.1. Definition of digital twins
included to discover the key issues, new solutions, challenges, and The digital twins concept is a development of modeling and simu
promising directions in the research of the digital twins technology in lation technology (Fig. 3). Traditional simulation methods are of limited
the SMS design process. Additionally, 33 additional papers are referred capabilities in evaluating system performance [7]. Digital twins tech
to make this survey more comprehensive. Finally, this survey includes nology represents the breakthrough of limitations on the modeling and
192 references, and the statistics of collected literature are shown in engineering analysis capabilities of simulation by integrating the IoT
Fig. 1. technology [9].
The virtual SMS is described by its digital models. Digital models
Table 1
Manufacturing system design in different stages of industry development.
Stages Industry 1.0 Industry 2.0 Industry 3.0 Industry 4.0
Period 1760s~1850s 1860s~1960s 1970s~2010s 2010s~Now

Mfg. Paradigm Mass Production Mass Production +Mass Customization +Mass Individualization
Mfg. Goal Scale +Speed +Knowledge +Intelligence
Mfg. System Type Mechanical Production Line Automated Mfg. Syst. Computer-Aided Mfg. Syst. Smart Mfg. Syst.
Design Elements Machinery +Automation +Digitalization +Complex Algorithms
Design Goal Tolerance +Interchangeability +Flexibility/Reconfigurability +Sustainability/Intelligence
Design Methods 2D Drawings 2D Drawings 3D Models & Simulations Digital Twins-based
Typical Blueprint – – Digital Prototype/Mock-Up Lifecycle Digital Twins
Major Challenges Huge Commissioning Cost Huge Commissioning Cost Complex Assemblies +Multi-dimension Couplings
Ref. – – – [4]
Notes: “+” is followed with additional items in the column compared to the former column; “-” denotes vacancy/unknown (Similarly hereinafter).
120
Fig. 1. Statistics of collected literature.
Fig. 2. The organization of this survey.
could be categorized into the digital replica, digital shadow, and digital medicine fields [14–16]. It is recognized as one of the technological
twin [10]. A digital replica emphasizes the automatic projection of pillars for achieving smart manufacturing and Industry 4.0 [17–20].
system constructions. The digital shadow emphasizes mathematical
modeling to describe the physical/chemical attributes of a system. A 2.2.2. The advantages of digital twins-based SMSD approach
digital twin is an integrated multi-physics and multi-scale simulation of Traditionally, simulation technology is restricted to a standard tool
a product/system that can model the mechanical, electrical, software, for supporting designers to solve specific engineering problems.
and other discipline-specific properties across its lifecycle. A digital twin Communication-By-Simulation is the core concept of Model-Based Sys
is supposed to be able to optimize the physical product/system based on tems Engineering. Extending the simulation from the design phase to
the updated real-time data synchronized from sensors [11]. Other subsequent lifecycle activities is the trend. As shown in Fig. 4, the
similar digital twins concept includes digital surrogates [12] and intel existing MSD methods lack an effective iterative optimization of
ligent beings [13]. Valckenaers [13] argued that real-world smart sys "execution-analysis-adjustment" in the design process. The design has
tems consist of digital twins and intelligent agents. The digital twins not been verified so it is difficult to guarantee the decision quality.
mirror the real systems without imposing artificial limitations, while Although simulation models are useful for supporting the design visu
intelligent agents convert NP-hard problems into alization, how to build a unified model across multi-discipline of SMS is
low-polynomial-complexity computations. Digital twins technology is challenging. The solution is to build a digital twin that reflects and ob
being developed to optimize manufacturing, aviation, healthcare, and serves the real-world SMS behavior and use it in all design procedures
121
Fig. 3. The evolution of simulation technology from the design view.
Fig. 4. The transition from conventional MSD to digital twins-based SMSD approach.
including system deployment and commissioning [21], which is named some industry specifications. The semi-physical system commis
as Digital twins-based Smart Manufacturing System Design (DT-SMSD) sioning is the focus of this stage.
in this survey. Digital twins are established based on combining a • Fine-tuning stage of digital twins (Intelligence model design and
multi-physics simulation model with the real-time data of a pro Performance model design): The digital twins are used to help build
duct/system [22]. Digital twin models could be timely modified and the prototype SMS and plan the system test scheme. The digital twins
validated to avoid bottlenecks that happen in the SMS simulation and could reduce trials and errors to accurately locate the points that
development processes. The design process of an SMS implies the con need to be intelligentized. The optimal performance could be ob
cepting, forming, and fine-tuning of corresponding digital twins. The use tained in a lean way. In another word, the digital twin approach
of digital twins in three stages is different, for example: could achieve the SMS validation purpose with fewer test times and
improve the design efficiency. The key point to the fine-tuning of
• Concepting stage of digital twins (Function model design and corresponding digital twins lies in the excellent reconfigurability and
Structure model design): Using multi-disciplinary simulations to flexibility of SMS to support alternative design/prototyping solutions
quickly verify the conceptual scheme. The goal in this stage is not the [23].
pursuit of simulation accuracy but a fast simulation.
• Forming stage of digital twins (Behavior model design and Control The most direct value of implementing digital twins is to replace the
model design): Key technical design parameters are determined by costly pure-physical commissioning and test. However, the ultimate goal
simulations. Although the purpose of digital twins is to replace of implementing digital twins is not simply validation but design inno
physical commissioning, necessary semi-physical system commis vation. The established digital twin in the SMS design phase can also be
sioning in practice needs to be retained, especially as required by further utilized for monitoring, optimizing activities, diagnostics, and
122
prognostics in the following-up lifecycle phases of SMS [24]. control the output of the process. It usually includes electrical &
pneumatic system designing, control & sensor networking, equip
ment controls designing, and multi-process coordination.
2.3. Function-structure-behavior-control-intelligence-performance
• Intelligence model. This model is supposed to able to describe,
framework
develop, and validate the learning ability, optimization ability, and
autonomy ability of an SMS from the control design perspective. It
The SMS is conceptually considered as an integration of complex
includes a systematic method and a variety of reasoning rules, ma
interconnected functional models. The essence of the SMSD is to
chine learning algorithms, and computational optimization algo
establish entities that meet individual demands, such as layout con
rithms for achieving sustainable performance in SMS.
straints, production capacity, used equipment integration, and
• Performance model. The differences in design decisions will result in
manufacturing efficiency. In the SMSD process, the personalized re
the fluctuation of manufacturing execution efficiency. This model
quirements from the customer and product domains are transferred to
usually includes the evaluation and optimization of the system per
the SMS configuration domain and corresponding execution domain,
formance such as efficiency, flexibility, reconfigurability, robustness,
including the customized industrial artificial intelligence models in SMS.
adaptivity, extensibility, and resilience.
This survey proposes a Function-Structure-Behavior-Control-
Intelligence-Performance (FSBCIP) framework to review the major
The input of SMSD includes personalized requirements and param
steps of DT-SMSD, as shown in Fig. 5.
eters. The design process includes the design of the above six models.
The output of the SMSD is the whole cyber models, motion scripts,
• Function model. It is a structured description of the activities for
control schemes, execution systems, and intelligent algorithms, which
manufacturing the product and the relationships among the activ
form the final optimized/fine-tuned digital twins. The proposed FSBCIP
ities. This model highlights the influence of manufacturer demand in
framework is a development of the conventional function-behavior-
the conceptual design on product manufacturability, assembly,
structure model [25]. Conventional MSD research usually focused on
maintainability, and safety. What kind of functions the SMS has is
the functional, structural, and behavior dimensions. The design scope
determined by the product features it faces and its development re
review in this paper is extended to the control, intelligence, and per
quirements. This model usually includes process planning, equip
formance dimensions, which are critical to efficiently realize the
ment selection, and mechanical design.
comprehensive design goals of SMSD in the context of Industry 4.0.
• Structure model. It is a structured description of the fusion,
Abstracting and making full use of similarities in the FSBCIP framework
connection, and assembly relations among the mechanical structures
at a higher level could effectively enhance the accuracy and efficiency of
that realize manufacturing functions, principles, and behaviors. The
the customized design of similar SMSs.
interrelation of the mechanical structure is the foundation for the
transferring and transformation of the material, energy, information,
and motion behavior of the manufacturing system. This model usu 3. Digital twins-based SMSD
ally includes topology definition, layout planning, fixture/carrier
designing, and buffer/cache designing in SMS. This section will review the research progress and issues in the
• Behavior model. It is a structured description of the mechanical critical steps of SMSD that could be enhanced by the digital twins
motion transmission, the motion form transformation, and their technologies from the proposed FSBCIP view, including 6 dimensions
mutual relations. The material, energy, and information connection and 16 sub-dimensions.
that existed among the SMS are mainly reflected in the transfer
relationship of the mechanical motion behavior. This model usually 3.1. Digital twins-based function design of SMS
includes equipment motion, WIP motion, material delivery/
handling, and process engagement in SMS. The essential elements of 3.1.1. Digital twins-based process planning
the motion behavior mainly include force, displacement, velocity, Manufacturing process planning is a process performed by engineers
acceleration, jerk, angular displacement, angular velocity, angular to analyze manufacturability, cost, and efficiency in the early phase of
acceleration, angular jerk, snap, and higher derivatives. MSD based on the manufacturing feature recognition [26] from product
• Control model. This model is built to deal with the structure, oper design. It includes manufacturing operation selection and operation
ation, or calculus of a process by statistical or engineering methods to sequencing. Few studies have been conducted on how to deal with the
Fig. 5. An FSBCIP framework for smart manufacturing system design.
123
random disturbances and complex loadings in the process plan process design of complex engineering systems from the perspective of the
[27]. Concerning current simulation-based process planning ap whole SMS. The digital twins model applied in the SMSD process should
proaches, the obtaining of adequate process information and searching naturally have the ability of optimization. Guo et al. [34,35] introduced
for optimal decisions needs 74 % of the planning time [28]. Process a digital twins-based Graduation Intelligent Manufacturing System that
planning is the first step in SMSD that could be enhanced by digital twins is capable of real-time task allocation and execution for fixed-position
technology. assembly islands. Generally, the digital twins-based coordination
The digital twin for manufacturing processes improves the optimi among production processes ensures that each equipment is allocated in
zation speed and accuracy of planning. Lu et al. [29] established a the right position and utilized to the right activities with enhanced
bidirectional cyber-physical mapping for planning and optimization of visibility.
re-manufacturing processes based on the discrete event simulation and
IoT technology. Liu et al. [30] proposed a digital twins-based process 3.2.2. Digital twins-based tooling and buffer designing
evaluation approach on the basis of the real-time mapping between the More than 20 % of the complex products’ manufacturing time is the
physical machining process and the design model. Further, to express auxiliary preparation time, which is thereby the key point to improve
the evolutionary characteristics of product processing, Liu et al. [31] manufacturing efficiency [36]. The tooling and buffer designing is a
proposed a digital twin process model based on the critical step affecting the auxiliary preparation time of product
knowledge-evolution machining features. Sierla et al. [32] derived a manufacturing. Through the takt time analysis and productivity
digital twin from a digital product description to concurrently conduct balancing analysis, the type and quantity of machines and buffer ca
the assembly design and planning, and automatically coordinate the pacity in each process could be determined. However, the stochastic
manufacturing resources in one manufacturing unit. The resulting ver logistics, massive disturbances, and various routing patterns in SMS
satile manufacturing system is supposed to be able to cope with a large make tooling and buffer designing a very challenging task.
variety of products at minimal reconfiguration cost. The application of The digital twins technology is a potential solution to tackle the
digital twins technologies in process planning is not yet a common difficulty of tooling/buffer designing and bridge the gap between
practice. More advantages, as well as deficits, are supposed to be designing and controlling. For instance, Liu et al. [36] proposed a digital
discovered in using digital twins in the manufacturing process planning, twins-based approach for dynamic clamping and positioning of flexible
as well as the rapid mining and learning of the collected data. tooling systems. The geometry data of the manufacturing parts are
collected in real-time to drive optimizing of the clamping forces and
3.1.2. Digital twins-based equipment mechanical design positioning. Wang et al. [37] proposed a digital twins-based CPS for the
A mechanical design of equipment should be conducted based on a design and control of roller conveyor lines, which is critical for
deep understanding of its machining process. It is necessary to simulate improving the sorting speed and delivery speed of workpieces in SMS
the physical machining processes of operations to design high- logistics.
performance equipment. For instance, in the design of laser processing
equipment, a digital twin model of the interaction process between laser 3.3. Digital twins-based behavior design of SMS
and materials should be established to have a deeper understanding of
the manufacturing process of laser materials through simulations. Haber 3.3.1. Digital twins-based equipment behavior design
et al. [33] presented a digital twins-based approach for optimizing the Equipment behavior design refers to defining the actions (including
controller parameters of the backlash peak amplitude/time and the high-level action patterns) of equipment (process equipment and inter
hysteresis amplitude of an ultra-precision motion stage system to guar mediate equipment) and the action sequence of multi-process coordi
antee an appropriate dynamic response in the presence of backlash and nation according to process route and process requirements. It should
friction in a cost-effective manner. Actually, the digital twinning of the avoid unnecessary collisions and deadlocks. Although building virtual
equipment implies the digital twinning of the physical process of the equipment is useful for simulating equipment’s capabilities safely and
equipment operation. The semi-physical simulation of the process could cost-effectively, it is difficult to exactly simulate the physical equipment.
be carried out to reflect all activities/status of the physical equipment. Establishing the digital twin of equipment could benefit the behavior
design, as well as the further diagnosis and prediction of failures/faults
3.2. Digital twins-based structure design of SMS in the operation of physical equipment. Vatankhah et al. [38] proposed a
digital twins-driven motion planning approach by integrating with
3.2.1. Digital twins-based layout planning and topology optimization agent-based rapid optimization algorithms in a robotic manufacturing
The layout planning (e.g., U-shaped, O-shaped, L-shaped, 1-shaped, cell. Cai et al. [39] utilized sensory data to model the actual status of the
and mixed type) of SMS is largely decided by the geometric con machine tools to develop their digital twins for enhancing accountability
straints of given site space. On the other hand, the logistics intension is a and manufacturing capabilities. Further, Cai et al. [40] presented an
critical index to evaluate the layout planning solution. Based on the augmented reality-based simulation approach for robotic toolpath
digital twins technologies, the physical manufacturing process could be planning based on communicating the layout information between a
mirrored and validated with its digital model in cyberspace promptly, reconfigurable additive manufacturing system and its’ digital twin.
and thus could support the automatic retrieving of layout information The intelligent machine tool (IMT) is a vision of digital twins-driven
and facilitate planning an optimized layout in the digital twin of phys digitalized, highly-efficient, and autonomous machine tools in smart
ical SMS. manufacturing blueprint. Tong et al. [41] presented a digital
According to the process route and the priority relationship between twins-based real-time machining data application service for IMT. Data
processes, the system topology and the connection relationship between transmission and storage are completed using the MTConnect protocol
equipment could be determined. The topology design of SMS is the basic and components. The digital twin of IMT is equipped with the optimi
idea of using numerical computation and optimization algorithms for zation functions related to the machining trajectory, machining status,
seeking to achieve the optimal topology structure and satisfy various energy consumption, machine tool dynamics, contour error estimation,
constraints in a given design space. Multi-disciplinary design optimiza and cutting-tool compensation. Also, as a key component of machine
tion is a well-known design direction aimed at solving the coupling and tools, the digital twin of the cutting tool is critical to the optimization of
tradeoff problems of multiple disciplines in the design of large-scale machining solutions [42].
complex engineering systems, including SMSs. It fully explores and
utilizes the synergistic mechanism of interaction in engineering systems, 3.3.2. Digital twins-based ergonomics design
considers the interaction between various disciplines, and optimizes the Due to the limited artificial intelligence level of software models and
124
the limited flexibility/reconfigurability of hardware in SMS, dynamic system to achieve excellent logistics traceability, controllability, and
demands and contextual randomness still require humans (workers) to efficiency via a design-for-manufacturing manner. For instance, Wang
participate in the highly flexible tasks. The ergonomics design in the et al. [53] proposed a proactive digital twins-based material handling
interactions between workers and machines is an indispensable method for CPS-enabled shop-floor to reflect the production status of
dimension to evaluate and optimize the system efficiency, flexibility, physical SMS. Banyai et al. [54] introduced a concept of matrix pro
and reconfigurability in the SMSD. Also, the advent of collaborative duction as a CPS concentrating on material handling and described an
robots makes the use of digital twins that involve humans grow in the extended and real-time optimization model of supply routing and
future. The discrete event simulation methods widely used during the assignment. Yerra and Pilla [55] created a digital twin for each physical
design stage cannot mirror the continuous movement behavior of part in smart material flow control in automotive composites body shop.
workers in SMSs. It is critical to enhancing the ergonomics simulation Instead of using the century-old assembly line philosophy in the
model with the continuous movement, working environment, and per manufacturing layouts, they proposed an assembly grid layout to
formance routines of workers in SMSs [43]. improve space utilization, inspired by a colony of ants working to
A form of hybrid automation is taking benefit from the synergistic accomplish a common goal. It has the potential to improve flexibility to
effect of human-robot collaboration. The need for adaptability and the product variabilities. Nevertheless, more realistic virtual models of WIP
dynamics of human presence are keeping the full potential of human- need to be developed to further fulfill the gap between design and
robot collaborative systems difficult to achieve [44]. When used in the manufacturing and to map the real and cyber worlds.
assembly, the requirement of flexibility, adaptability, and safety makes
the design and redesign of human-robot collaborative systems a complex 3.4. Digital twins-based control design of SMS
and prone to error process [45]. Based on the data collected from
wearable sensors on workers, virtual reality-based digital twins tech 3.4.1. Digital twins-based process optimization and equipment control
nology is the key to designing, monitoring, and optimizing the ergo designing
nomics performance of manual work activities, as well as guiding the Virtual equipment is usually built to simulate the machining toolpath
improvement of the working conditions in human-machine interactions. of various workpieces/parts, and thereby emulate each machining
For instance, Graessler and Poehler [46] presented a digital twin process to guide the generation of controls. If the digital twin of physical
approach to enable workers on the shop floor to take part in computa equipment is established, designers could not only confirm its func
tional decision-making. Latif and Starly [47] proposed a data-driven tionality to greatly decrease the design cycle but also conveniently make
simulation algorithm to model the complex and manual innovations in manufacturing processes. For instance, Zhu et al. [56]
manufacturing process in a generic-reusable way, from which the pro developed a functional simulation system to establish digital twins of the
duction managers can make more informed early decisions. Greco et al. physical machining tools. It is capable of generating the robot manipu
[48] proposed a methodological human digital twin framework that lator path to produce a given workpiece, which could be straightly
enables the monitoring and optimizing of the ergonomics performances controlled by the CAD tool employed in the product design phase. Cao
of workers. The actions and personal preferences of workers should be et al. [57] proposed a STEP-NC standard-based digital twin system for a
captured in the digital twin [49]. Malik and Bilberg [50] presented a computer-aided manufacturing system. Zhao et al. [58] proposed a
digital twin approach to help the designing and controlling of digital twin-driven cyber-physical system for autonomously controlling
human-machine interactions/cooperation’s in manual assembly work. of micro punching system. Gohari et al. [59] developed the digital twin
After the SMSD phase, the digital twin of ergonomics should be inspection system to optimizing the inspection controls of geometrically
continuously updated in the following operating activities of the SMS complex surfaces of workpieces. The closed-loop digital twins could
from a comprehensive cyber-physical-social-connected systematic view avoid the randomness involved in the inspection process based on the
[51]. The realization of digital twins for human-robot collaborative high-volume contextual data collected from distributed sensors in
tasks is facing challenges in the lack of high-fidelity simulation models at workpieces.
various detailed levels of human endeavors, as well as the seamless From the workpiece view, incorporating the geometry assurance
integration of different modules. model into digital twins is critical for realizing the process optimization
and controls designing, thereby guarantying the desired geometry of the
3.3.3. Digital twins-based WIP/material handling product. For instance, Tabar et al. [60] proposed a digital twin approach
Unsynchronized production can easily cause problems such as the emphasizing the geometry assurance to secure the geometry of the in
backlog of intermediate warehouses, unsmooth production, and long dividual sheet metal assemblies based on a spot-welding sequence
production cycles. Synchronized production helps to improve overall optimization method. Based on an overview of employing the digital
efficiency and reduce waste. It requires the planning of the material twins approach in geometrical variations management in the context of
handling, production logistics path, movement pattern, suspension, and industry 4.0 [61], Schleich et al. [62] presented a Skin Model Shapes
cache mode of the WIP (i.e., work-in-process) according to the action concept-oriented reference model serving as the digital twin of the
and behavior mode of the equipment. Material handling refers to the physical product to simultaneously optimize its design and
unloading, distributing, and delivery of raw materials to the production manufacturing. Cerrone et al. [63] established a finite element model of
unit and warehouse associated with the entire manufacturing process. the as-manufactured specimen to resolve the crack-path ambiguity, and
Robots are common tools for material delivery/handling in SMSs due to thus predict the crack path. The modeling of as-manufactured properties
their high flexibility. Using simulation/emulation is an effective is essential to establish the digital twins of workpieces.
approach of optimization designing of WIP and material handling of Equipment calibration and subsequent controller-based compensa
manufacturing system [52]. An adaptable material handling solution is tion are critical steps in the initial controls design to improve the
supposed to be able to deal with the dynamic-change product volumetric performances in the ultra-precision SMS domain. Digital
manufacturing requirements under the mass individualization para twins-based technology is a potential solution. For instance, Montavon
digm. The optimization design of WIP and material handling cannot be et al. [64] adopted a digital view regarding machine tool calibration and
separated from the consideration of three aspects: 1) the impact of the defined the layered semantics in the CPS of SMS, in which the controller
manufacturing process on the logistics; 2) the requirement of of equipment additionally acts as the edge computing device to coor
manufacturing capacity on the logistics facilities; 3) the impact of pro dinate with upper management systems in the manufacturing process.
duction takt time on the material flow. Although some performance checking models have been proposed for
These dynamic factors and impacts could be more efficiently evalu predicting the volumetric performance degradation and thereby avoid
ated by digital twins technologies. Digital twins technology enables the ing undesirable breakdown, the insights in the equipment calibration
125
are still limited due to the fragmentation of sensory data. SMS during semi-physical commissioning. Leng et al. [70] proposed
digital twins-based remote semi-physical commissioning of flow-type
3.4.2. Digital twins-based system controls commissioning smart manufacturing systems. As shown in Fig. 6, the I/O addresses
The design and runtime validation of controls are critical for product and signals of distributed machines (hardware PLC) are configured and
quality management but challenging since SMSs in Industry 4.0 vision mapped to the software PLC of digital equipment models in the digital
are highly complex and heterogeneous. The system controls commis twin system via Industrial Internet protocols. When the designers start
sioning is a complex task requiring modeling of low-level and high-level the simulation of the whole material flow, workers put the needed raw
control functions in different software environments and connecting material in the machines according to the instructions. Except for the
them through specific interfaces, which is a critical step to verify and simulation of whole material flow, many situations of simulation (e.g.,
validate the basis design models. machine kinetics inside equipment, and control compatibility) may be
The commissioning of an SMS mainly includes three approaches: performed via a no-load running test manner [70].
virtual commissioning, physical commissioning, and semi-physical A prerequisite is a current virtual plant model of the structured,
commissioning. The physical commissioning needs to wait for the mechatronic resource components of the real plant during the different
completion of physical equipment assembly and control logic to com design phases. Talkhestani et al. [71,72] presented a cross-domain
plete the physical debugging work, resulting in waiting-time waste. mechatronic data structure-based deviations detection method of the
There will inevitably be design errors, logic errors, coding errors, and virtual manufacturing cells in the automotive industry in the context of
other problems, resulting in irreversible consequences such as me RAMI 4.0. Kang et al. [73] proposed a digital twin framework named
chanical equipment misoperations. There exist many virtual simulation DigTwinOps for runtime verification of CPS. The DigTwinOps is char
commissioning software tools, such as Siemens Tecnomatix, Rockwell acterized by an execution engine that drives the cyber model to map
Emulate3D, VisualComponents, and Simumatik3D. The results obtained states of physical manufacturing. Therefore, a designer could validate
from virtual commissioning suffer from low confidence. Simply writing the controls before deploying them to the physical SMS.
the model behavior logic on the simulation platform is unable to debug Generally, the greatest difference between the simulation-based
and verify whether the real controller can drive the physical equipment design method and the digital twins-based design method lies in the
correctly and reasonably [65]. After the design of the actual difference between conventional simulation-based virtual commis
manufacturing system is adjusted, the behavior logic prepared by the sioning and digital twins-based semi-physical commissioning. However,
simulation model also needs to be changed accordingly, which cannot it is still challenging to evaluate the environmental and human factors of
meet the iterative optimization design process of the physical the equipment control in a digital twin model.
equipment.
Therefore, it is essential to carry out the semi-physical (hardware-in-
the-loop) commissioning. The application of hardware-in-the-loop 3.5. Digital twins-based intelligence model design of SMS
commissioning is to create reasonable and effective virtual equipment,
which requires the combination of control technology and digital twins Intelligence is an important manifestation of the digital twins model.
technology. The digital twin of the physical equipment is taken as the It is a common vision that an SMS could realize self-perception, self-
debugging object to run the control logic in the virtual environment, to decision, self-execution, autonomy, self-learning, and self-improvement
optimize and verify the rationality and feasibility of the logic. A digital in its production process, including the optimal control of parameters in
twin can truly reflect the changes of the physical equipment, and the production process. Its optimization not only includes classic and
continuously accumulate relevant knowledge, optimize and analyze the modern optimization control technology but also integrates big data and
current situation under specific controls, and thus support the design artificial intelligence technology. Without digital twin’s accurate
decision. modeling of the actual system, the intelligence of an SMS cannot be
In the development process of an SMS, many sets of equipment may analyzed and reasoned. The industrial artificial intelligence that exists in
be supplied by different manufacturers, and it requires a concurrent SMS could be categorized into data-driven machine learning models and
control design of multiple virtual machines. Conventionally, the control computational optimization models.
program/code written by engineers relying on their personal experience
is neither conducive to others’ understanding of the logical action flow 3.5.1. Digital twins-based machine learning model design of SMS
nor conducive to the rapid response of the iterative optimization of the The SMS is characterized by its complex, dynamic, and chaotic be
control debugging process. Automated control code generation is critical haviors [74]. Machine learning saw fast developments in terms of not
and, in the long run, makes it easier to modify and upgrade programs only promising results but also usability to satisfy the demand for
and software. For instance, Julius et al. [66] proposed a method to high-quality customized products efficiently. The intelligent control
automatically convert the Grafcet specification language into the ST text models based on machine learning (e.g., naive Bayes, decision tree,
code of Programmable Logic Controller (PLC) while preserving the hi random forest, support vector machine, dimensionality reduction algo
erarchical structure described by the processing logic. A methodology of rithms, and neural network) exist in the different levels from
automatic generation, formal verification, and implementation of safety manufacturing equipment, production line, workshop, and factory to
PLC program can save the time of equipment design/verification, and the cross-factory level [75].
thus avoid the duplication of tasks and design errors [67]. The selection and design of an appropriate machining learning sys
Developing a digital twins-based commissioning system that could tem to handle dynamic events and enable intelligent process control is
automatically generate the control programs (e.g., PLC codes) is of great an essential issue in SMSD. The machine learning model design of SMS is
significance to further shorten the preparation time of integrated system to develop and validate a variety of machine learning algorithms to
commissioning and continuously iterative design optimization. For achieve sustainable performance. For instance, to account for the
instance, Orive et al. [68] presented a digital twins-based design multimodal and time-varying properties of system dynamics in SMSs,
approach to conveniently emulate controlling reactions to disturbances Sun et al. [76] proposed a hybrid machine learning modeling approach
and misoperations during system commissioning. Toivonen et al. [69] based on an extended/comprehensive state-space descriptive system, in
introduced a digital twins-based versatile learning environment for which the system control parameters are learned for responding
supporting engineers to design and program controls in cyberspace different running conditions. The machine learning systems can act as an
before commissioning physical SMS. Digital twins could cut down the intelligent execution engine in a digital twin of the physical process.
reconfiguration cost since it enables the timely discovery of design However, if problems in the industry could not be properly modeled, this
errors/flaws in the functional, structural, and behavioral dimensions of technology will not yield true value but misleading.
126
Fig. 6. The rationale of digital twins-based system controls commissioning [70].
3.5.2. Digital twins-based computational optimization model design of SMS alternative artificial intelligence, computational optimization, and
Computational optimization in SMSs includes a variety of algorithms hybrid autonomy models to find better production patterns/rules [79].
in control and management optimization, such as large-scale optimiza Generally, the abstraction of the optimization problems and their
tion, complexity theory, network optimization, and robustness/sensi coupling relationships contained in manufacturing systems is a critical
tivity analysis. The computation cost and the numerical accuracy of design step.
computational optimization algorithms are influenced by the initiali
zation of hyper-parameters, which are sensitive to application situa
3.6. Digital twins-based performance design of SMS
tions. However, the current mass individualization paradigm implies
significant changes in SMS. In SMSD practice, designers are searching
Various blueprints of SMS have envisioned it to be highly flexible,
for robust computational optimization algorithms to decrease the
intelligent, adaptive, and autonomous [80]. These SMS performance
sensitivity to the changes/disturbances as much as possible [77].
goals should be evaluated and optimized in the design phase to deal with
Therefore, the main goal of the computational optimization model
more unpredictable changes and demands [81]. Scholars are making
design of SMS is to find a robust design that could handle a large variety
efforts to improve system performance (e.g., quality, flexibility, and
of system changes and disturbances. Leng et al. [78] proposed a digital
reconfigurability) based on digital twins technology.
twins-driven warehousing optimization model for a large-scale auto
mated high-rise warehouse under a mass individualization paradigm.
3.6.1. Quality-oriented SMSD based on digital twins
The digital twins technology provides a highly confident runtime envi
Product quality is an essential metric to evaluate the performance of
ronment to simulate the complex system for searching robust compu
a newly design SMS. On one hand, from the product perspective, the
tational optimization models.
mathematical modeling of tolerance specifications is critical for
The hybrid autonomy models that integrate data-driven machine
enabling overall tolerance analysis, which could be enhanced by the
learning models and computational optimization models are also com
digital twins technology. Luo et al. [82] proposed a parametric space
mon practices. Real-time data-based digital twin systems could support
envelope-based tolerance modeling approach that can facilitate
127
different computer-aided tools to satisfy individualized tolerancing re manufacturers are faced with the challenge of increasing the sustain
quirements. On the other hand, from the perspective of the ability of the SMS [93]. Optimal machining parameter design/decision
manufacturing system, the full-field displacement perceiving of key is imperative for energy conservation, emission reduction, and cost
parts is desirable in the ultra-precision manufacturing domain. Zidek saving of SMS [94]. The SMS should have the ability to continuously
et al. [83] discussed how to create a digital twin for an experimental coordinate the relationship between multiple components according to
assembly system based on a belt conveyor system and an automatized the characteristics of its internal structure, to ensure that the internal
line for quality production check. Liang et al. [84] presented an online coordination can always be maintained during the continuous devel
full-field displacement perceiving model by incorporating the matrix opment process. However, little attention has been paid to formally
completion theory with multi-points displacement tracking technology. define and optimize energy efficiency and carbon emission reduction in
Experiments on the assembly of large aircraft showed that the maximum the design phase of SMS [95].
displacement error could satisfy the ultra-precision manufacturing de In Industry 4.0, the digital twins technology is a powerful tool to
mands efficiently. However, these two studies did not address how to understand and optimize the interaction between all relevant material,
use the digital twins technology to simulate and optimize the product energy, and resource flows in an SMS [96]. For instance, Barni et al. [97]
manufacturing quality in the early design and configuration stages of an proposed a digital twins-based sustainability evaluation framework that
SMS. could simulate achievable results for different process plans. Wang et al.
[98] presented a digital twins-based system for supporting the design
3.6.2. Flexibility-oriented SMSD based on digital twins and recovery decisions in the waste electrical and electronic equipment
In the mass customization/individualization/personalization para recovery. Generally, the digital twins play a critical role in providing a
digm, the changeover of the SMS is extremely frequent, which calls for data-rich runtime for realizing the sustainability-oriented SMSD.
fast system design/reconfiguration technologies as well as high flexi
bility of SMS. The SMS should be able to ensure the scalability and 3.6.5. Security/Safety-oriented SMSD based on digital twins
compatibility of production to meet the needs of continuous product Process safety is a serious issue for SMS control. The digital twin can
development and services for users. predict system failures [99], which could be used to optimize the process
Many studies have been conducted on MSD that embody production safety and information flow of SMSs in the design phase. For instance,
flexibility for different product orders. In the blueprint of Industry 4.0, Becue et al. [100] presented a digital twins-based Cyber-Factory blue
each element in SMS (e.g., machine tool, workpiece, robot) is supposed print that could realize the iterative design of the safety function of
to be described with the Asset Administrative Shell technique. It is a cobots to reinforce the ability on quality monitoring of avionics elec
knowledge structure to model the functionality and interactions of each tronics for withstanding manipulation attacks. Kloibhofer et al. [101]
element in SMS, which could be recognized as the key techniques of the proposed a digital twins-based smart factory blueprint for operating
digital twins model. Seif et al. [85,86] created AASs to seamlessly with secured product configuration and secured data communication.
integrate assets into an SMS and presented a Model Factory system to Lee et al. [102] applied a system thinking methodology for imple
enhance the connectivity and the agility in a Mini Factory context. menting the digital twins to improve the safety of manufacturing sys
Gallego et al. [87] proposed a holistic capacity management model for tems. A standardized language/ontology is critical for enabling the
manufacturing organizations to change or manage their capacity to reasoning to associate modules and processes in SMS. From the control
react to all potential demand scenarios. Generally, an effective security perspective, fault tolerance is a critical capability of controllers
flexibility-oriented SMSD solution should evaluate the upgrading of to adjust system behavior to cope with faults without manual inter
products and the combination of different generations and architectures vention. Designing active control strategies could improve the capability
of products [88]. of fault tolerance of SMS, which could also be called robustness of the
system. Rodriguez et al. [103] used digital twins technology to design
3.6.3. Reconfigurability-oriented SMSD based on digital twins adaptive control strategies to increase the robustness of controllers.
The conventional methods to SMSD are usually limited in In fact, it is difficult for many manufacturers to find a real SMS that
manufacturing one single type/generation/family of products and suffer can fully realize all the above goals. And in the actual SMS, due to the
from tremendous cost for redesigning/reconfiguration when new type/ diversity and contradiction of various goals, the improvement of some
generation/family of products/models are introduced. As the upgrading performances will lead to the decline and deterioration of other per
and renewing of a product become faster, this “new-product-then-new- formance goals. For instance, flexibility and productivity are usually
system-design” logic is less ineffective. An optimized design for a given contradictory. Therefore, how to reasonably configure and coordinate
product may be inexecutable for the updated generation of the product various goals to jointly realize the overall high performance of the SMS
since the reconfigurability and adaptivity of the design are unclear in is a complicated issue. Understanding the couplings among different
terms of the indeterministic and unpredictable product changes, which design objectives is a premise for solving this issue [104]. A unified
makes it challenging to identify a design that is adaptive for different environment that could be utilized to decouple and coordinate different
type/generation/family of products/models. design objectives is favorable to increase design efficiency [105].
Digital twins technology provides a solution for reconfigurability-
oriented SMSD. For instance, Leng et al. [89] proposed a digital 4. Digital twins-based SMSD models
twins-driven fast reconfiguration method for SMSs based on the high
reconfigurability of the open-architecture machine tool. Jeon and Suh Few existing models can fully meet the design requirements of SMSs,
[90] presented a smart factory architecture based on big data and digital while two models of digital twins-based SMSD are discussed as the po
twins technology. Zhang et al. [91] presented 5D digital twin tential solution.
model-based service function block strategies to increase the reusability
of functions/algorithms in an automatic reconfiguration of 4.1. Quad-play CMCO model
robotics-based SMSs. Still, there is a need to integrate the computational
optimization techniques into the digital twins technology to improve the A similarity transferring mechanism exists among the function
design effectiveness and efficiency so that reconfigurability can be model, structure model, behavior model, control model, intelligence
automatically and intelligently optimized [92]. model, and performance model of SMS. Using this similarity mechanism
could effectively improve design efficiency. Liu et al. [106] proposed a
3.6.4. Sustainability-oriented SMSD based on digital twins digital twins-based quad-play CMCO (Configuration design-Motion
With increasing pressures to reduce energy consumption, planning-Control development-Optimization decoupling) architecture
128
for supporting the rapidly customized designing of flow-type SMSD shop-floor (Fig. 9). It mainly includes the design and configuration of a
(Fig. 7). Configuration Model was defined as the topological layout and physical shop-floor, a virtual shop-floor, and a shop-floor service system.
resource configuration of the flow-type SMS. Motion Model was defined The physical shop floor refers to the physical entities (e.g., machine
as the motion form of equipment and the logistics form of WIP in the tools) in the shop floor. The physical shop floor receives manufacturing
flow-type SMS. Control Model was defined as the data acquisition and instructions and performs manufacturing tasks. To be able to understand
processing model and control system structure of the flow-type SMS. The the status and consequence of manufacturing operations, including the
optimization model was defined as the coupled optimization in the machines and production plans required to perform specific operations,
execution of the flow-type SMS. it is necessary to design and configure the embedded system for auto
In CMCO, the Configuration Design dimension includes requirements matic identification and data perception. The design of physical
analysis, process plan, typology definition, layout plan, and the config shop-floor and virtual shop-floor refers to establishing high-precision
uration of equipment and buffer. The Motion Planning dimension in digital simulation models and interactions of various shop-floor
cludes defining the motion of the machine tool, the material handling, elements.
and the buffering/caching. The Control Development dimension includes Digital twins could help realize the correlation and dynamic
designing the bi-directional sensory and controlling mechanism. The adjustment between manufacturing planning and implementation, as
Optimization Decoupling dimension includes solving the coupled process well as promote fault prediction, diagnosis, and maintenance in a digital
optimization problems and exporting the execution engine to drive the way [107]. To establish a digital twin workshop, it is necessary to pro
SMS operation. The dynamics of an SMS during its operation undoubt vide modeling tools (e.g., Plant Simulation and Demo3D) and initialize
edly determine its performance. Usually, designers can only consider the the digital twin model of physical shop-floor using a reference model
SMS in the design and development stage based on the limited cognition with static properties, motion scripts, control schemes, and communi
of the dynamics, but actually, they use the static intellectual entity cation interfaces. On this basis, the real-time operating performance of
thinking to investigate the SMS. Concurrently conducting the SMSD the workshop could be simulated and predicted, including process
tasks and developing a unified model to effectively imitated every planning, manufacturing operations, and abnormal/error. After estab
interaction and behavior of the manufacturing process are challenging. lishing the digital model, an interoperability channel between cyber
To cope with this challenge, the CMCO model includes four typical space and physical space could be built from a data-driven view. The
iterative design paradigms (Fig. 8) that could be implemented based on design of the shop-floor service system includes establishing a data
practical design goals. The CMCO model is a bi-level architecture, in channel for cyber-physical interoperability. The development of various
which the “FW” and “FB” are two symbols to separate the physical data service applications (such as the deep neural network) makes
design level (including Configuration design and motion planning) and manufacturing data valuable for operation decision support in the
logic design level (including Control development and optimization shop-floor service system.
decoupling). The CMCO model is an iterative "execution-analysis-
adjustment" optimization for balancing the static configuration scheme 5. Key enabling technologies
and the dynamic execution engine. Therefore, the CMCO architecture is
a promising approach for effectively improving design efficiency in DT- The integration of various key enabling technologies such as the
SMSD. Industrial Internet of Things, cloud computing, artificial intelligence
algorithms makes the digital twin system more reliable and efficient for
SMSD. This section will review these key enabling technologies shown in
4.2. Digital twin shop-floor model Fig. 10.
As a basic constituent part of SMS, the design and optimization of the 5.1. Industrial internet of things
shop floor are important. Qi and Tao [107] proposed the concept of
digital twin shop-floor and discussed the connotation, configuration The digital twin of a physical process allows simulations over the
mechanism, operation mechanism, and key technologies of digital twin process in a reliable environment [108]. The Industrial Internet of
Things (IIoT), as a network composed of massive various types of sensor
and actuator nodes to realize the dynamic perceiving of the physical
system, is a carrier for feeding the digital twin with high-quality in-situ
data for high-fidelity virtual modeling and simulation computing [109].
For instance, Chhetri et al. [110] proposed an IIoT-based model to
establish digital twins by employing indirect side-channels (to percept
the energy consumptions/emissions) of the SMS, which can precisely
predict the product quality. Li et al. [111] proposed a communication
architecture for connecting heterogeneous industrial automation sub
systems in an Industry 4.0 SMS by integrating the Open Platform
Communications Unified Architecture (OPC UA) and Time-Sensitive
Networking technologies. Liu et al. [112] presented a cyber-physical
machine tools platform with standard and interoperable interfaces by
integrating the OPC UA and MTConnect technologies. Generally,
concurrently conducting the design and semi-physical commissioning
allows designers to use the IIoT data from the prototype SMS to enable
the digital twin system to validate the designs in advance in the early
development phase, and then carry out corresponding redesign or
adjustment decisions if inefficiencies and errors are identified [113].
5.2. Multi-domain physical-chemical modeling
The SMS is a multi-field physical (mechanical-electric-hydro-ther

Fig. 7. The architecture of the quad-play CMCO design model [106]. mal-magnetic-control, etc.) integrated system. The traditional way of the
129
Fig. 8. Four patterns in the iterative design logic of the quad-play CMCO model [106].
Fig. 9. Conceptual model of digital twin shop-floor [107].
manufacturing system design is that designers in various fields design design, but the models lack unified coding and can not be shared.
their relatively independent subsystems separately and then integrate Modeling work is still in the "chimney" mode of information trans
them. Although there are overall considerations, discussions, and ne mission, forming model information silos without a suitable system
gotiations in the design process, it is still difficult to grasp the complex engineering workflow. Based on the Unified Modeling Language, Model-
coupling of various parts of the system. The models of various profes Based Systems Engineering (MBSE) is a common approach for
sional disciplines have been widely used in various aspects of SMS describing SMSs. The MBSE becomes the hub of the models of various
130
Fig. 10. An overview of key enabling technologies in DT-SMSD.
professional disciplines. For instance, a software provider developed a 5.3. Real-time synchronization and discrete event simulation
corresponding support SysML (System Modeling Language) tool
modeling language, which could be integrated with existing professional The model simulation is the core technology of creating/running a
analysis software such as Finite Element Analysis and CAD. Designers digital twin and ensuring an effective closed-loop between the digital
from all parties carry out requirements analysis, system design, simu twin and corresponding physical entity. Simulation is a technology that
lation, and other work around the system model, to facilitate the simulates the physical world by transforming a model containing
collaborative work of the SMSD team. Therefore, a new generation of deterministic laws and mechanisms into the software. As long as the
digital design methods based on multi-domain modeling and multi- model is correct and fed with complete context information and envi
disciplinary simulation knowledge is needed. ronmental data, it can accurately reflect the characteristics and pa
The digital twin model of an SMS may contain models reflecting rameters of the physical world. The real-time cyber-physical
physical dynamics/characteristics in different dimensions, such as fluid, synchronization is a major difference between the digital twins method
structural, thermodynamic, stress, fatigue damage, and material state (e. and conventional simulation methods. Network latency is a key metric
g., stiffness and strength). How to correlate these dimensions is critical to to evaluate the synchronization capability of a digital twin system. For
the establishment of digital twins. Models and data are the cores of the instance, Szabo et al. [121] suggested a 5 G ultra-reliable low-latency
digital twins. The digital twin model should reflect the inner operating communication solution for establishing an accurate digital twin of a
law of the physical system. The profound understanding of the nature of robot cell. Leng et al. [122] utilized a software-comprehensible PLC
the laws, phenomena, and processes occurring in the SMSs would allow model to connect the cyber models, physical PLC, and other systems.
one to develop and apply relevant multi-domain physical-chemical Discrete Event Simulation (DES) is useful in the process flow
models. For example, the construction of a digital twin model of laser designing and planning of SMSs. Dos et al. [123] analyzed the feasibility
processing equipment should go deep into the mechanism of the inter of the DES to support building the digital twins of non-automated pro
action between laser and material. cesses. Silva et al. [124] presented a Petri Net-based DES method to
If the digital twin model has the simulation ability including the enable digital twins of SMSs. Sierla et al. [32] presented an automatic
power/spectrum/force/heat distribution and coupling relationships, it assembly planning model using UML and a 3D simulation environment
can guarantee the output design solution with better processing per employing the Automation Markup Language. To get rid of the issue of
formance. Lin et al. [114] proposed a cyber-physical SMS simulation poor-quality data from heterogeneous sources, Mieth et al. [125]
architecture based on the integration of aggregate software controllers introduced a real-time indoor localization system to provide reliable
in cyberspace and distributed hardware controllers in physical space. manufacturing data into the digital twin. To obtain a deep cognition of
Zhang et al. [115] proposed an information modeling approach for CPPS material properties of pharmaceutical powders in the manufacturing
based on the standardized format AutomationML. Ghosh et al. [116] stage, Bhalode and Ierapetritou [126] incorporated material calibration
adopted hidden Markov models for the construction and abstraction of into a DES-based feeder simulation model to achieve a particle-level
dynamics in digital twins of the manufacturing systems. Malykhina et al. understanding of the impacts of material properties on feeder perfor
[117] proposed an information-measuring system model to extract mance. These digital twin methods provided deep micro/meso/macro
knowledge from massive manufacturing information in digital twin insights into the system. Still, automatic simulation model generation is
models. Zheng et al. [118] proposed a three-level Ele critical for material flow simulation in the discrete-type SMS [127].
ment-Behavior-Rule modeling method for SMS. Kim et al. [119] pro
posed a three-stage modeling environment with an automatic module 5.4. Visualization & virtual/augmented reality
synthesis engine for supporting user-level customizing of an SMS.
Despite the remarkable achievements, further improvements of the SMSD is a complicated and prone-to-error human-computer inter
mathematical tools are associated with the need to improve the effi action process. Traditional 3D simulation systems usually cannot sup
ciency of multi-threaded calculations in solving multi-parametric design port the designers to experience the final SMS as an end-user in an
optimization problems of SMSs [120]. immersive manner. Some intelligent systems with virtual reality (VR),
augmented reality (AR) or mixed reality (MR) have been proposed to
support the engineers to design and operate. The boundary between
131
"virtual" and "real" becomes blurred in such a human-machine system 5.6. Industrial blockchains and smart contracts
supplemented by VR/AR/MR technologies [128]. Scholars are making
efforts to improve design efficiency by enhancing the digital twin system Since the SMSD involves multi-field (e.g., mechanical, electric,
with the VR/AR/MR technologies. For instance, Han et al. [129] hydro, thermal, magnetic, and control) knowledge and expertise, it is
developed a framework to extract and visualize 3D models that could be necessary to establish a reliable and interconnected runtime that enable
used in the AR/VR of an industrial plant in real time. Fera et al. [130] designers from all parties to collaboratively carry out analysis, design,
combined wearable sensors and simulation tools for analyzing and simulation work around one unified model. The existing digital twin
manufacturing performances of SMS at varying demands. Malik et al. systems are mostly centralized and suffer from the shortcomings of low
[45] developed a unified VR-enhanced human-robot simulation frame data auditability and traceability. The data security in this inter
work to estimate the human-robot cycle times and thereby helping de connected runtime is a critical issue under cyber-attack [140]. The
signers to make process plans, optimize system layouts, and program traditional cybersecurity methods need to evolve to address the attacks
controls. Perez et al. [131] presented an immersive VR-enhanced pro that threaten the collaboration network [141].
cess automation design methodology to create a digital twin of SMS for Blockchains provide a new solution for secure collaborative SMSD.
guiding its physical implementation. Generally, introducing immersive Blockchains use cryptographic technology and distributed consensus
VR/AR/MR-enhanced simulation could provide safe cyberspace for protocols to ensure the security of network transmission and access, and
commissioning and validating thus easing the SMSD. thereby to enhance the cyber-credit among various participants [142].
Much research on blockchains-empowered smart manufacturing appli
5.5. Big data analytics and industrial artificial intelligence cations in Industry 4.0 had been conducted [143,144], including
blockchains-secured data sharing in collaborative DT-SMSD. Hasan et al.
The direct goal of SMSD is to satisfy the individualized requirements, [24] proposed a blockchain-based digital twin solution to ensure
such as the site space constraint, manufacturing capacity, product auditability, accessibility, and traceability of design data. They used
quality, and production efficiency. The individualization requirements smart contracts to manage and trace data transactions in digital twins.
from the product domain are translated to the SMS domain step-by-step. Chirico et al. [145] exploited the assume-guarantee reasoning method
There exists a similarity in translating the six dimensions of the FSBCIP by using smart contracts to model a virtual production line for efficiently
model of SMS, which can be utilized to improve the design efficiency. and accurately simulating the manufacturing operations. Leng et al.
The development of comprehensive knowledge models in the digital [122] proposed an iterative bi-level hybrid intelligence framework
twin of manufacturing processes/operations is critical to make use of called ManuChain to avoid the inconsistency between planning and
these design similarities and make accurate predictions on the overall execution in individualized SMS. Spellini et al. [146] argued to use a
process dynamics for each design parameter/decision [132]. smart contract-based design mechanism to cope with the increased
The conventional knowledge-based systems could be categorized automation of SMSD. However, introducing blockchain technologies
into case-based reasoning systems and rule-based reasoning systems. A into SMSD has been greatly hindered by practical issues such as system
better way to learn in the digital intelligence age is to effectively and scalability and flexibility [147], which should be further addressed in
efficiently discover hidden knowledge and patterns in large-scale case the future.
data to enhance design efficiency [133]. There are so many underlying
interrelated factors and correlations in a complex SMS that it is often 5.7. Cloud computing and web services for distributed digital twins
difficult for domain experts to realize, which could be, however,
captured by big data analytics and deep learning methods. In short, the Besides blockchains, another kind of key enabling technology to
digital twin model with learning ability can make the SMSD more and realize the collaborative SMSD is the cloud computing and web services
more intelligent. The growing availability of data coming from historical for supporting the analysis, design, and simulation work in a distributed
design cases and contextual operations enables data analytics towards digital twin environment. In the context of collaborative SMSD, both the
the creation of added value [134]. Ringsquandl et al. [135] introduced a designers and computing resources are geographically distributed
representation learning model for synchronizing semantics from existing [148]. The cloud computing and web service systems allow to cooper
manufacturing knowledge graphs and time-dependent operational data. atively share designing activities across multiple similar units located in
Huerkamp et al. [136] presented a cyber-physical production remote areas [149]. The demand for high-performance computing for
system-based framework that integrates simulation and machine simulation work in SMSD is another driver to introducing cloud
learning techniques to form a smart digital twin. A computing technology, which could provide low-cost computing
Finite-Element-Method surrogate model was established for predicting resource sharing, dynamic allocation, and flexible expansion [150].
the interface bond strength quality independence of the process settings. Integrating web services with cloud computing enables designers to
Cavalcante et al. [137] discussed how and when machine learning and conduct the SMSD work remotely and distributed.
simulation can be combined to create digital twins for supporting the
design of risk mitigation strategies in supply disruption management 6. Digital twins-based SMSD cases
models.
However, considering the randomness in the manufacturing process, Under the increasingly diverse product demands and the evolution of
the proper design of digital twins remains difficult to enable its technologies, the manufacturing paradigm has evolved from mass pro
robustness and adaptability [138]. For instance, even small deviations of duction, mass customization to mass individualization/personalization.
the temperature settings may result in great defects in the composite The corresponding manufacturing systems are being transformed from
structure [136]. Deep learning (e.g., deep reinforcement learning), as dedicated production lines, flexible/reconfigurable manufacturing sys
one of the most promising artificial intelligence technologies [139], tems, to SMSs. Table 2 provides an overview of manufacturing system
could be used to stimulate and expand human intelligence in the SMSD. design cases for different manufacturing paradigms. The flexibility,
Today’s digital twin models are mostly limited to the understanding and autonomy, and sustainability of SMS are critical for enabling the low-
simple reasoning of design decisions of SMS. In the future, incorporating cost mass-individualized production. Open architecture is recognized
more high-level industrial artificial intelligence algorithms into the as a potential solution of SMS architecting, which is characterized by a
digital twin models could be expected to realize the ability to auto standardized platform and swappable individualized modules [89].
matedly create and generate new design solutions. Designers can flexibly redesign and reconfigure the SMS following
process planning by swapping individualized modules onto the platform
[151]. As the open architecture SMSD paradigm is enriched with the
132
Table 2 twins-based SMSD.

Manufacturing system design cases for different manufacturing paradigms.
Paradigms Mass Mass Customization Mass 7.1. Integrating the SMSD with product development
Production Individualization
Typical Mfg. Dedicated Flexible/ Smart From the product development perspective, SMSD could be catego
Systems Production Reconfigurable Manufacturing rized into the SMSD for a given product and the SMSD for a new product.
Line Manufacturing System System
SMSD and new product development are highly coupled [170], which,
Design Goal Efficiency and + Flexibility and + Sustainability,
Quality, etc. Reconfigurability, etc. etc.
however, are conventionally sequentially conducted without coordina
Mfg. System Unified Modular Architecture +Open tion [171]. The concept of digital twins promises high potentials for
Architecture Architecture Architecture integrating product development and SMSD [172]. Integrating the SMS
IP Protection Closed Source Closed Source +Open Source function, structure, behavior, and control modeling with product design
Major System +Module Reusbility, +Cost Effectivity,
could help to achieve a balance between product cost and performance
Challenges Reliability, etc. etc. etc.
Typical Mobile Phone, Furniture, Automobile Apparel, Printed efficiently and effectively from a holistic product lifecycle view [173].
Industries Chemicals Circuit Board This direction has received wide attention. Kuliaev et al. [174] pre
Ref. – [119,152] [80,89,122,151, sented a digital twins-based modeling & assembly planning framework
153]
for product-centric design. Eddy et al. [175] introduced an optimal
composition of simulation models in the product development for more
wild innovation of crowds in the community, the scale, and variety of precisely predicting the manufacturing system performance based on
module databases will grow significantly. The open architecture in the digital twins model, which could thus cut down errors in SMS
SMSD is supposed to go into a deeper open-source level ultimately commissioning at the product development phase. Ma et al. [164] pro
instead of simply build unified interfaces. Simultaneously, the intellec posed a digital twins-driven production life-cycle management system
tual property protection consensus will be updated accordingly to make framework to support a CPS of manufacturing workshop, including the
it more suitable for the distributed design and open innovations. design and manufacturing stage. Loizou et al. [152] proposed a platform
The SMS will suffer from unexpected situations with different forms that supports the consumers and the manufacturers to co-design prod
and data [154]. The design elements, design goals, and major challenges ucts based on the integration of view-based modeling, visualization, and
of different levels (e.g., machine tool/center, manufacturing units, pro monitoring technologies, and thereby creating digital twin models of
duction line, workshop, factory) of SMSD differ from each other. Table 3 SMS towards more fact-based manufacturing decisions. Schuetzer et al.
provides an overview of different levels of SMSD. The SMSD at the [176] proposed a smart digital twin model to integrate the product twin
production line level has been intensively researched. For instance, and the twin of its development process.
Gericke et al. [155] proposed a digital twins-based optimization model Actually, product development involves supply chain integration. A
for decreasing the bottleneck/delay in a water-bottling production line. product goes through a group of manufacturers involved in a series of
Liu et al. [156] presented a digital twins-driven approach for fast indi coordinated manufacturing activities before finished. Although the
vidualized designing of the furniture production line. Further, Liu et al. digital twins technology offers great superiorities in this research di
[106] utilized the designing of a hollow glass production line as a case to rection, however, many manufacturers develop their digital twins in
demonstrate the formerly-mentioned CMCO design model. Actually, the their own business processes. Existing models for the implementation of
proposed fundamental SMSD solution for this hollow glass production digital twins are limited in isolated physics domains inside one enter
line mainly includes the design of a computational decoupling model for prise. Manufacturers need continually adapt their supply chain to
a coupled optimization of "grouping-blanking-loading-planning" that changing product designs. Therefore, the development of digital twins
affects the production performance. The types of the workshop in for automating businesses towards supply chain integration is critical
Table 3 could be further categorized into job shop [29,123,127], flow but in absence. Currently, the lack of real-time data available and
shop [5,156], and open shop. For instance, Leng et al. [151] presented a responsive designing/planning systems make the adaptation of digital
digital twins-driven manufacturing cyber-physical system for a twins difficult. A unifying of digital twin models in integrating smart
board-type product manufacturing flow-shop. Despite the remarkable product development and supply chain integration with SMSD is
achievements, the existing DT-SMSD model is still insufficient to meet promising in future product R&D since it will allow us to holistically
the requirements of precisely and completely interconnecting and fusing optimize the lifecycle activities of the product [177].
the virtual model/data and physical SMS [157]. An important step to
wards the success of DT-SMSD is the establishment of practical standards 7.2. Standards for digital twins-based SMSD
and reference architectures [158].
Implementing digital twins in small and medium-size manufacturers
7. Research directions is time-consuming since the desired standards are unclear [178].
Moreover, there are often large differences between actual physical
This section suggests four directions for future research on digital manufacturing and the cyber system. Standards should be developed to
answer the question of how to realize a good DT-SMSD and to define the
Table 3
Different levels of smart manufacturing system design.
Levels Machine Tool/Center Mfg. Units Production Line Workshop Factory Distributed System
Design Parts/Components, +Robots, AGVs, etc. +Layout, Typology, + Material + Load Balancing, + Coordination
Elements Controls, etc. Performance, etc. Handling, etc. Intelligence, etc. Mechanism, etc.
Design Goal Product Quality +Autonomy +Reconfigurability +Productivity +Flexibility +Sustainability
Typical Open Architecture Autonomous Mfg. Reconfigurable Production Smart workshop, Smart Factory, etc. e.g., Decentralized
Blueprint Machine tools, etc. Units, etc. Line, etc. etc. Autonomous Mfg.
Major Manufacturing Accuracy, Deadlock Bottleneck Prevention, etc. Adaptive Layout, Compatibility with New Consensus Efficiency, etc.
Challenges etc. Prevention, etc. etc. Products, etc.
Ref. [39,41,64,75,112,159, [56,154,155,161, [5,36,130,145,163] [151,156,164] [80,86,87,90,100,132, [149,169]
160] 162] 157,165–168]
133
evaluation criteria. Such standards could be used to calibrate the dif business models to offer not only a single design solution but also ser
ference/gap between the DT-SMSD results and the final operating per vices extended to the operation phase to satisfy individualized adjust
formance of real SMS. Even if the difference/gap is large, the results can ment needs. This servitization trend has driven an IT-driven business
be considered reasonable as long as the gap is constant and can be paradigm. Similar to the concept of service-oriented manufacturing and
calculated and refined by a certain data or method. This "precision" is smart product-service system [187,188], the servitization of DT-SMSD
determined by standards instead of software, calculation methods, or aims to provide on-demand design capabilities to gain individualized
processing methods. With the help of DT-SMSD standards, the calibra satisfaction with less environmental impact by utilizing data-driven
tion deviation of design results compared to the final operating perfor digital twins as the media and tool. Servitization of DT-SMSD is a
mance of real SMS could be identified and revised. The ultimate goal of customer-oriented trend that achieves the continuous conversion of
developing DT-SMSD standards is to make sure each design result can be manufacturers’ individualized requirements into the elements, features,
reproduced and traced, instead of pursuing the design results to look and parameters of SMS. Detailed knowledge of SMSD must be provided
more like the real SMS. In another word, if we cannot keep the consis [189]. Considering the dynamic behavior and interactions in SMS [190],
tency of design results, this DT-SMSD method is unreliable. the trend is towards mining decision-support knowledge and patterns
Developing DT-SMSD standards is a process to establish the specific from large-scale historical case data to enhance SMSD services [191]
provision of technical guidelines and enforcement documents, including and context-aware value co-creation [192].
technology principles, premise assumptions, processing methods, soft
ware tools, model definitions, boundary conditions, design steps, vari 8. Concluding remarks
able controls, result evaluations, and result calibrations in DT-SMSD.
The following evaluation dimensions may be includes in the developing This article surveys how the digital twins technologies are integrated
DT-SMSD standards: 1) digital thread breadth, 2) data synchronization into and promote the SMS design based on a literature search in the Web
frequency, 3) artificial intelligence type/level, 4) simulation capabil of Science database. Based on the definitions of SMSD and the advantage
ities, 5) reference model richness, and 6) user interface [179]. Compared of DT-SMSD, a Function-Structure-Behavior-Control-Intelligence-
with the use of digital twin-based CAD software, DT-SMSD standards are Performance (FSBCIP) framework is proposed to review the major
much more complex and more difficult to form, which requires not only steps of DT-SMSD. Major design steps in SMSD that could be enhanced
a lot of test verification of the designing process but also a lot of error by the digital twins technology are reviewed. New blueprint models
calibration work. including the CMCO design architecture and digital twin shop-floor are
discussed, yet there still is an urgent need for further DT-SMSD meth
7.3. Data schema and reference architecture of digital twins-based SMSD odology development. Key enabling technologies such as IIoT, multi-
domain physical-chemical modeling, virtual reality, data analytics, in
Critical to the creation and evaluation of a DT-SMSD solution is a dustrial artificial intelligence, blockchains, and cloud computing for
digital thread that allows the continuous collection and linking of a wide supporting the DT-SMSD are analyzed. Design cases in different
range of relevant data and analytical models throughout the lifecycle of manufacturing paradigms and different manufacturing system levels are
a physical SMS. However, the lack of a universal data schema for syn presented. Based on the insights obtained from the analyses and dis
chronizing, coordinating, and integrating cross-domain digital twin cussions, we then suggest four future directions of research in DT-SMSD.
models of SMS results in a massive investment. It is critical to capture It is expected that this survey will shed new light on urgent industrial
scheme variances in the specific domains of electrics, mechanics, ergo concerns in developing new SMSs in the Industry 4.0 era.
nomics, and software and build a consistent interface standard and data
model of the digital twins [165]. This direction has also received
Declaration of Competing Interest
attention. Park et al. [180] designed a reference activity model-based
data schema for supporting the development of a cloud-based digital
The authors declare that they have no known competing financial
twin manufacturing system. Kumar et al. [181] presented a federated
interests or personal relationships that could have appeared to influence
multimodal data platform integrating modern semantic and big data
the work reported in this paper.
technologies.
Although significant strides have been made to the digital twin ar
Acknowledgments
chitecture development [153,182–186], there is still a need for
comprehensive reference architectures that can help guide the imple
This work was supported by the National Key R&D Program of China
mentation of DT-SMSD. The methodological, technological, operative,
under Grant No. 2018AAA0101704 and 2019YFB1706200; the National
and business aspects of developing, operating, and evaluating DT-SMSD
Natural Science Foundation of China under Grant No. 52075107 and
solutions should be defined in the reference architecture. The reference
U20A6004; the Science and Technology Planning Project of Guangdong
architecture answers the question of where and by what means to use
Province of China under Grant No. 2019A1515011815,
digital twins. How to define the design task in which digital twins can
2019B090916002, and 2019A050503010.
exert the greatest value and ensure correct decision guidance when
necessary is the mission of establishing reference architecture. The
References
DT-SMSD reference architecture is supposed to address requirements of
reusability, interoperability, interchangeability, maintainability, exten [1] Fowler JW, Rose O. Grand challenges in modeling and simulation of complex
sibility, and autonomy across the entire digital twin lifecycle. In addi manufacturing systems. Simulation 2004;80(9):469–76.
tion, considering that the maturity of enterprise digital twin system is [2] Rooks B. A shorter product development time with digital mock-up. Assem
Autom 1998;18(1):34–8.
gradually evolving, the reference architecture should not only stipulate
[3] Wu D, et al. Cloud-based design and manufacturing: a new paradigm in digital
"what digital twins implementation work should be done when the manufacturing and design innovation. Comput Des 2015;59:1–14.
SMSD is done with fully-mature digital twin system", but also stipulate [4] Mahmoud MA, Grace J. A generic evaluation framework of smart manufacturing
"what design functions should be tailored at different maturity levels of systems. Procedia Comput Sci 2019;161:1292–9.
[5] Zhang H, et al. A digital twin-based approach for designing and multi-objective
the digital twin system". optimization of hollow glass production line. IEEE Access 2017;5:26901–11.
[6] Wang K, Lee T, Hsu Y. Revolution on digital twin technology-a patent research
7.4. Servitization of digital twins-based SMSD approach. Int J Adv Manuf Technol 2020;107(11–12):4687–704.
[7] Nujoom R, Mohammed A, Wang Q. A sustainable manufacturing system design: a
fuzzy multi-objective optimization model. Environ Sci Pollut Res - Int 2018;25
Servitization of SMSD is a trend in industries, which adopts service (25):24535–47.
134
[8] Choi S, et al. Cyber-physical systems: a case study of development for [43] BAINES TS, KAY JM. Human performance modelling as an aid in the process of
manufacturing industry. Int J Comput Appl Technol 2017;55(4):289–97. manufacturing system design: a pilot study. Int J Prod Res 2002;40(10):2321–34.
[9] Frontoni E, et al. In: DePaolis LT, Bourdot P, DePaolis LT, Bourdot P, editors. [44] Malik AA, Brem A. Digital twins for collaborative robots: a case study in human-
Cyber physical systems for industry 4.0: towards real time virtual reality in smart robot interaction. Robot Comput Integr Manuf 2021;68:102092.
manufacturing, in lecture notes in computer science. CHAM: Springer [45] Malik AA, Masood T, Bilberg A. Virtual reality in manufacturing: immersive and
International Publishing Ag: Cham; 2018. p. 422–34. collaborative artificial-reality in design of human-robot workspace. Int J Comput
[10] Zakoldaev DA, et al. In: Sapozhkov SB, Sapozhkov SB, editors. Digital forms of Integr Manuf 2020;33(1):22–37.
describing Industry 4.0 objects, in IOP Conference series-materials science and [46] Graessler I, Poehler A. Integration of a digital twin as human representation in a
engineering. Bristol: Iop Publishing Ltd; 2019. scheduling procedure of a cyber-physical production system. IEEE; 2017.
[11] Negri E, Fumagalli L, Macchi M. A review of the roles of Digital Twin in CPS- [47] Latif H, Starly B. A simulation algorithm of a digital twin for manual assemble
based production systems. Procedia Manuf 2017;11:939–48. process. Procedia Manuf 2020;48:932–9.
[12] Shao GD, Kibira D. Digital manufacturing: requirements and challenges for [48] Greco A, et al. Digital twin for monitoring ergonomics during manufacturing
implementing digital surrogates. In: Winter Simulation Conference Proceedings; production. Appl Sci 2020;10(775821).
2018. p. 1226–37. [49] Graessler I, Poehler A. Intelligent control of an assembly station by integration of
[13] Valckenaers P. Perspective on holonic manufacturing systems: PROSA becomes a digital twin for employees into the decentralized control system. Procedia
ARTI. Comput Ind 2020;120(103226). Manuf 2018;24:185–9.
[14] Barricelli BR, Casiraghi E, Fogli D. A survey on digital twin: definitions, [50] Malik AA, Bilberg A. Digital twins of human robot collaboration in a production
characteristics, applications, and design implications. IEEE Access 2019;7: setting. Procedia Manuf 2018;17:278–85.
167653–71. [51] Frazzon EM, et al. Manufacturing networks in the era of digital production and
[15] Zobel-Roos S, et al. Accelerating biologics manufacturing by modeling or: is operations: a socio-cyber-physical perspective. Annu Rev Control 2020;49:
approval under the qbd and pat approaches demanded by authorities acceptable 288–94.
without a digital-twin? Processes 2019;7(942). [52] Bučková M, et al. Designing of logistics systems with using of computer
[16] Chen Y, et al. Digital twins in pharmaceutical and biopharmaceutical simulation and emulation. Transp Res Procedia 2019;40:978–85.
manufacturing: a literature review. Processes 2020;8(10889). [53] Wang W, Zhang Y, Zhong RY. A proactive material handling method for CPS
[17] Tao F, et al. Digital twin in industry: state-of-the-art. IEEE Trans Ind Inform 2019; enabled shop-floor. Robot Comput Integr Manuf 2020;61(101849).
15(4):2405–15. [54] Banyai A, et al. Smart cyber-physical manufacturing: extended and real-time
[18] Lu Y, et al. Digital twin-driven smart manufacturing: connotation, reference optimization of logistics resources in matrix production. Appl Sci 2019;9(12877).
model, applications and research issues. Robot Comput Integr Manuf 2020;61 [55] Yerra VA, Pilla S. IIoT-enabled production system for composite intensive vehicle
(101837). manufacturing. SAE Int J Engines 2017;10(2):209–14.
[19] Yan J, Li B. Research hotspots and tendency of intelligent manufacturing. Chin Sci [56] Zhu CY, Pires JN, Azar A. A novel multi-brand robotic software interface for
Bull-Chin 2020;65(8):684–94. industrial additive manufacturing cells. Ind Robot 2020;47(4):581–92.
[20] Cimino C, Negri E, Fumagalli L. Review of digital twin applications in [57] Cao X, Zhao G, Xiao W. Digital Twin-oriented real-time cutting simulation for
manufacturing. Comput Ind 2019;113(103130). intelligent computer numerical control machining. Proceed Instit Mech Eng Part
[21] Qi Q, Tao F. Digital twin and big data towards smart manufacturing and Industry B 2020. 0954405420937869.
4.0: 360 degree comparison. IEEE Access 2018;6:3585–93. [58] Zhao R, et al. Digital twin-driven cyber-physical system for autonomously
[22] Tao F, et al. Digital twins and cyber-physical systems toward smart controlling of Micro Punching system. IEEE Access 2019;7(1):9459–69.
manufacturing and Industry 4.0: correlation and comparison. Engineering 2019;5 [59] Gohari H, Berry C, Barari A. A digital twin for integrated inspection system in
(4):653–61. digital manufacturing. IFAC Papersonline 2019;52(10):182–7.
[23] Konstantinov S, et al. The cyber-physical E-machine manufacturing system: [60] Tabar RS, et al. Efficient spot welding sequence optimization in a geometry
virtual engineering for complete lifecycle support. Procedia CIRP 2017;63: assurance digital twin. J Mech Des 2020;142:10200110.
119–24. [61] Schleich B, et al. Geometrical variations management 4.0: towards next
[24] Hasan HR, et al. A blockchain-based approach for the creation of digital twins. generation geometry assurance. Procedia CIRP 2018;75:3–10.
IEEE Access 2020;8:34113–26. [62] Schleich B, et al. Shaping the digital twin for design and production engineering.
[25] Qian L, Gero JS, et al. Function-behavior-structure paths and their role in CIRP Ann Manuf Technol 2017;66(1):141–4.
analogy-based design. AI EDAM 1996;10(4):289–312. [63] Cerrone A, et al. On the effects of modeling as-manufactured geometry: toward
[26] Zhang Y, et al. Intelligent feature recognition for STEP-NC-compliant digital twin. Int J Aerosp Eng 2014;(439278). 2014.
manufacturing based on artificial bee colony algorithm and back propagation [64] Montavon B, et al. A digital perspective on machine tool calibration. Int J Autom
neural network. J Manuf Syst 2021. https://doi.org/10.1016/j. Technol 2020;14(3SI):360–8.
jmsy.2021.01.018). [65] Berger T, et al. Arezzo-flexible manufacturing system: a generic flexible
[27] Zhao P, et al. The modeling and using strategy for the digital twin in process manufacturing system shop floor emulator approach for high-level control virtual
planning. IEEE Access 2020;8:41229–45. commissioning. Concurr Eng Res Appl 2015;23(4):333–42.
[28] Uhlemann T, Lehmann C, Steinhilper R. The digital twin: realizing the cyber- [66] Julius R, et al. Transformation of GRAFCET to PLC code including hierarchical
physical production system for industry 4.0. Procedia CIRP 2017;61:335–40. structures. Control Eng Pract 2017;64:173–94.
[29] Lu Y, et al. An IoT-enabled simulation approach for process planning and [67] Niang M, et al. A methodology for automatic generation, formal verification and
analysis: a case from engine re-manufacturing industry. Int J Comput Integr implementation of safe PLC programs for power supply equipment of the electric
Manuf 2019;32(4–5SI):413–29. lines of railway control systems. Comput Ind 2020;123:103328.
[30] Liu J, et al. Dynamic evaluation method of machining process planning based on [68] Orive D, et al. Fault injection in Digital Twin as a means to test the response to
digital twin. IEEE Access 2019;7:19312–23. process faults at virtual commissioning. In: IEEE International Conference on
[31] Liu J, et al. Dynamic design method of digital twin process model driven by Emerging Technologies and Factory Automation-ETFA; 2019. p. 1230–4.
knowledge-evolution machining features. Int J Prod Res 2021:1–19. https://doi. [69] Toivonen V, et al. The FMS Training Center - a versatile learning environment for
org/10.1080/00207543.2021.1887531. engineering education. Procedia Manuf 2018;23:135–40.
[32] Sierla S, et al. Automatic assembly planning based on digital product descriptions. [70] Leng J, et al. Digital twins-based remote semi-physical commissioning of flow-
Comput Ind 2018;97:34–46. type smart manufacturing systems. J Clean Prod 2021;306:127278.
[33] Haber Guerra R, et al. Digital twin-based optimization for ultraprecision motion [71] Talkhestani BA, Schloegl W, Weyrich M. Synchronization of digital models -
systems with backlash and friction. IEEE Access 2019;7:93462–72. Application of an anchor-point method for production cells. ATP Edition 2017;
[34] Guo D, et al. Digital twin-enabled graduation intelligent manufacturing system (7–8):62–9.
for fixed-position assembly islands. Robot Comput Integr Manuf 2020;63 [72] Talkhestani BA, et al. Consistency check to synchronize the Digital Twin of
(101917). manufacturing automation based on anchor points. Procedia CIRP 2018;72:
[35] Guo D, et al. Graduation Intelligent Manufacturing System (GiMS): an Industry 159–64.
4.0 paradigm for production and operations management. Ind Manag Data Syst [73] Kang S, Chun I, Kim H. Design and implementation of runtime verification
2020. framework for cyber-physical production systems. J Eng 2019;(2875236). 2019.
[36] Liu J, et al. A digital twin-based approach for dynamic clamping and positioning [74] Leng J, et al. Contextual self-organizing of mass individualization process under
of the flexible tooling system. Procedia CIRP 2019;80:746–9. social manufacturing paradigm: a cyber-physical-social system approach. Enterp
[37] Wang P, et al. Digital twin-driven system for roller conveyor line: design and Inf Syst 2020;14(8):1124–49.
control. J Ambient Intell Humaniz Comput 2020. [75] Wojcicki J, Bianchi G. A Smart Spindle Component concept as a standalone
[38] Vatankhah Barenji A, et al. A digital twin-driven approach towards smart measurement system for Industry 4.0 Machine Tools. New York, USA: IEEE; 2020.
manufacturing: reduced energy consumption for a robotic cellular. Int J Comput p. 278–82.
Integr Manuf 2020. [76] Sun B, et al. A comprehensive hybrid first principles/machine learning modeling
[39] Cai Y, et al. Sensor data and information fusion to construct digital-twins virtual framework for complex industrial processes. J Process Control 2020;86:30–43.
machine tools for cyber-physical manufacturing. Procedia Manuf 2017;10: [77] Orosz T, et al. Robust design optimization and emerging technologies for
1031–42. electrical machines: challenges and open problems. Appl Sci 2020;10(665319).
[40] Cai Y, Wang Y, Burnett M. Using augmented reality to build digital twin for [78] Leng J, et al. Digital twin-driven joint optimisation of packing and storage
reconfigurable additive manufacturing system. J Manuf Syst 2020;56:598–604. assignment in large-scale automated high-rise warehouse product-service system.
[41] Tong X, et al. Real-time machining data application and service based on IMT Int J Comput Integr Manuf 2019:1–18.
digital twin. J Intell Manuf 2020;31(5):1113–32.
[42] Botkina D, et al. Digital twin of a cutting tool. Procedia CIRP 2018;72:215–8.
135
[79] Lugaresi G, Travaglini D, Matta A. In: A LEGO (R) Manufacturing System as [114] Lin WD, et al. Integrated cyber physical simulation modelling environment for
Demonstrator for a Real-Time Simulation Proof of Concept, in Winter Simulation manufacturing 4.0. In: International Conference on Industrial Engineering and
Conference Proceedings; 2019. p. 2025–36. Engineering Management IEEM; 2018. p. 1861–5.
[80] Park KT, et al. Design and implementation of a digital twin application for a [115] Zhang H, Yan Q, Wen Z. Information modeling for cyber-physical production
connected micro smart factory. Int J Comput Integr Manuf 2019;32(6):596–614. system based on digital twin and AutomationML. Int J Adv Manuf Technol 2020;
[81] Gu P, Rao HA, Tseng MM. Systematic design of manufacturing systems based on 107(3–4):1927–45.
axiomatic design approach. CIRP Ann-Manuf Technol 2001;50(1):299–304. [116] Ghosh AK, Ullah AMMS, Kubo A. Hidden Markov model-based digital twin
[82] Luo C, et al. A framework for tolerance modeling based on parametric space construction for futuristic manufacturing systems. Artif Intell Eng Des Anal Manuf
envelope. J Manuf Sci Eng-Trans ASME 2020;142(0610076). 2019;33(3):317–31.
[83] Zidek K, et al. Digital twin of experimental smart manufacturing assembly system [117] Malykhina GF, Vasilyev AN, Tarkhov DA. Information measurement systems in
for industry 4.0 concept. Sustainability 2020;12(36589). the digital society. J Phys Conf Ser 2019;1379(1):12037.
[84] Liang B, et al. A displacement field perception method for component digital twin [118] Zheng Y, et al. Digital twin for geometric feature online inspection system of car
in aircraft assembly. Sensors 2020;20(516118). body-in-white. Int J Comput Integr Manuf 2020. https://doi.org/10.1080/
[85] Seif A, Toro C, Akhtar H. Implementing Industry 4.0 asset administrative shells in 0951192X.2020.1736637.
mini factories. Procedia Comput Sci 2019;159:495–504. [119] Kim BS, Jin Y, Nam S. An integrative user-level customized modeling and
[86] Toro C, Seif A, Akhtar H. Modeling and connecting asset administrative shells for simulation environment for smart manufacturing. IEEE Access 2019;7:
mini factories. Cybern Syst 2020;51(2SI):232–45. 186637–45.
[87] Gallego-Garcia S, Reschke J, Garcia-Garcia M. Design and simulation of a [120] Golubev VO, et al. Systems and aids of mathematical modeling of the alumina
capacity management model using a digital twin approach based on the viable refinery methods: problems and solutions. Non-Ferrous Met 2019;(1):40–7.
system model: case study of an automotive plant. Appl Sci 2019;9(556724). [121] Szabo G, et al. Digital twin: network provisioning of mission critical
[88] Ko J, Hu SJ. Manufacturing system design considering stochastic product communication in cyber physical production systems. New York, USA: IEEE;
evolution and task recurrence. J Manuf Sci Eng-Trans ASME 2009;131(0510125). 2019. p. 37–43.
[89] Leng J, et al. Digital twin-driven rapid reconfiguration of the automated [122] Leng J, et al. ManuChain: combining permissioned blockchain with a holistic
manufacturing system via an open architecture model. Robot Comput Integr optimization model as Bi-Level intelligence for smart manufacturing. IEEE Trans
Manuf 2020;63(101895). Syst Man Cybern-Syst 2020;50(1):182–92.
[90] Jeon B, Suh S. Design considerations and architecture for cooperative smart [123] Dos Santos CH, et al. Use of simulation in the industry 4.0 context: creation of a
factory: MAPE/BD approach. Procedia Manuf 2018;26:1094–106. Digital Twin to optimise decision making on non-automated process. J Simul
[91] Zhang C, et al. Digital twin-enabled reconfigurable modeling for smart 2020.
manufacturing systems. Int J Comput Integr Manuf 2019. https://doi.org/ [124] Silva JM, Javales R, Silva JR. A new requirements engineering approach for
10.1080/0951192X.2019.1699256. manufacturing based on petri nets. IFAC Papersonline 2019;52(10):97–102.
[92] Kurniadi KA, Lee S, Ryu K. Digital twin approach for solving reconfiguration [125] Mieth C, Meyer A, Henke M. Framework for the usage of data from real-time
planning problems in RMS. IFIP advances in information and communication indoor localization systems to derive inputs for manufacturing simulation.
technology. Cham: Springer International Publishing; 2018. p. 327–34. Procedia CIRP 2019;81:868–73.
[93] Mawson VJ, Hughes BR. The development of modelling tools to improve energy [126] Bhalode P, Ierapetritou M. Discrete element modeling for continuous powder
efficiency in manufacturing processes and systems. J Manuf Syst 2019;51:95–105. feeding operation: calibration and system analysis. Int J Pharm 2020;585
[94] Xiao Q, et al. Meta-reinforcement learning of machining parameters for energy- (119427).
efficient process control of flexible turning operations. IEEE Trans Autom Sci Eng [127] Reinhardt H, Weber M, Putz M. A survey on automatic model generation for
2021;18(1):5–18. material flow simulation in discrete manufacturing. Procedia CIRP 2019;81:
[95] Supekar SD, et al. A framework for quantifying energy and productivity benefits 121–6.
of smart manufacturing technologies. Procedia CIRP 2019;80:699–704. [128] Simoes B, et al. Cross reality to enhance worker cognition in industrial assembly
[96] Ma S, et al. Data-driven sustainable intelligent manufacturing based on demand operations. Int J Adv Manuf Technol 2019;105(9SI):3965–78.
response for energy-intensive industries. J Clean Prod 2020;274(123155). [129] Han Y, et al. 3D CAD data extraction and conversion for application of
[97] Barni A, et al. Exploiting the digital twin in the assessment and optimization of augmented/virtual reality to the construction of ships and offshore structures. Int
sustainability performances. New York, USA: IEEE; 2018. J Comput Integr Manuf 2019;32(7):658–68.
[98] Wang XV, Wang L. Digital twin-based WEEE recycling, recovery and [130] Fera M, et al. Towards digital twin implementation for assessing production line
remanufacturing in the background of Industry 4.0. Int J Prod Res 2019;57(12SI): performance and balancing. Sensors 2020;20(971).
3892–902. [131] Perez L, et al. Digital twin and virtual reality based methodology for multi-robot
[99] He B, Bai K. Digital twin-based sustainable intelligent manufacturing: a review. manufacturing cell commissioning. Appl Sci 2020;10(363310).
Adv Manuf 2020. [132] Taylor C, et al. Defining production and financial data streams required for a
[100] Becue A, et al. CyberFactory#1-Securing the Industry 4.0 with cyber-ranges and factory digital twin to optimise the deployment of labour. Communications in
digital twins. New York, USA: IEEE; 2018. computer and information science. Singapore: Springer Singapore; 2018. p. 3–12.
[101] Kloibhofer R, Kristen E, Jakšić S. Safety and security in a smart production [133] Moyne J, Iskandar J. Big data analytics for smart manufacturing: case studies in
environment. Lecture notes in computer science. Cham: Springer International semiconductor manufacturing. Processes 2017;5(393).
Publishing; 2018. p. 190–201. [134] Gao RX, et al. Big data analytics for smart factories of the future. CIRP Ann-Manuf
[102] Lee J, Cameron I, Hassall M. Improving process safety: what roles for Technol 2020;69(2):668–92.
digitalization and Industry 4.0. Process Saf Environ Prot 2019;132:325–39. [135] Ringsquandl M, et al. Knowledge fusion of manufacturing operations data using
[103] Rodriguez-Guerra J, et al. On fault-tolerant control systems: a novel representation learning. In: Lodding H, Lodding H, editors. IFIP advances in
reconfigurable and adaptive solution for industrial machines. IEEE Access 2020;8: information and communication technology. Cham: Springer International
39322–35. Publishing; 2017. p. 302–10.
[104] Cochran DS, et al. A decomposition approach for manufacturing system design. [136] Huerkamp A, et al. Combining simulation and machine learning as digital twin for
J Manuf Syst 2001;20(6):371–89. the manufacturing of overmolded thermoplastic composites. J Manuf Mater
[105] Hibino H, et al. The development of an object-oriented simulation system basedon Process 2020;4(923).
the thought process of the manufacturing system design. Int J Prod Econ 1999; [137] Cavalcante IM, et al. A supervised machine learning approach to data-driven
(60–61):343–51. simulation of resilient supplier selection in digital manufacturing. Int J Inf
[106] Liu Q, et al. Digital twin-based designing of the configuration, motion, control, Manage 2019;49:86–97.
and optimization model of a flow-type smart manufacturing system. J Manuf Syst [138] Wang J, et al. Digital Twin for rotating machinery fault diagnosis in smart
2021;58:52–64. manufacturing. Int J Prod Res 2019;57(12SI):3920–34.
[107] Tao F, Zhang M. Digital twin shop-floor: a new shop-floor paradigm towards [139] Leng J, et al. A loosely-coupled deep reinforcement learning approach for order
smart manufacturing. IEEE Access 2017;5:20418–27. acceptance decision of mass-individualized printed circuit board manufacturing
[108] Souza V, et al. A digital twin architecture based on the industrial internet of in Industry 4.0. J Clean Prod 2021;280(2):124405.
things technologies. In: IEEE International Symposium on Consumer Electronics; [140] Danilczyk W, Sun YL, He H. In: ANGEL: An Intelligent Digital Twin Framework
2019. for Microgrid Security, in North American Power Symposium; 2019.
[109] Cheng J, et al. DT-II:digital twin enhanced Industrial Internet reference [141] Gupta N, et al. Additive manufacturing cyber-physical system: supply chain
framework towards smart manufacturing. Robot Comput Integr Manuf 2020;62 cybersecurity and risks. IEEE Access 2020;8:47322–33.
(101881). [142] Leng J, et al. Makerchain: a blockchain with chemical signature for self-
[110] Chhetri SR, et al. In: Ramachandran GS, Ortiz J, Ramachandran GS, Ortiz J, organizing process in social manufacturing. J Clean Prod 2019;234:767–78.
editors. QUILT: quality inference from living digital twins in IoT-enabled [143] Leng J, et al. Blockchain-secured smart manufacturing in Industry 4.0: a survey.
manufacturing systems. New York, USA: Assoc Computing Machinery; 2019. IEEE Trans Syst Man Cybern Syst 2021;51(1):237–52.
p. 237–48. [144] Leng J, et al. Blockchain-empowered sustainable manufacturing and product
[111] Li Y, et al. Practical implementation of an OPC UA TSN communication lifecycle management in industry 4.0: a survey. Renew Sustain Energy Rev 2020;
architecture for a manufacturing system. IEEE Access 2020;8:200100–11. 132:110112.
[112] Liu C, et al. A cyber-physical machine tools platform using OPC UA and [145] Chirico R, et al. A contract-based methodology for production lines validation. In:
MTConnect. J Manuf Syst 2019;51:61–74. IEEE International Conference on Industrial Informatics; 2019. p. 695–8.
[113] Zhang K, et al. Digital twin-based opti-state control method for a synchronized [146] Spellini S, et al. Languages and formalisms to enable EDA techniques in the
production check toroperation system. Robot Comput Integr Manuf 2020:63. context of industry 4.0.. New York, USA: IEEE; 2019.
136
[147] Leng J, et al. Blockchain security: a survey of techniques and research directions. [170] Tomiyama T, et al. Development capabilities for smart products. CIRP Ann-Manuf
IEEE Trans Serv Comput 2020:1. Technol 2019;68(2):727–50.
[148] Leng J, Jiang P. Evaluation across and within collaborative manufacturing [171] Steimer C, Aurich JC. Analysis of information interdependencies between product
networks: a comparison of manufacturers’ interactions and attributes. Int J Prod development and manufacturing system planning in early design phases. Procedia
Res 2018;56(15):5131–46. CIRP 2016;50:460–5.
[149] Avventuroso G, Silvestri M, Pedrazzoli P. A networked production system to [172] Schleich B, et al. Shifting value stream patterns along the product lifecycle with
implement virtual enterprise and product lifecycle information loops. IFAC digital twins. Procedia CIRP 2019;86:3–11.
Papersonline 2017;50(1):7964–9. [173] Promyoo R, Alai S, El-Mounayri H. Innovative digital manufacturing curriculum
[150] Cohen Y, et al. Design and management of digital manufacturing and assembly for industry 4.0. Procedia Manuf 2019;34:1043–50.
systems in the Industry 4.0 era. Int J Adv Manuf Technol 2019;105(9SI):3565–77. [174] Kuliaev V, et al. Towards product centric manufacturing: from digital twins to
[151] Leng J, et al. Digital twin-driven manufacturing cyber-physical system for parallel product assembly. In: IEEE International Conference on Industrial Informatics;
controlling of smart workshop. J Ambient Intell Humaniz Comput 2019;10(3SI): 2019. p. 164–71.
1155–66. [175] Eddy D, et al. Framework for design from manufacturing data mapping. New
[152] Loizou S, et al. In: Filipe J, editor. A smart product Co-design and monitoring York, USA: ASME; 2020.
framework via gamification and complex event processing; 2019. p. 237–44. [176] Schuetzer K, et al. Contribution to the development of a Digital Twin based on
Setubal, Portugal. product lifecycle to support the manufacturing process. Procedia CIRP 2019;84:
[153] Park KT, et al. Digital twin-based cyber physical production system architectural 82–7.
framework for personalized production. Int J Adv Manuf Technol 2020;106(5–6): [177] Wagner R, et al. Challenges and potentials of digital twins and Industry 4.0 in
1787–810. product design and production for high performance products. Procedia CIRP
[154] He YC, Zhang NS, Wang AM. Digital twin process and simulation operation 2019;84:88–93.
control technology for intelligent manufacturing unit. In: IOP Conference Series- [178] Kurfess TR, et al. A review of modern communication technologies for digital
Materials Science and Engineering; 2020. manufacturing processes in industry 4.0. J Manuf Sci Eng-Trans ASME 2020;142:
[155] Gericke GA, et al. Design of digital twins for optimization of a water bottling 11081511.
plant. IEEE industrial electronics society. New York: IEEE; 2019. p. 5204–10. [179] Bevilacqua M, et al. Digital twin reference model development to prevent
[156] Liu Q, et al. Digital twin-driven rapid individualised designing of automated flow- operators’ risk in process plants. Sustainability 2020;12(10883).
shop manufacturing system. Int J Prod Res 2019;57(12SI):3903–19. [180] Park Y, Woo J, Choi S. A cloud-based digital twin manufacturing system based on
[157] Chen G, et al. The framework design of smart factory in discrete manufacturing an interoperable data Schema for smart manufacturing. Int J Comput Integr
industry based on cyber-physical system. Int J Comput Integr Manuf 2020;33(1): Manuf 2020.
79–101. [181] Kumar VS, et al. A federated, multimodal digital thread platform for enabling
[158] Redelinghuys AJH, Basson AH, Kruger K. A six-layer architecture for the digital digital twins. Nav Eng J 2020;132(1):47–56.
twin: a manufacturing case study implementation. J Intell Manuf 2020;31(6): [182] Ahuett-Garza H, Urbina Coronado PD. A reference model for evolving digital
1383–402. twins and its application to cases in the manufacturing floor. Smart Sustain Manuf
[159] Jeon B, et al. The architecture development of Industry 4.0 compliant smart Syst 2019;3(2):1–13.
machine tool system (SMTS). J Intell Manuf 2020;31(8):1837–59. [183] Catarci T, et al. In: Bertino E, editor. A conceptual architecture and model for
[160] Wang K, Lee Y, Angelica S. Digital twin design for real-time monitoring - a case smart manufacturing relying on service-based digital twins. New York, USA: IEEE;
study of die cutting machine. Int J Prod Res 2020. https://doi.org/10.1080/ 2019. p. 229–36.
00207543.2020.1817999. [184] Damjanovic-Behrendt V, Behrendt W. An open source approach to the design and
[161] Wang B, et al. Intelligent welding system technologies: state-of-the-art review and implementation of Digital Twins for Smart Manufacturing. Int J Comput Integr
perspectives. J Manuf Syst 2020;56:373–91. Manuf 2019;32(4–5SI):366–84.
[162] Zhang X, Zhu W. Application framework of digital twin-driven product smart [185] Moyne J, et al. A requirements driven digital twin framework: specification and
manufacturing system: a case study of aeroengine blade manufacturing. Int J Adv opportunities. IEEE Access 2020;8:107781–801.
Robot Syst 2019;16. 17298814198806635. [186] Stark R, Fresemann C, Lindow K. Development and operation of Digital Twins for
[163] Lin WD, Low MYH. Concept and implementation of a cyber-physical digital twin technical systems and services. CIRP Ann-Manuf Technol 2019;68(1):129–32.
for a SMT line. In: International Conference on Industrial Engineering and [187] Zhang G, et al. An integrated framework for active discovery and optimal
Engineering Management; 2019. p. 1455–9. allocation of smart manufacturing services. J Clean Prod 2020;273(123144).
[164] Ma J, et al. A digital twin-driven production management system for production [188] Zheng P, et al. A systematic design approach for service innovation of smart
workshop. Int J Adv Manuf Technol 2020;110(5–6):1385–97. product-service systems. J Clean Prod 2018;201:657–67.
[165] Talkhestani BA, et al. A concept in synchronization of virtual production system [189] Olivotti D, et al. Creating the foundation for digital twins in the manufacturing
with real factory based on anchor-point method. Procedia CIRP 2018;67:13–7. industry: an integrated installed base management system. Inf Syst E-Business
[166] Nafors D, et al. Application of a hybrid digital twin concept for factory layout Manag 2019;17(1):89–116.
planning. Smart Sustain Manuf Syst 2020;4(2SI):231–44. [190] Li Y, Li L. Enhancing the optimization of the selection of a product service system
[167] Cheng Y, et al. Cyber-physical integration for moving digital factories forward scheme: a digital twin-driven framework. Strojniski Vestnik-J Mech Eng 2020;66
towards smart manufacturing: a survey. Int J Adv Manuf Technol 2018;97(1–4): (9):534–43.
1209–21. [191] Leng J, et al. Combining granular computing technique with deep learning for
[168] Longo F, Nicoletti L, Padovano A. Ubiquitous knowledge empowers the Smart service planning under social manufacturing contexts. Knowl Based Syst 2018;
Factory: the impacts of a Service-oriented Digital Twin on enterprises’ 143:295–306. https://doi.org/10.1016/j.knosys.2017.07.023.
performance. Annu Rev Control 2019;47:221–36. [192] Cong J, et al. A holistic relook at engineering design methodologies for smart
[169] Schimanski CP, et al. Pushing digital automation of configure-to-order services in product-service systems development. J Clean Prod 2020;272(122737).
small and medium enterprises of the construction equipment industry: a design
science research. Appl Sci 2019;9(8):3780.
137

178 - Digital Twins-Based Smart Manufacturing System Design in Industry 4.0A Review

Uploaded by

Copyright:

Available Formats

178 - Digital Twins-Based Smart Manufacturing System Design in Industry 4.0A Review

Uploaded by

Document Information

Original Description:

Original Title

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

178 - Digital Twins-Based Smart Manufacturing System Design in Industry 4.0A Review

Uploaded by

Copyright:

Available Formats

Journal of Manufacturing Systems 60 (2021) 119–137

Contents lists available at ScienceDirect

Journal of Manufacturing Systems

Digital twins-based smart manufacturing system design in Industry 4.0:

Period 1760s~1850s 1860s~1960s 1970s~2010s 2010s~Now

Fig. 1. Statistics of collected literature.

Fig. 2. The organization of this survey.

Fig. 3. The evolution of simulation technology from the design view.

Fig. 5. An FSBCIP framework for smart manufacturing system design.

Fig. 6. The rationale of digital twins-based system controls commissioning [70].

5.2. Multi-domain physical-chemical modeling

The SMS is a multi-field physical (mechanical-electric-hydro-ther

Fig. 9. Conceptual model of digital twin shop-floor [107].

Fig. 10. An overview of key enabling technologies in DT-SMSD.

Table 2 twins-based SMSD.

You might also like

178 - Digital Twins-Based Smart Manufacturing System Design in Industry 4.0A Review

Uploaded by

Copyright:

Available Formats

178 - Digital Twins-Based Smart Manufacturing System Design in Industry 4.0A Review

Uploaded by

Document Information

Original Description:

Original Title

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

178 - Digital Twins-Based Smart Manufacturing System Design in Industry 4.0A Review

Uploaded by

Copyright:

Available Formats

Journal of Manufacturing Systems 60 (2021) 119–137

Contents lists available at ScienceDirect

Journal of Manufacturing Systems

Digital twins-based smart manufacturing system design in Industry 4.0:

Period 1760s~1850s 1860s~1960s 1970s~2010s 2010s~Now

Fig. 1. Statistics of collected literature.

Fig. 2. The organization of this survey.

Fig. 3. The evolution of simulation technology from the design view.

Fig. 5. An FSBCIP framework for smart manufacturing system design.

Fig. 6. The rationale of digital twins-based system controls commissioning [70].

5.2. Multi-domain physical-chemical modeling

The SMS is a multi-field physical (mechanical-electric-hydro-ther­

Fig. 9. Conceptual model of digital twin shop-floor [107].

Fig. 10. An overview of key enabling technologies in DT-SMSD.

Table 2 twins-based SMSD.

You might also like

The SMS is a multi-field physical (mechanical-electric-hydro-ther