Journal of Manufacturing Systems 56 (2020) 312–325

Contents lists available at ScienceDirect

Journal of Manufacturing Systems

journal homepage: www.elsevier.com/locate/jmansys

Review

Smart manufacturing process and system automation – A critical review of T

the standards and envisioned scenarios
Yuqian Lua,*, Xun Xua, Lihui Wangb
a
Department of Mechanical Engineering, The University of Auckland, New Zealand
b
KTH Royal Institute of Technology, Stockholm, Sweden

A R T I C LE I N FO A B S T R A C T

Keywords: Smart manufacturing is arriving. It promises a future of mass-producing highly personalized products via re-
Smart manufacturing sponsive autonomous manufacturing operations at a competitive cost. Of utmost importance, smart manu-
Standard facturing requires end-to-end integration of intra-business and inter-business manufacturing processes and
Automation systems. Such end-to-end integration relies on standards-compliant and interoperable interfaces between dif-
Mass personalization
ferent manufacturing stages and systems. In this paper, we present a comprehensive review of the current
Industry 4.0
landscape of manufacturing automation standards, with a focus on end-to-end integrated manufacturing pro-
Self-organizing manufacturing network
Digital thread cesses and systems towards mass personalization and responsive factory automation. First, we present an au-
thentic vision of smart manufacturing and the unique needs for next-generation manufacturing automation. A
comprehensive review of existing standards for enabling manufacturing process automation and manufacturing
system automation is presented. Subsequently, focusing on meeting changing demands of efficient production of
highly personalized products, we detail several future-proofing manufacturing automation scenarios via in-
tegrating various existing standards. We believe that existing automation standards have provided a solid
foundation for developing smart manufacturing solutions. Faster, broader and deeper implementation of smart
manufacturing automation can be anticipated via the dissemination, adoption, and improvement of relevant
standards in a need-driven approach.

1. Introduction end-to-end engineering integration.

Among all the recognized research gaps [3], standards-compliant
Today’s manufacturers face ever-increasing demands of product interoperability and integration between manufacturing processes and
variability (i.e., personalization), smaller lot sizes, and turbulent market systems are of utmost importance [7]. With an overwhelming portfolio
needs. Manufacturers will have to adopt new technologies and manu- of industry standards relevant to smart manufacturing applications [8],
facturing theories to help them quickly adapt to rapid changes and there is a need to review the current standard landscape for smart
elevate product quality while optimizing the use of energy and re- manufacturing process and system automation to enable efficient pro-
sources [1]. Over the past few years, international initiatives have duction of a variety of personalized products simultaneously. Therefore,
collaboratively advocated a new generation of manufacturing – smart this paper details well-recognized standards that are essential for
manufacturing [2–4], characterized by on-demand responsive autono- manufacturing process and system integration and presents some
mous manufacturing operations via advanced sensing, data processing, emerging manufacturing automation scenarios based on these stan-
and decision-making technologies [5]. Similarly, the Industry 4.0 in- dards.
itiative [6] aims to develop efficient and low-cost production with The rest of the paper is organized as follows. Section 2 briefs the
flexible workflows for producing high-quality personalized products at drivers for smart manufacturing innovation and the impact on manu-
low costs. Industry 4.0 uses cyber-physical systems (CPS) in manu- facturing automation from two aspects, i.e., manufacturing process
facturing processes with advanced intelligence and flexibility. In par- automation and manufacturing system automation. The review of the
ticular, manufacturing automation needs personalized-product-based current landscape for smart manufacturing automation is presented in
(i.e., batch size one) manufacturing process automation and vertical two streams. Section 3 presents the current standards for enabling
integration of manufacturing systems. Both collectively form dynamic smart manufacturing process automation throughout the product

⁎
Corresponding author.
E-mail address: [email protected] (Y. Lu).

https://doi.org/10.1016/j.jmsy.2020.06.010
Received 13 May 2020; Received in revised form 18 June 2020; Accepted 19 June 2020
0278-6125/ © 2020 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

development lifecycle. Section 4 reviews existing standards for manu- Edge Computing, and Fog Computing allow resource-constrained in-
facturing system automation. Some piloting smart manufacturing au- dustrial devices and systems to gain powerful remote computation
tomation application scenarios for achieving end-to-end integration in capabilities, enabling near real-time in-process decision-making. With
smart manufacturing are presented in Section 5. Section 6 presents our the Internet of Things, dispersed manufacturing processes and systems
discussions on current smart manufacturing standardization activities can be networked and connected with ubiquitous communication and
and our suggestions on accelerating the smart manufacturing journey. coordination.
Section 7 concludes the paper. The merge of a strong desire for mass personalization and under-
pinning technology explosion has made it timely to revisit manu-
2. Smart manufacturing facturing automation theories and practices. It is believed that manu-
facturing is entering the next era – Industry 4.0 or, more specifically,
The term smart manufacturing is controversial as it has been seen as smart manufacturing.
a general concept describing manufacturing systems or processes with
advanced intelligence, such as in [9] and, more importantly, refers to a 2.2. Definition
specific manufacturing paradigm [5,10]. This paper refers to the latter.
This section briefs the vision of smart manufacturing and new re- Smart manufacturing is devised by the Smart Manufacturing
quirements for manufacturing automation, as well as highlights the role Leadership Coalition (SMLC) as a set of manufacturing practices that
of standards in facilitating the next generation of manufacturing auto- respond to a new wave of networked data and information technology
mation. capability destined to shape future manufacturing operations [2].
Wallace and Riddick [17] describe smart manufacturing as “a data-in-
2.1. Drivers tensive application of information technology at the shop floor level and
above to enable intelligent, efficient, and responsive operations.” According
A decisive driving force for manufacturing paradigm shifts is the to Davis et al. [10], smart manufacturing is the dramatically intensified
continued market desire for personalized products. The market has application of “manufacturing intelligence” throughout the manu-
never stopped its hunt for personalized products. However, this journey facturing and supply chain enterprise. NIST defines smart manu-
has been extremely long because of the limitations of manufacturing facturing systems as “fully-integrated, collaborative manufacturing systems
automation technologies, as shown in Fig. 1. Only since 2010, mass that respond in real-time to meet changing demands and conditions in the
personalization [11] (or mass individualization [12]) has become a factory, in the supply network, and in customer needs [5].”
realistic goal. Today, manufacturers desire to have the capability of These definitions all highlight the use of Information and
concurrently producing highly personalized products at dynamic batch Communication Technology (ICT) and advanced data analytics to
sizes with the efficiencies of mass production. achieve intelligent and flexible manufacturing operations at all levels
Alongside the dramatic shift from mass customization to mass per- from the shop floor, through factory level to supply chain. Integrating
sonalization, underpinning technologies for manufacturing automation all these concepts, we believe smart manufacturing is “fully-integrated,
have also advanced at an unprecedented rate, as shown in Fig. 2. After collaborative and responsive operations that respond in real-time to
the 3rd industrial revolution, the explosion of algorithms, connectivity, meet changing demands and conditions in the factory, in the supply
and computation technologies have collectively laid a solid foundation network, and in customer needs via data-driven understanding, rea-
for a connected and smart industry. In particular, the rise of Artificial soning, planning, and execution of all aspects of manufacturing pro-
Intelligence (AI) algorithms, such as Reinforcement Learning [14] and cesses, facilitated by the pervasive use of advanced sensing, modeling,
Knowledge Graph [15], can make manufacturing control more in- simulation, and analytics technologies.” The characteristics of smart
telligent and intuitive [16]. Cyber-physical Systems, Cloud Computing, manufacturing include (1) digitalization and service-orientation, (2)

Fig. 1. Manufacturing paradigm shifts and the drivers [13].

313
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

Fig. 2. Algorithm, connectivity and computation technology explosion.

smart and connected automation devices, and (3) collaborative manu- product.
facturing networks, to enable cost-effective, flexible, and resilient mass For smart manufacturing system automation, devices, machines,
personalization. systems, and people can be connected via machine-to-machine (M2M)
communication channels, creating a manufacturing network in which
2.3. Impact on manufacturing automation information carriers exchange device data in near real-time. Data-
driven distributed intelligence will enable the rapid configuration of a
As discussed in Section 2.1, two game-changing drivers in the new network of manufacturing things to cost-effectively produce a variety of
era of manufacturing are predominant – (1) the manufacturing para- personalized products with dynamic batch sizes simultaneously. In
digm shift from mass production and mass customization to mass per- particular, smart manufacturing systems should have the following
sonalization, and (2) rapid development of smart algorithms, con- features:
nectivity, and computation technologies. Traditionally, automated
manufacturing processes and systems sacrifice flexibility for pro- • Context-awareness: Smart manufacturing systems can recognize,
ductivity. In the context of mass personalization, rigid, fully automated interpret and analyze intentions of objects, systems, and partici-
systems do not work anymore. Manufacturing automation needs to be pating users in the application domain, which allows for self-
both flexible and productive to mass-produce personalized products at a awareness regarding knowledge about its situation, status, and op-
reasonable cost. Therefore, future-proofing manufacturing automation tions for actions to be taken.
will have to become smart in two aspects in parallel for achieving smart • Modularity: Smart manufacturing systems are modular, being able
manufacturing: to configure sub-components to form different system configurations
to produce new personalized products cost-effectively.

• Personalized-product-based manufacturing process automation • Self-organization: Participants (e.g., systems, processes, and
– manufacturing processes will be integrated and automated from people) can communicate with each other and coordinate their ac-
design to inspection for each unique product instead of product fa- tions in a purposeful matter without external involvement. Self-or-
milies. ganization is typically embodied as self-configuration, self-optimi-

• Networked self-organizing manufacturing systems – traditional

•
zation, and self-healing capabilities.
Data-driven decision-making: Smart manufacturing makes ex-
dedicated hierarchical manufacturing pyramid will evolve to in-
tegrated networks of autonomous manufacturing things with self- tensive use of insights learned from big engineering data to make
configuration, self-optimization, and self-healing capabilities. intelligent and adaptive decisions according to changing external
and internal conditions.
For smart manufacturing process automation, product family-based
manufacturing process configuration and integration will not apply 2.4. Smart manufacturing standard dimensions
ubiquitously as each product can be unique, and the batch size can be as
small as one in the context of mass personalization. Therefore, product Standardization is an essential requirement for integrating different
development lifecycle tracking, process configuration, and product in- systems and processes. Standardization is the top challenge in im-
spection all need to be bound to a unique product instance, e.g., a plementing smart manufacturing on a large scale [7]. Given the colla-
unique product ID. This will require deep integration and seamless bi- borative and integrated nature of smart manufacturing, the standardi-
directional data flow between all the design and manufacturing stages zation of architectures, data exchange formats, semantics, and
of the product development lifecycle without data loss [18]. A product interfaces is critical to maximizing business outputs between different
digital twin [19] needs to be created to track all the relevant in- technologies and solutions in smart manufacturing [20]. The focus of
formation (e.g., design specification, materials, manufacturing pro- international standardization initiatives should, therefore, be placed on
cesses, manufacturing facilities, and inspection logs) about a unique interoperable interfaces between smart manufacturing processes and

314
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

Fig. 3. Smart manufacturing automation dimensions (MoM: Manufacturing Operations Management; SCADA: Supervisory Control And Data Acquisition).

systems. manufacturing [26]. The Geometric Dimensioning & Tolerancing (GD&

The standardization activities in the field of smart manufacturing T) data through AP 242 can be automatically consumed by downstream
are incredibly dynamic, with several international initiatives working applications such as Computer-aided Process Planning (CAPP), Com-
on smart manufacturing-related standards. According to DIN – German puter-aided Inspection (CAI), Computer-aided Tolerance Systems
Institute for Standardization [21], there are more than 600 standards (CATS), and Coordinate Measuring Machines (CMM) [27].
related to smart manufacturing. All these dynamics make it impossible
to systematically identify the ultimate list of recommended standards
for smart manufacturing. Several studies thus far have investigated the 3.2. Standards for manufacturing
landscape of standards for smart manufacturing with distinct focuses
and recommendations [5,22,23]. Standards for planning and execution of manufacturing are also
Our review primarily focuses on the standards for enabling smart integral to smart manufacturing. In this sub-section, we view manu-
manufacturing processes and system automation. Reference archi- facturing in the narrow sense as the step to convert raw material to the
tectures of smart manufacturing itself, such as RAMI 4.0, are not in the final product based on product specification. In the era of smart man-
scope. IT-oriented standards on communication and security are not in ufacturing, manufacturing focuses on one-of-a-kind production that can
the scope, either. Fig. 3 illustrates two dimensions of concerns that are potentially involve a collection of fabrication methods, such as NC
manifest in manufacturing automation – manufacturing process auto- machining, robotic machining, and additive manufacturing. Thus, in-
mation and manufacturing system automation. Our review in Sections 3 teroperability between manufacturing systems is required to achieve a
and 4 are organized in these two streams accordingly. flexible organization of manufacturing activities under changing con-
ditions [28]. More importantly, manufacturing equipment needs to be
capable of interpreting manufacturing requirements from a CAD file at
3. Standards for manufacturing process automation
the semantic level and generating adaptive manufacturing strategies.
Take CNC machine tools as an example. The ultimate machine tool can
Smart manufacturing requires distributed manufacturing businesses
directly engage with CAD files and generate feasible manufacturing
to work together to fulfill a highly personalized product development
plans without going through the off-line process planning stage [18].
need. To achieve this, manufacturing companies need to be capable of
In this regard, ISO 14649 [29] and ISO 10303-238 [30] (also known
exchanging product data throughout the product development process
as STEP-NC), aims to replace the RS274D (ISO 6983) G and M code via
without interoperability issues.
a modern associative language that directly connects the CAD design
Fig. 4 shows a timeline-based depiction of standards for different
data with the downstream processes. Unlike G-code, STEP-NC describes
product development stages, including design, planning, manu-
“tasks” to be carried out (what-to-do information) instead of “methods”
facturing, and inspection. The plethora of standards in this space makes
to do the job (how-to-do information) for a machine tool. STEP-NC
it challenging to coalesce a shared vision across an organization. In this
relies on machine tools to interpret machine-dependent machining in-
work, our discussion focuses on a collection of standards with minimal
structions based on the local machining conditions. This shift of inter-
interoperability risks to enable streamlined product data exchange
preting local machining instructions into individual CNC controllers
during the design and manufacturing stages.
maximizes the interoperability between machine tools.
To this end, numerous studies have been conducted to achieve
3.1. Standards for product data exchange STEP-NC compliant manufacturing with applications in milling
[31–33], turning [34–36], electro-erosion [37], industrial robotics
ISO 10303, commonly known as STEP, is an international standard [38–40], and additive manufacturing [41–43]. These works have used
designed to exchange product data between CAD systems with a neutral STEP-NC in the loop of CAD/CAM/CNC, but the industry still waits for
file format and data structure. STEP has just completed a significant a truly intelligent and adaptive STEP-NC controller. The ultimate
development of STEP AP242 (as shown in Fig. 5) for “Managed Model- adaptive STEP-NC controller can optimize machining in real-time by
Based 3D Engineering”, focusing on representing 3D model data, geo- considering in-process machine conditions. An adaptive CNC controller
metric tolerance, and PMI (Product Manufacturing Information) to can fully understand a product’s design intent and its quality require-
enable global design and manufacturing collaboration [25]. STEP ment and devise an optimal manufacturing strategy according to the
AP242 can enable streamlined product design, process planning, and real-time local manufacturing environment. It can also adjust the

315
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

Fig. 4. Timeline-based depiction of standards for different product lifecycle stages ([19], enriched based on [24]).

fabrication process and parameters according to online inspection re- “workplans”. The resultant STEP-NC file is then passed on to an in-
sults. telligent controller that can interpret the complex data structure and
Smart manufacturing needs an adaptive CNC controller that can derive the necessary local machine movement commands. The in-
directly take a STEP-NC file and can communicate the as-executed formation flow is bidirectional since complete product information is
product model back to CAD/CAM system. Fig. 6 depicts the STEP-NC carried onto the CNC controller itself. More importantly, the bidirec-
compliant CAD/CAM/CNC chain, enabling interoperable design and tional information flow serves as a basis for digital twin manufacturing
manufacturing in a distributed design and manufacturing scenario. The by enabling product changes and final product parameters to be synced
chain starts with a CAD system generating a 3D CAD model with pro- back to the product design file.
duct manufacturing information in STEP AP242 standard. The 3D CAD
model is then passed on to a CAM system, which adds manufacturing
information to the geometry and stores the extended information in an 3.3. Standards for manufacturing process monitoring
AP238 file. The AP238 file contains the AP242 design information with
the addition of manufacturing operations, manufacturing features, Monitoring the manufacturing execution progress is vital to en-
tools, and machining strategies organized in “workingsteps” and suring traceability and product quality. Manufacturing process mon-
itoring has been around using equipment-dependent monitoring

Fig. 5. The high-level scope of ISO 10303 AP242.

316
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

Fig. 6. Information flow in a STEP-NC compliant manufacturing chain.

Fig. 7. Integrated metrology process with QIF standard.

technologies from individual manufacturing equipment suppliers. This machining process parameters, it could have lost the feature ID after the
method does not provide the required interoperable information flow machining planning and NC programming stage.
between manufacturing devices and processes. This limitation is not a
significant issue in a factory that mass produces standard products.
However, the rapid production of highly personalized products needs 3.4. Standards for smart inspection
the manufacturing process monitoring signals to match with product
design and manufacturing plans to derive meaningful monitoring in- Product inspection is an integral part of the product development
formation. This requirement inevitably needs standardized manu- process. It verifies product quality on-site at various stages of the pro-
facturing process monitoring and data fusion with product design files duction process from receiving the raw materials to pre-shipping.
and manufacturing plans. Currently, inspection activities are mostly performed either as off-line
The MTConnect [44] standard enables manufacturing equipment to inspection or online post-process inspection, due to the difficulties and
capture execution monitoring data and transmitted to external sources cost in setting up online in-process inspection along with the production
in a structured XML format. Since its inception, MTConnect has been process. In the future, online in-process inspection that integrates with
widely used to monitor machine tools status [45], with limited efforts the product fabrication process is a must for producing highly perso-
on contextualizing the collected machine data. In a smart manu- nalized products at small batches. The production of highly persona-
facturing context, manufacturing progress monitoring data will need to lized products cannot afford the costs incurring from manual or semi-
be associated with the unique work-in-progress part. In this regard, automated inspection.
Bernstein et al. [24] overlaid real feed rate gathered from MTConnect The Quality Information Framework (QIF) is an ANSI (the American
streams onto planned feed rate to examine machining process accuracy National Standards Institute) standard that defines an integrated set of
and used the aggregation of real feed rates and cutting parameters to XML information models to enable the exchange of metrology data
estimate machining costs of a machining feature. Helu et al. [46] also throughout the entire manufacturing quality measurement process –
demonstrated that manually aligning MTConnect data onto a virtual from product design, through inspection planning and execution, to
product model can directly reveal design insight to a decision-maker. analysis and reporting [48]. QIF provides a complete and accurate 3D
By automatically alignment of manufacturing execution data, such as product definition with semantic geometric and dimensional tolerances,
spindle speed and feed rate to numerical control code, Monnier et al. definitions for measurement resources, template for measurement rules,
[47] demonstrated the benefits of verifying process plans and real ex- and statistical functionality.
ecution data. However, this alignment still requires substantial human As shown in Fig. 7, a QIF-compliant metrology process starts with
interpretation and intervention. This is mostly because the features and the generation of CAD and PMI data exported as QIF Model-Based
the identification of design and machining features are not maintained Design (MBD) product model. Quality planning systems import the
consistently throughout the data conversion chain from design to ma- product model and generate the measurement plans according to the
chining. For example, though MTConnect can output near real-time quality requirements and manufacturing processes. The inspection re-
sources and rules are considered at this stage. Programming systems

317
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

import the measurement plans to create Dimensional Measurement exchange and generic, extensible, and object-oriented modeling cap-
Equipment (DME) specific programs. Dimensional measurement abilities for the information a production system wants to expose.
equipment executes the programs and evaluates the measurements as Similar to OPC UA, MTConnect is used to access real-time data from
measurement results. Analysis systems import single part results and shop floor manufacturing equipment such as machine tools. ISA 88 is a
generate the analysis of multiple part batches as QIF statistics data. standard for the batch processing industry and defines the reference
The integration between quality information using QIF and pro- models and data models for batch control. BatchML is an XML im-
duction process data saves resources on non-value-adding activities plementation of ISA 88. PackML models standardized machines and
related to the translation of data between different components of a SCADA for batch control in the packaging industry.
dimensional metrology system. This integration will enable integrated
manufacturing and inspection taking place simultaneously, which in
the end fulfills the rapid production and QA (Quality Assurance) of 4.2. Communication standards
highly personalized products with minimum human involvement.
The research of using QIF standards across the product lifecycle is Industrial communication goes beyond physical data packet trans-
still in its infancy with just some theoretical discussions on harnessing mission between manufacturing things. According to [52], commu-
QIF data into the product design and manufacturing processes [24,49]. nication technologies should be able to:
Michaloski et al. [50] investigated the possibility of gathering con-
tinuous quality inspection results during the product machining process • Send and receive messages – at this physical level, manufacturing
via integrating QIF standard with MTConnect. Their pilot provided a devices must communicate over agreed physical and network layers
feasible solution for collecting and representing necessary process data to be able to send and receive objects that represent messages;
and quality measurement data at the same time. With the same intent, • Parse the messages – at the syntactic level, manufacturing devices
STEP Tools, Inc [51]. recently developed a digital thread solution that must parse messages to correctly decode the message to its parts,
keeps the design, manufacturing, and inspection data connected via such as message content, language, sender, and must be able to
deep integration between STEP, STEP-NC and QIF standard. parse the content of the message;
More research needs to be carried out to understand the benefits • Understand the messages – at the semantic level, manufacturing
and limitations of automating the product manufacturing and inspec- devices must interpret and reason about the parsed information in
tion process by utilizing the recently developed QIF standards. the same context in the same way, in which context-aware actions
can be taken.
4. Standards for manufacturing system automation
At the physical level, messages can be exchanged via a physical
Apart from the need for mass-personalization-oriented smart pro- communication channel, wired or wireless [53]. The foundation of
cess integration from design to inspection, smart manufacturing also M2M communication is the reliable exchange of information. Since the
needs vertically integrated smart factories with ubiquitous connectivity 1980s, industrial communication networks have evolved through sev-
between manufacturing devices and adaptive manufacturing opera- eral stages. According to [53], industrial communications started with
tions, thus moving away from traditional rigid fully-automated manu- dedicated Fieldbus networks, such as PROFIBUS and Modbus, to enable
facturing systems. This section reviews the essential standards that fa- M2M communications. However, many Fieldbus protocols were de-
cilitate smart manufacturing system automation for ensuring flexible signed to operate on different physical media and have wide compat-
and efficient production of a variety of personalized products simulta- ibility issues with the OSI (Open Systems Interconnection) model [19].
neously. As Internet technologies matured, Ethernet-based networks, such as
Industry 4.0 will see factory automation architecture changing from EtherNet and EtherCAT, became popular for facilitating communica-
a conventional dedicated automation pyramid to a networked auto- tions at a higher level. Since 2000, influenced by Internet of Things
mation structure, in which intelligent manufacturing things con- (IoT) and Wireless Sensor Networks (WSN) application needs, some
tinuously and collaboratively self-optimize their setups and configura- modern approaches have adopted new standards such as IEEE 802.11,
tions via self-awareness, reasoning, planning, and execution. In this IEEE 802.15.1, and IEEE 802.15.4. Detailed discussions on these
network, the information from these behaviors must flow between any transport-oriented network communication channels and their latest
connected nodes. Therefore, data exchange can be between any two developments can be found in [53].
layers in the conventional automation pyramid, not just adjacent layers. The above transport-oriented communication technologies provide
However, for presentation simplicity, we still present the relevant industrial data exchange with guaranteed reliability, availability, and
standards in a manufacturing pyramid structure. In Table 1, we divide time-critical behavior. However, meaningful data exchange in a specific
standards based on the ISA 95 hierarchy and classify their use to three domain needs more technologies, one of which is a meta-model that
primary categories – modeling, communication, and execution. explicitly defines the data content to be exchanged [54].
OPC UA developed by OPC Foundation provides a cross-platform
4.1. Modeling standards M2M communication mechanism for data exchange between industrial
devices and systems [55]. OPC UA also supports an object-oriented
From the top level of the automation pyramid, several standards information modeling approach. It allows the development of domain-
define the reference architecture, framework, core constructs, activities, specific companion models, for example, in [56–58]. In the manu-
interfaces, and transactions of a factory. IEC 62714 (AutomationML) facturing domain, MTConnect is a widely-used semantic vocabulary for
models linked production systems, enabling a transfer of engineering exchanging contextualized data of manufacturing equipment [44]. It
data of these systems across domains and companies in a heterogeneous has been widely used for monitoring machining processes [59] and
engineering tool landscape. At the MoM level, IEC 62264 defines ac- machining simulation [60].
tivities models, function models, and object models in the MoM do- At the semantic level, semantic communication technologies are
main, and B2MML serves the implementation of IEC 62264 to link ERP required for interpreting message content and building contextual un-
and Supply Chain Management systems with manufacturing systems derstandings within a domain or across domains. Semantic Web tech-
such as Manufacturing Execution Systems (MES). nologies, such as RDF, RDF Schema, and OWL can be used at a higher
At the SCADA level, OPC UA (OPC Unified Architecture) can be level of the automation pyramid to enable semantic understanding
used for connecting components in a production system. It also defines between devices and systems across domains [61–63].
platform-independent communication mechanisms for online data

318
 Table 1
Manufacturing system automation standards by automation pyramid levels and functions.
Y. Lu, et al.

Standard Description Enterprise Level MoM SCADA Device Modeling Communication Execution
Level Level Level

IEC PAS 63088 Smart manufacturing - Reference architecture model industry 4.0 (RAMI4.0) ✓ ✓
IEC 62832 Industrial-process measurement, control, and automation - Digital factory framework defines the ✓ ✓
general principles of the Digital Factory framework including a set of model elements and rules for
modeling production systems
ISO 15704 Enterprise modeling and architecture — Requirements for enterprise-referencing architectures and ✓ ✓
methodologies
ISO 19439 Enterprise Integration - Framework for enterprise modeling, provides a unified conceptual basis for ✓ ✓
model-based enterprise engineering that enables consistency, convergence, and interoperability of the
various modeling methodologies and supporting tools
ISO 19440 Enterprise integration - Constructs for enterprise modeling, specifies the characteristics of the core ✓ ✓
constructs necessary for computer-supported modeling of enterprises conforming to ISO 19439.
ISO 20140 Automation systems and integration - Evaluating energy efficiency and other factors of manufacturing ✓ ✓
systems that influence the environment.
OAGIS Open Applications Group Integration Specification (OAGIS) defines a common content model and ✓ ✓ ✓ ✓
common messages for communication between business applications, including application-to-
application (A2A) and business-to-business (B2B) integration.
BPMN Business Process Model and Notation (BPMN) is a graphical representation for specifying business ✓ ✓
processes in a business process model.
DMN Decision Modelling Notation (DMN) from OMG provides the required constructs for modeling ✓ ✓
decisions to close the gap between the business decision design and decision implementation. DMN
notation is designed to be useable alongside the standard BPMN business process notation.
B2MML Business To Manufacturing Markup Language (B2MML) is an XML implementation of data models in ✓ ✓
IEC/ISO 62264. It is a common data definition to link ERP and supply chain management systems
with manufacturing systems such as Industrial Systems and Manufacturing Execution Systems.

319
IEC 62714 AutomationML (Automation Markup Language) is a neutral data format based on XML for the storage ✓ ✓
and exchange of plant engineering information.
IEC 62264 Enterprise-control system integration - defines manufacturing hierarchical model, and describes the ✓ ✓
manufacturing operations management domain and its activities, the interface content and associated
transactions within Level 3 and between Level 3 and Level 4 and Level 3 objects. This standard is
based upon ANSI/ISA-95.
IEC 62541 OPC Unified Architecture - an industrial M2M communication protocol for interoperability developed ✓ ✓ ✓ ✓ ✓
by the OPC Foundation
IEC 61512 ISA 88 defines terminology, reference models, data models for batch control as used in the process ✓ ✓ ✓
industries.
IEC/PAS 62030 (Modbus) Modbus is a de facto standard providing serial communications protocol to connect industrial ✓ ✓
electronic devices; Modbus is often used to connect a supervisory computer with a remote terminal
unit (RTU)/PLC in supervisory control and data acquisition systems.
PackML Packaging Machine Language defines a common approach, or machine language, for automated ✓ ✓ ✓
machines. PackML was adopted as part of the ISA88 industry standard in August 2008.
BatchML BatchML is an XML implementation of the ISA-88 ✓ ✓ ✓
MTConnect MTConnect is a lightweight, open, and extensible protocol designed for the exchange of data from ✓ ✓ ✓
shop floor equipment to software applications used for monitoring and data analysis.
IEC 61158 The standard specifies industrial communication networks - Fieldbus, including ControlNet and ✓ ✓
Profibus.
IEC 61784 This standard specifies a set of protocol specific communication profiles based on the IEC 61158 series ✓ ✓
and real-time Ethernet communication profiles. It is to be used in the design of devices involved in
communications in factory manufacturing and process control.
ISO 11898 Controller Area Network (CAN) - a serial communication protocol that supports distributed real-time ✓ ✓
control and multiplexing
IEC 62591/HART The standard specifies Wireless communication network and communication profiles - WirelessHART. ✓ ✓
IEC 61499 The standard describes a generic modeling approach for distributed control applications enabling ✓ ✓ ✓
interoperability, reconfigurability, and portability for distributed control systems, facilitated through
event-driven Function Blocks
(continued on next page)
Journal of Manufacturing Systems 56 (2020) 312–325
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

Execution 4.3. Execution standards

✓
Manufacturing execution can be categorized into two levels – (1)
production control at the enterprise and MoM level and (2) device
Communication

control. Standards for manufacturing execution have been focusing on

enabling adaptive and interoperable control of industrial processes and
systems. For production control, ISA 88 provides a consistent set of
standards and terminology for batch control. It defines the physical
models, procedures, and recipes. PackML specialized in batch control
Modeling

for the packaging industry, provides (1) standard machine states and
operational flow, and (2) flexible recipe schemes and common SCADA
✓

or MoM inputs.
At the device level, G and M codes are currently widely used to
control CNC machines and 3D printers. However, G code’s machine-
Device
Level

dependent nature stops it from being shared between machines.

✓

Instead, STEP-NC, as discussed in Section 3.2, can embed a complete set

of machining information, allowing manufacturing organizations to
share machining information between machines.
SCADA
Level

IEC 61499 function block specification is an IEC standard for dis-

tributed industrial processes and control systems, particularly for PLC
control. It is based on an explicit event-driven model and provides for
data flow and finite-state automata-based control. Based on previous
Level
MoM

research on function blocks [64–66], function blocks can be used for

machine-level monitoring, shop floor execution control, and CNC con-
Enterprise Level

trol. Research has demonstrated the effectiveness of delivering generic

manufacturing process plans in the form of function blocks [67] to
machine controllers on the physical shop floor for execution. Thanks to
the event-driven model and ability to embed multiple control logics
into a function block, manufacturing process plans in a function block
will not be tied to a specific machine but is portable to any alternative
machines if needed. Similar to STEP-NC, when CNC controllers natively
support function blocks, there will be no need to generate G-code but to
run algorithms directly for adaptive machining.

5. Smart manufacturing automation scenarios

Physical device control – Data model for computerized numerical controllers

This section presents several future-proofing manufacturing process

and system automation scenarios enabled by the various standard de-
velopment efforts and application innovations. The scenarios focus on
(1) personalized-product-based manufacturing process automation, and
(2) networked self-organizing manufacturing system automation.

5.1. Manufacturing digital thread

Cutting tool data representation and exchange

Manufacturing process automation in the context of mass persona-

lization requires a manufacturing digital thread that facilitates up-
stream and downstream flow between the manufacturing processes by
seamlessly fusing data from design, planning, manufacturing, and in-
spection via standard integration interfaces. Such a digital thread will
enable informed design and analysis, collaborative manufacturing
process planning and control, full-process traceability, and automated
online quality inspection [68].
As shown in Fig. 8, manufacturing digital thread takes a model-
Description

based approach to describing both each step as well as the connectivity

and interoperability of manufacturing processes, both digital and phy-
sical. Product design starts with an as-designed model in STEP AP242
format. Based on the principles in Section 3.2, the as-designed model is
transferred to an as-planned model with all the planning rules and
ISO 10303-238 / ISO 14649

coding procedures (e.g., STEP-NC and G-code) that are necessary for
executing the production process. All the data that describe the actual
Table 1 (continued)

build event in the manufacturing environment are tracked in an as-

(STEP-NC)

executed model. At the manufacturing phase, in-situ monitoring and

product inspection constantly update the product model, thus building
ISO 13399
Standard

the as-inspected model. Post-manufacturing, product assembly, and

final inspection occur. The final QA report using QIF standard is in-
corporated into the final product digital twin before commissioning.

320
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

Fig. 8. Manufacturing digital thread with seamless error-free and zero data-loss information flow upstream and downstream the manufacturing processes.

The above manufacturing digital thread outlines a feasible approach network can perceive it is not operating correctly and, without ex-
to manufacturing process automation for a unique product. With this ternal assistance, make the necessary adjustments to restore itself to
manufacturing thread, the overall manufacturing process of a perso- normal operation. A self-healing manufacturing network can
nalized product can be automated via standard-compliant data ex- proactively monitor and identify a potential variance from its
change interfaces between different manufacturing stages. These data standard parameters, validate it with a degree of confidence and
exchange interfaces can seamlessly link together disparate processes, resume normal operations without human intervention.
systems, and companies into a single coherent ecosystem. Of all the
enablers, the discussed smart manufacturing standards are the foun- Fig. 9 depicts an abstract framework of a self-organizing manu-
dation for a streamlined manufacturing digital thread. However, ex- facturing network. A self-organizing manufacturing network consists of
isting integration technologies cannot establish a true manufacturing three components:
thread yet without repetitive manual assistance between various man-
ufacturing stages, such as manual process planning and data conversion • Semantically interoperable communications between manufacturing
from STEP AP242 to AP238. More algorithms and technologies need to things
be developed to contextualize and link the data from multiple stan- • Accurate observation of manufacturing environment via self-sensing
dards, as shown in [46,47]. • Cognitive manufacturing control algorithms to enable self-config-
uration, self-optimization, and self-healing
5.2. Self-organizing manufacturing network
For establishing a manufacturing network, M2M communication
Smart manufacturing will see conventional manufacturing pyramid protocols will be used to exchange information between the connected
changing to a self-organizing manufacturing network with cognitive nodes of a network. Apart from the physical level communications,
decision-making capabilities, enabling resilient and responsive manu- Semantic Web technologies [63] and resource virtualization models,
facturing operations. Self-organizing manufacturing network can be such as Industry 4.0 Administration Shell [71], can also be used to
defined as “a network of autonomous manufacturing things (e.g., create semantic understandings between heterogenous connected
manufacturing software tools, manufacturing equipment, and manufacturing “things” across domains and organizations.
operators) connecting in situation-dependent ways that can For modeling and observations of the environment, the Industrial
change their internal structure, organization, and functions with IoT can collect raw environmental data from various sensors and de-
minimum external intervention to achieve optimal manufacturing vice-level monitoring systems. The multi-source, multi-scale, hetero-
operations and system performance in response to unforeseen genous, and multi-granularity sensor data can be fused to derive a
conditions and evolution along time.” In general, self-organization trustful observation of the environment. This observation can be
aims at increasing autonomy and decreasing dependencies for faster quantitatively represented using potential system modeling metrics,
responses to dynamic context [69]. A self-organizing manufacturing such as order status, material stock, and machine workload, thus
network includes the following three aspects of autonomous functions: forming an observed state of the environment feeding into the
Cognitive Agent.
• Self-configuration – Self-configuration strives towards the “plug- Cognitive algorithms can enable self-organizing manufacturing
and-produce” paradigm in the way that all manufacturing “things” control to adaptively schedule optimal manufacturing routes for a
shall be capable of integrating to form a feasible production system variety of personalized products simultaneously. Traditional heuristic
via standard hardware and software interfaces. A manufacturing algorithms, such as genetic algorithms and ant colony optimization
network can be dynamically re-established via adding or deleting a algorithms, can be used. More advanced cognitive capabilities can be
connecting node or reconfiguring the links between existing nodes. developed for all the connected manufacturing things via learning-
• Self-optimization – Every node in a manufacturing network can based algorithms, such as Reinforcement Learning. In a Reinforcement
automatically optimize its parameters to seek optimal performance Learning setting, the Cognitive Agent seeks to maximize the cumulative
of the manufacturing network, based on observations of its state as reward to optimize its action-selection strategy. The reward value can
well as the state of the connected node in the network. be linked to some system objective factors together, such as less order
• Self-healing – Self-healing refers to the ability that a manufacturing completion time, lower cost, and higher equipment utilization rate.

321
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

Fig. 9. Abstract framework of self-organizing manufacturing network in Reinforcement Learning system model (Adapted from [70]).

A fully-fledged self-organizing manufacturing network will become Company B is created for the creation of a virtual MES. The machine
a ubiquitous manufacturing system organization pattern that can be from Company B is connected dynamically to Company A’s MES.
implemented for a manufacturing system, a factory, or an inter-en- Company A can control (1) in what sequence A’s jobs are executed, and
terprise production network. The continuous learning and optimization (ii) how manufacturing should be executed on Machine B by injecting
control loop of a self-organizing manufacturing network will be able to Company A’s know-how into Machine B. Company A also gets shopfloor
adequately deal with uncertainties from (1) dynamic concurrent orders production progress report from Machine B regularly.
of highly personalized products, (2) manufacturing network dynamics, In this scenario, manufacturing assets become services thanks to
and (3) emergent instructions from human supervisors. intelligent technologies, such as service-oriented architecture and ma-
Existing work on self-organization principles for manufacturing chine communication standards. More importantly, vendor-neutral data
system automation, such as holonic manufacturing systems [72–75], format – STEP/STEP-NC and Function Block play a crucial role in en-
multi-agent manufacturing systems [76–79], and bionic manufacturing abling passing manufacturing jobs between companies and machines.
systems [80,81], has provided some answers to autonomous manu- Because of the ability to describe generic manufacturing jobs using
facturing control in a distributed environment, in particular in semi- STEP-NC standard, a manufacturer can easily interpret received man-
structured and partial-hierarchical settings [82]. These approaches face ufacturing requirements and generate adaptive machine control stra-
the problem of dynamic environment changes but introducing a closed tegies based on local machine conditions and setups.
world assumption [83]. They lack universal adaptability to an extend-
able, collaborative, and sometimes chaotic manufacturing network.
While self-organizing manufacturing network could still adopt a partial- 6. Discussions
hierarchical control structure [84] (a mixture of heterarchical and
hierarchical control structure) to achieve both system flexibility and We detailed in Sections 3–5 selected standards for enabling manu-
optimality, a key research goal can be advancing fully decentralized facturing process and system automation, as well as some pioneering
learning-based algorithms, such as Multi-agent Reinforcement Learning manufacturing automation scenarios. This section discusses some of the
algorithms [85–87], to achieve satisfactory system performance in notable new standardization initiatives targeting smart manufacturing
chaotic manufacturing networks. automation and our views on future research activities related to smart
manufacturing standards development and application implementa-
tion.
5.3. Cloud-based manufacturing equipment as service Though many standards are already in place, there is ongoing work
from various standardization bodies reviewing, enriching, and devel-
As the Industrial Internet and interoperable machine control inter- oping standards for Industry 4.0 manufacturing automation. For ex-
face matures, manufacturing equipment can become an on-demand ample, ISO Smart Manufacturing Coordinating Committee (ISO/SMCC)
manufacturing service via connectivity and control over the Internet is coordinating international standards review and implementation re-
[88]. This will be a practical end-to-end implementation of cloud commendations in this space. IEC Systems Committee smart manu-
manufacturing principles [89]. Cloud-based equipment as a service can facturing (IEC/SyC) started in 2018, aims to identify gaps and overlaps
significantly contribute to the required manufacturing process auto- relating to the collaboration between relevant standards organizations.
mation and manufacturing system automation. This is because the ISO and IEC also set up a joint working group to draw up, publish, and
possible service-oriented architecture of cloud-based manufacturing update Industry 4.0 related standards. In particular, being the most
equipment can be directly embedded into a manufacturing digital active technical committee in automation, ISO Technical Committee
thread or a self-organizing manufacturing network. 184 (ISO/TC 184) and IEC Technical Committee 65 (IEC/TC 65) formed
Fig. 10 depicts a simplified concept of encapsulating a physical asset a joint working group – ISO/IEC JWG 21, to bring about the harmo-
as a consumable manufacturing service. Manufacturing assets are con- nization of smart manufacturing reference models and various stan-
nected to the Internet in the form of cyber-physical production systems dards.
via machine communication standards, such as MTConnect or OPC-UA. Beyond the possible standard gaps for smart manufacturing, three
At a scheduled time, a control network between Company A and other barriers to standard adoption hinder the growth of smart

322
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

Fig. 10. Cloud-based manufacturing equip-

ment as service via standardized machine
control and monitoring interfaces (The green-
colored background indicates A’s temporary
scope of ownership of equipment) (For inter-
pretation of the references to colour in this
figure legend, the reader is referred to the web
version of this article).

manufacturing: (1) lack of tracking of standard adoption, (2) overlap implementation projects should be encouraged.
and redundancies between standards, and (3) inadequate co-develop-
ment of standards with user communities. Regarding the lack of 7. Conclusions
tracking of standard adoption, the significant number of standards can
be misleading and not easy to use. Standardization parties often view This paper presents a critical review of standards for enabling the
the success of a standardization program as publishing the standards, next generation manufacturing process and system automation in the
leaving potential adopters blindly navigating the standards. Huge pro- context of smart manufacturing in two parts. First, we traced the au-
blem also exists in tracking the adoption rate and effectiveness of a thentic concept and vision of smart manufacturing and the impact on
standard in the industry. next-generation manufacturing automation. Second, a comprehensive
Regarding the overlap and redundancies between standards, it is review of existing standards for enabling manufacturing process and
unfortunate to see many standardization bodies are competing revising system automation is presented, highlighting a sensible approach to
and writing standards to claim their authority in smart manufacturing developing standard-compliant manufacturing automation solutions.
compatibility. However, the full picture of Industry 4.0/smart manu- Several smart manufacturing automation scenarios are also presented.
facturing is yet to be clear. These practices cause overlap and re- Smart manufacturing promises a future of mass-producing highly
dundancy. As an example, the draft standard ISO 23247 – Automation personalized products via responsive autonomous manufacturing op-
systems and integration – Digital Twin framework for manufacturing, erations at competitive costs. In our humble view, smart manu-
primarily focuses on software architecture reference model by in- facturing, though, as a new concept has become a buzzword with in-
tegrating existing modeling and communication standards. Our review consistent interpretations, commendable research advancement on the
found that the existing standards are good enough for implementing the core research issues of smart manufacturing is still limited. Importantly,
required smart manufacturing process and system automation towards smart manufacturing automation needs to focus on two fundamental
flexible and reconfigurable mass personalization. themes – personalized-product-based manufacturing process automa-
Regarding the last barrier, top-down standardization efforts need to tion and networked self-organizing manufacturing system automation.
embrace bottom-up need-driven community-based research activities. We believe existing standards have provided a solid foundation for
Since the inception of Industry 4.0, academic research activities and developing smart manufacturing automation solutions in these two
industrial innovations have shown remarkable successes in developing themes. Together with advanced algorithms, connectivity, and com-
de facto standards and applications for smart manufacturing, such as putation technologies, much work can be done to achieve networked
MTConnect, openAAS (open Asset Administration Shell) [90], NIST self-organizing manufacturing with advanced intelligence in re-
Smart Manufacturing Systems (SMS) Test Bed [91] and Eclipse IoT sponding to dynamic changes in the manufacturing environment.
[92]. These open-source projects have made a significant contribution
to the up-take of smart manufacturing reference implementations. They Declaration of Competing Interest
will provide first-hand learnings to any necessary upcoming standar-
dization activities.
The authors declared no potential conflicts of interest with respect
To summarize, the two overarching transformative manufacturing
to the research, authorship and/or publication of this article.
signals from Industry 4.0 – (1) mass personalization and (2) CPS-based
automation have not changed the application grounds for existing
manufacturing automation standards to play. Research focus, on the Acknowledgments
one hand, should be seeking improvements to existing standards to
ensure their applicability to manufacturing automation in the context of The authors are grateful to Mr. Zhaojun Qin, Department of
smart manufacturing and, on the other hand, needs to diversify the Mechanical Engineering, The University of Auckland, for his valuable
dissemination mechanisms of standard-based manufacturing automa- inputs. The authors also thank the reviewers for their valuable com-
tion exemplars in practice. In particular, need-driven open-source ments that helped to significantly improve this article. The article is an
enhanced and extended version of the following conference article “Lu

323
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

Y, Huang H, Liu C, Xu X. Standards for smart manufacturing: A review. 09511920600627170.

IEEE International Conference on Automation Science and Engineering, [32] Suh SH, Lee BE, Chung DH, Cheon SU. Architecture and implementation of a
shopfloor programming system for STEP-compliant CNC. Comput Des
vol. 2019- August, IEEE Computer Society; 2019, p. 73–78.” 2003;35:1069–83. https://doi.org/10.1016/S0010-4485(02)00179-3.
[33] Nassehi A, Newman ST, Allen RD. The application of multi-agent systems for STEP-
References NC computer aided process planning of prismatic components. Int J Mach Tools
Manuf 2006;46:559–74. https://doi.org/10.1016/j.ijmachtools.2005.06.005.
[34] Xu XW. Realization of STEP-NC enabled machining. Robot Comput Integr Manuf
[1] Wang L, Shih AJ. Challenges in smart manufacturing. J Manuf Syst 2016;40:1. 2006;22:144–53. https://doi.org/10.1016/J.RCIM.2005.02.009.
https://doi.org/10.1016/j.jmsy.2016.05.005. [35] Shin S-J, Suh S-H, Stroud I. Reincarnation of G-code based part programs into STEP-
[2] Coalition SML. Implementing 21st century smart manufacturing. Workshop sum- NC for turning applications. Comput Des 2007;39:1–16. https://doi.org/10.1016/J.
mary report; 2011. CAD.2006.08.005.
[3] Kagermann H, Helbig J, Hellinger A, Wahlster W. Recommendations for im- [36] Suh S-H, Chung D-H, Lee B-E, Shin S, Choi I, Kim K-M. STEP-compliant CNC system
plementing the strategic initiative INDUSTRIE 4.0: securing the future of German for turning: data model, architecture, and implementation. Comput Des
manufacturing industry; final report of the industrie 4.0 working group. 2006;38:677–88. https://doi.org/10.1016/J.CAD.2006.02.006.
Forschungsunion; 2013. [37] ISO 14649-13:2013 - Automation systems and integration – physical device control
[4] Li L. China’s manufacturing locus in 2025: With a comparison of “Made-in-China – data model for computerized numerical controllers – part 13: process data for wire
2025” and “Industry 4.0. Technol Forecast Soc Change 2017. electrical discharge machining (wire-EDM). 2013.
[5] Lu Y, Morris KC, Frechette S. Current standards landscape for smart manufacturing [38] Zivanovic S, Slavkovic N, Milutinovic D. An approach for applying STEP-NC in
systems. National Institute of Standards and Technology, NISTIR; 2016. p. 22–8. robot machining. Robot Comput Integr Manuf 2018;49:361–73. https://doi.org/10.
8107. 1016/J.RCIM.2017.08.009.
[6] MacDougall W. Industrie 4.0: smart manufacturing for the future. Germany Trade & [39] Solvang B, Refsahl LK, Sziebig G. STEP-NC based industrial robot CAM system. IFAC
Invest; 2014. Proceedings Volumes 2009:245–50. https://doi.org/10.3182/20090909-4-JP-
[7] Bitkom V. ZVEI (2013):Tendenzumfrage der Plattform Industrie 4.0. Abgerufen Am; 2010.00043. 42.
2013. 9. [40] Toquica JS, živanović S, Alvares AJ, Bonnard R. A STEP-NC compliant robotic
[8] DIN Deutsches Institut für Normung e. V. Industry 4.0 standard overview; 2019 machining platform for advanced manufacturing. Int J Adv Manuf Technol
(accessed April 15, 2020). https://www.din.de/en/innovation-and-research/ 2018;95:3839–54. https://doi.org/10.1007/s00170-017-1466-8.
industry-4-0. [41] Um J, Rauch M, Hascoët J-Y, Stroud I. STEP-NC compliant process planning of
[9] Kusiak A. Smart manufacturing. Int J Prod Res 2018;56:508–17. https://doi.org/10. additive manufacturing: remanufacturing. Int J Adv Manuf Technol
1080/00207543.2017.1351644. 2017;88:1215–30. https://doi.org/10.1007/s00170-016-8791-1.
[10] Davis J, Edgar T, Porter J, Bernaden J, Sarli M. Smart manufacturing, manu- [42] Bonnard R, Mognol P, Hascoët J-Y. A new digital chain for additive manufacturing
facturing intelligence and demand-dynamic performance. Comput Chem Eng processes. Virtual Phys Prototyp 2010;5:75–88. https://doi.org/10.1080/
2012;47:145–56. https://doi.org/10.1016/J.COMPCHEMENG.2012.06.037. 17452751003696916.
[11] Hu SJ. Evolving paradigms of manufacturing: from mass production to mass cus- [43] AP 242 Edition 2 capabilities for Additive Manufacturing interoperability n.d.
tomization and personalization. Procedia CIRP 2013;7:3–8. https://doi.org/10. http://www.ap242.org/additive-manufacturing (accessed July 24, 2018).
1016/j.procir.2013.05.002. [44] Sobel W. MTConnect standard. Part 1—overview and protocol. 2015. 10.
[12] Gu X, Koren Y. Manufacturing system architecture for cost-effective mass-in- [45] Liu C, Vengayil H, Lu Y, Xu X. A cyber-physical machine tools platform using OPC
dividualization. Manuf Lett 2018;16:44–8. https://doi.org/10.1016/j.mfglet.2018. UA and MTConnect. J Manuf Syst 2019;51:61–74. https://doi.org/10.1016/j.jmsy.
04.002. 2019.04.006.
[13] Koren Y. The global manufacturing revolution: product-process-business integration [46] Helu M, Joseph A, Hedberg T. A standards-based approach for linking as-planned to
and reconfigurable systems. Wiley; 2013. as-fabricated product data. CIRP Ann. 2018;67:487–90. https://doi.org/10.1016/J.
[14] Mnih V, Kavukcuoglu K, Silver D, Rusu AA, Veness J, Bellemare MG, et al. Human- CIRP.2018.04.039.
level control through deep reinforcement learning. Nature 2015;518:529–33. [47] Monnier L, Bemstein WZ, Foufou S. A proposed mapping method for aligning ma-
https://doi.org/10.1038/nature14236. chine execution data to numerical control code. IEEE International Conference on
[15] Chen X, Jia S, Xiang Y. A review: knowledge reasoning over knowledge graph. Automation Science and Engineering, Vol. 2019- August 2019. p. 66–72. https://
Expert Syst Appl 2020;141:112948https://doi.org/10.1016/j.eswa.2019.112948. doi.org/10.1109/COASE.2019.8842832.
[16] Wang L. From intelligence science to intelligent manufacturing. Engineering [48] DMSC. Home - QIF Standard n.d. http://qifstandards.org/ (accessed July 7, 2018).
2019;5:615–8. https://doi.org/10.1016/j.eng.2019.04.011. [49] Morse E, Heysiattalab S, Barnard-Feeney A, Hedberg T. Interoperability: Linking
[17] Wallace E, Riddick F. Panel on enabling smart manufacturing. 2013. Design and Tolerancing with Metrology. Procedia CIRP 2016;43:13–6. https://doi.
[18] Xu X. Machine Tool 4.0 for the new era of manufacturing. Int J Adv Manuf Technol org/10.1016/J.PROCIR.2016.04.106.
2017;92:1893–900. https://doi.org/10.1007/s00170-017-0300-7. [50] Michaloski JL, Zhao YF, Lee BE, Rippey WG. Web-enabled, real-time, quality as-
[19] Lu Y, Liu C, KI-K Wang, Huang H, Xu X. Digital Twin-driven smart manufacturing: surance for machining production systems. Procedia CIRP 2013;10:332–9. https://
connotation, reference model, applications and research issues. Robot Comput doi.org/10.1016/j.procir.2013.08.051.
Integr Manuf 2020;61:101837https://doi.org/10.1016/j.rcim.2019.101837. [51] STEP tools I. Digital thread for manufacturing. 2017 (accessed July 7, 2018).
[20] Kagermann Henning, Anderl Reiner, Gausemeier Jürgen, Schuh Günther, Wahlster https://www.steptoolsI.com/sln/thread/.
W. Industrie 4.0 in a global context: strategies for cooperating with international [52] Lu Y, Asghar MR. Semantic communications between distributed cyber-physical
partners (acatech STUDY) Munich 2016. systems towards collaborative automation for smart manufacturing. J Manuf Syst
[21] Industry 4.0 Standards. DIN - German Institute for Standardization n.d. https:// 2020;55:348–59. https://doi.org/10.1016/j.jmsy.2020.05.001.
www.din.de/en/innovation-and-research/industry-4-0/standards (accessed June [53] Wollschlaeger M, Sauter T, Jasperneite J. The Future of Industrial Communication:
28, 2018). Automation Networks in the Era of the Internet of Things and Industry 4.0. IEEE Ind
[22] Adolph L. German standardization roadmap: industry 4.0. 2016. Electron Mag 2017;11:17–27. https://doi.org/10.1109/MIE.2017.2649104.
[23] National smart manufacturing standards architecture construction guidance. 2015. [54] Rowley J. The wisdom hierarchy: representations of the DIKW hierarchy. J Inf Sci
[24] Bernstein WZ, Hedberg TD, Helu M, Barnard Feeney A. Contextualising manu- 2007;33:163–80. https://doi.org/10.1177/0165551506070706.
facturing data for lifecycle decision-making. Int J Prod Lifecycle Manag [55] OPC Foundation. Unified Architecture - OPC Foundation n.d. https://
2017;10:326–47. https://doi.org/10.1504/IJPLM.2017.090328. opcfoundation.org/about/opc-technologies/opc-ua/ (accessed April 25, 2019).
[25] ISO. ISO 10303-242: industrial automation systems and integration - Product data [56] Miyazawa I, Murakami M, Matsukuma T, Fukushima K, Maruyama Y, Matsumoto
representation and exchange - Part 242: application protocol: managed model- M, et al. OPC UA information model, data exchange, safety and security for IEC
based 3D engineering. 2014. 61131–3. SICE Annual Conference 2011. 2011. p. 1556–9.
[26] Wardhani R, Xu X. Model-based manufacturing based on STEP AP242. 2016 12th [57] Trnka P, Kodet P, Havlena V. OPC-UA information model for large-scale process
IEEE/ASME International Conference on Mechatronic and Embedded Systems and control applications. IECON 2012-38th Annual Conference on IEEE Industrial
Applications (MESA) 2016. p. 1–5. https://doi.org/10.1109/MESA.2016.7587187. Electronics Society. 2012. p. 5793–8.
[27] Venkiteswaran A, Hejazi SM, Biswas D, Shah JJ, Davidson JK. Semantic inter- [58] Maka A, Cupek R, Rosner J. OPC UA object oriented model for public transportation
operability of GD&T data through ISO 10303 step AP242. Volume 2B: 42nd Design system. 2011 UKSim 5th European Symposium on Computer Modeling and
Automation Conference 2016. https://doi.org/10.1115/DETC2016-60133. p. Simulation 2011. p. 311–6. https://doi.org/10.1109/EMS.2011.84.
V02BT03A018. [59] Edrington B, Zhao B, Hansel A, Mori M, Fujishima M. Machine monitoring system
[28] Mabkhot M, Al-Ahmari A, Salah B, Alkhalefah H. Requirements of the smart factory based on MTConnect technology. Procedia CIRP 2014;22:92–7. https://doi.org/10.
system: a survey and perspective. Machines 2018;6:23. https://doi.org/10.3390/ 1016/J.PROCIR.2014.07.148.
machines6020023. [60] Shin S-J, Woo J, Kim DB, Kumaraguru S, Rachuri S. Developing a virtual machining
[29] ISO. ISO 14649-1: industrial automation systems and integration - Physical device model to generate MTConnect machine-monitoring data from STEP-NC. Int J Prod
control - Data model for computerized numerical controllers - Part 1: overview and Res 2016;54:4487–505. https://doi.org/10.1080/00207543.2015.1064182.
fundamental principles. 2003. [61] Grangel-Gonzalez I, Halilaj L, Coskun G, Auer S, Collarana D, Hoffmeister M.
[30] ISO. ISO 10303-238: industrial automation systems and integration - Product data Towards a semantic administrative Shell for industry 4.0 components. 2016 IEEE
representation and exchange - Part 238: application protocol: application inter- Tenth International Conference on Semantic Computing (ICSC) 2016. p. 230–7.
preted model for computerized numerical controllers. 2007. https://doi.org/10.1109/ICSC.2016.58.
[31] Hardwick M, Loffredo D. Lessons learned implementing STEP-NC AP-238. Int J [62] Grangel-Gonzalez I, Halilaj L, Auer S, Lohmann S, Lange C, Collarana D. An RDF-
Comput Integr Manuf 2006;19:523–32. https://doi.org/10.1080/ based approach for implementing industry 4.0 components with administration

324
Y. Lu, et al. Journal of Manufacturing Systems 56 (2020) 312–325

shells. 2016 IEEE 21st International Conference on Emerging Technologies and [77] Leitão P. Agent-based distributed manufacturing control: a state-of-the-art survey.
Factory Automation (ETFA) 2016. p. 1–8. https://doi.org/10.1109/ETFA.2016. Eng Appl Artif Intell 2009;22:979–91. https://doi.org/10.1016/j.engappai.2008.
7733503. 09.005.
[63] Lu Y, Xu X. Resource virtualization: a core technology for developing cyber-physical [78] Yeung WL. Behavioral modeling and verification of multi-agent systems for man-
production systems. J Manuf Syst 2018;47:128–40. https://doi.org/10.1016/j.jmsy. ufacturing control. Expert Syst Appl 2011;38:13555–62. https://doi.org/10.1016/j.
2018.05.003. eswa.2011.04.067.
[64] Wang L, Feng HY, Song C, Jin W. Function block design for adaptive execution [79] Chou YC, Cao H, Cheng HH. A bio-inspired mobile agent-based integrated system
control of job shop machining operations. Int J Prod Res 2009;47:3413–34. https:// for flexible autonomic job shop scheduling. J Manuf Syst 2013;32:752–63. https://
doi.org/10.1080/00207540701666212. doi.org/10.1016/j.jmsy.2013.01.005.
[65] Wang L, Keshavarzmanesh S, Feng HY. Design of adaptive function blocks for dy- [80] Tharumarajah A, Wells AJ, Nemes L. Comparison of the bionic, fractal and holonic
namic assembly planning and control. J Manuf Syst 2008;27:45–51. https://doi. manufacturing system concepts. Int J Comput Integr Manuf 1996;9:217–26.
org/10.1016/j.jmsy.2008.06.003. https://doi.org/10.1080/095119296131670.
[66] Wang L, Adamson G, Holm M, Moore P. A review of function blocks for process [81] Ueda K, Vaario J, Ohkura K. Modelling of biological manufacturing systems for
planning and control of manufacturing equipment. J Manuf Syst 2012;31:269–79. dynamic reconfiguration. CIRP Ann Manuf Technol 1997;46:343–6. https://doi.
https://doi.org/10.1016/j.jmsy.2012.02.004. org/10.1016/s0007-8506(07)60839-7.
[67] Wang L, Cai N, Feng HY, Ma J. ASP: An adaptive setup planning approach for [82] Wang L, Haghighi A. Combined strength of holons, agents and function blocks in
dynamic machine assignments. IEEE Trans Autom Sci Eng 2010;7:2–14. https://doi. cyber-physical systems. J Manuf Syst 2016;40:25–34. https://doi.org/10.1016/j.
org/10.1109/TASE.2008.2011919. jmsy.2016.05.002.
[68] Hedberg T, Lubell J, Fischer L, Maggiano L, Feeney AB. Testing the digital thread in [83] Caesar B, Grigoleit F, Unverdorben S. (Self-)adaptiveness for manufacturing sys-
support of model-based manufacturing and inspection. J Comput Inf Sci Eng tems: challenges and approaches. Software-intensive cyber-physical systems vol.
2016:16. https://doi.org/10.1115/1.4032697. 34. Springer; 2019. p. 191–200. https://doi.org/10.1007/s00450-019-00423-8.
[69] Zhang Y, Qian C, Lv J, Liu Y. Agent and cyber-physical system based self-organizing [84] Brennan RW, Norrie DH. Evaluating the performance of reactive control archi-
and self-adaptive intelligent shopfloor. IEEE Trans Industr Inform 2017;13:737–47. tectures for manufacturing production control. Comput Ind 2001;46:235–45.
https://doi.org/10.1109/TII.2016.2618892. https://doi.org/10.1016/S0166-3615(01)00108-7.
[70] Chen B, Wan J, Lan Y, Imran M, Li D, Guizani N. Improving cognitive ability of edge [85] Buşoniu L, Babuška R, De Schutter B. A comprehensive survey of multiagent re-
intelligent IIoT through machine learning. IEEE Netw 2019;33:61–7. https://doi. inforcement learning. IEEE Trans Syst Man Cybern Part C 2008;38:156–72. https://
org/10.1109/MNET.001.1800505. doi.org/10.1109/TSMCC.2007.913919.
[71] Adolphs P, Auer S, Bedenbender H, Billmann M, Hankel M, Heidel R, et al. Structure [86] Panait L, Luke S. Cooperative multi-agent learning: the state of the art. Auton Agent
of the administration shell. Continuation of the development of the reference model Multi Agent Syst 2005;11:387–434. https://doi.org/10.1007/s10458-005-2631-2.
for the industrie 4.0 component. 2016. [87] Lowe R, Wu Y, Tamar A, Harb J, Abbeel P, Mordatch I. Multi-agent actor-critic for
[72] Gou L, Luh PB, Kyoya Y. Holonic manufacturing scheduling: architecture, co- mixed cooperative-competitive environments. Advances in Neural Information
operation mechanism, and implementation. Comput Ind 1998;37:213–31. https:// Processing Systems, Vol. 2017- December, Neural Information Processing Systems
doi.org/10.1016/S0166-3615(98)00100-6. Foundation 2017:6380–91.
[73] Leitão P, Restivo F. ADACOR: A holonic architecture for agile and adaptive man- [88] Lu Y, Xu X. Cloud-based manufacturing equipment and big data analytics to enable
ufacturing control. Comput Ind 2006;57:121–30. https://doi.org/10.1016/j. on-demand manufacturing services. Robot Comput Integr Manuf 2019;57:92–102.
compind.2005.05.005. https://doi.org/10.1016/j.rcim.2018.11.006.
[74] Zambrano Rey G, Pach C, Aissani N, Bekrar A, Berger T, Trentesaux D. The control [89] Xu X. From cloud computing to cloud manufacturing. Robot Comput Integr Manuf
of myopic behavior in semi-heterarchical production systems: a holonic framework. 2012;28:75–86.
Eng Appl Artif Intell 2013;26:800–17. https://doi.org/10.1016/j.engappai.2012. [90] Open Asset Administration Shell n.d. http://acplt.github.io/openAAS/ (accessed
08.011. May 12, 2020).
[75] Jana TK, Bairagi B, Paul S, Sarkar B, Saha J. Dynamic schedule execution in an [91] Smart Manufacturing Systems (SMS) Test Bed | NIST n.d. https://www.nist.gov/
agent based holonic manufacturing system. J Manuf Syst 2013;32:801–16. https:// laboratories/tools-instruments/smart-manufacturing-systems-sms-test-bed (ac-
doi.org/10.1016/j.jmsy.2013.07.004. cessed May 12, 2020).
[76] Monostori L, Váncza J, Kumara SRT. Agent-based systems for manufacturing. CIRP [92] Eclipse IoT - Leading open source community for IoT innovation n.d. https://iot.
Ann Manuf Technol 2006;55:697–720. https://doi.org/10.1016/j.cirp.2006.10. eclipse.org/ (accessed May 12, 2020).
004.

325