See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/373919291

Towards Ontologizing a Digital Twin Framework for Manufacturing

Chapter · September 2023

DOI: 10.1007/978-3-031-43666-6_22

CITATIONS READS

3 393

6 authors, including:

Milos Drobnjakovic Guodong Shao

National Institute of Standards and Technology National Institute of Standards and Technology
17 PUBLICATIONS 69 CITATIONS 13 PUBLICATIONS 49 CITATIONS

SEE PROFILE SEE PROFILE

Boonserm Kulvatunyou Simon Frechette

National Institute of Standards and Technology National Institute of Standards and Technology
116 PUBLICATIONS 1,196 CITATIONS 39 PUBLICATIONS 840 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Boonserm Kulvatunyou on 07 October 2023.

The user has requested enhancement of the downloaded file.

 Towards Ontologizing a Digital Twin
Framework for Manufacturing

Milos Drobnjakovic(B) , Guodong Shao , Ana Nikolov, Boonserm Kulvatunyou,

Simon Frechette, and Vijay Srinivasan

National Institute of Standards and Technology, Gaithersburg, MD, USA

{milos.drobnjakovic,boonserm.kulvatunyou}@nist.gov

Abstract. This paper presents a timely opportunity for manufacturing researchers

to ontologize a digital twin framework for manufacturing by using recently pub-
lished international standards. An ISO/IEC 21838 series of standards was released
recently to address ‘top-level’ ontologies. These standards have been used by
industrial consortia to release standards called ‘mid-level’ ontologies. One such
‘mid-level’ ontology standards is the Industrial Ontologies Foundry (IOF) Core.
Around the same time, an ISO 23247 series of standards was released to stan-
dardize a digital twin framework, specifically for the manufacturing domain. This
paper proposes to apply existing top-level and mid-level ontologies to create a
‘domain-level’ ontology for a digital twin framework for manufacturing. This
domain-level ontology can then be used to create a manufacturing sector-specific
digital twin ontology, called the ‘application-level’ ontology. In this paper, that
application-level is the biomanufacturing sector. This paper also calls for a col-
laborative effort to create and deploy these two bottom-level ontologies and the
digital twin standards associated with them.

Keywords: ontology · digital twin · framework · standards · manufacturing ·

biomanufacturing

1 Introduction
The past few years have seen some important advances in the standardization of ontolo-
gies, and digital twins are gaining great interest in the manufacturing industry. On the
ontological front, a joint effort by ISO (International Organization for Standardization)
and IEC (International Electrotechnical Commission) has resulted in the publication of
a series of top-level ontology standards [1]. This original series of standards has been
picked up by other standards organizations, such as the OAGi (Open Applications Group
Inc.). OAGi used these top-level standards as the foundation for building and releasing
its own mid-level ontology standard [2] that is meant to benefit primarily the manufac-
turing industry. In the meantime, ISO has also released a series of standards on a digital
twin framework for manufacturing [3].
Both the ontology standards and the digital twin standards are of great interest to the
manufacturing sector. As a result, it is natural for the manufacturing industry to inquire
© IFIP International Federation for Information Processing 2023
Published by Springer Nature Switzerland AG 2023
E. Alfnes et al. (Eds.): APMS 2023, IFIP AICT 690, pp. 317–329, 2023.
https://doi.org/10.1007/978-3-031-43666-6_22
318 M. Drobnjakovic et al.

what benefits would accrue if these standards were used to ontologize a digital twin
framework for manufacturing. This paper examines such benefits and uses them to pro-
pose a collaborative research effort to create an ontology-based digital twin framework
for manufacturing. If successful, this framework will benefit not only existing manufac-
turing industries but also many emerging and important manufacturing industries, such
as biomanufacturing.
The rest of the paper is organized as follows. Section 2 gives a brief description of
the existing top-level and mid-level ontology standards. A similar briefing is given in
Sect. 3 about manufacturing-specific digital twin framework standards. The main task of
Sect. 4 is to propose how the existing domain-independent ontology standards described
in Sect. 2 can be used to create a domain-level ontology for the digital twin framework
described in Sect. 3. In Sect. 5, application-level ontologies are proposed for emerging
and important areas using biomanufacturing as an example. Section 6 provides some
concluding remarks.

2 A Hierarchy of Ontologies
The term ‘Ontology’ has a rich and colorful history in the human world, dating back to
the development of philosophy, metaphysics, and theology. Since its recent appearance in
the computational world, however, the term ‘ontology’ has taken a decisively utilitarian
turn. It has a new focus, which is no longer about philosophy, metaphysics, or theology;
now it is about using rigorous mathematics and logic to create data-driven informational
terms and data-driven relational expressions. Both live in the digital world, and both are
associated with the data being collected in the real world. Those resulting and varied
data repositories are now being used across a wide range of scientific and industrial
disciplines.
This new ontology-based focus was first triggered by earlier needs in two scientific
disciplines: biological sciences and life sciences. The first joint need was a gene ontol-
ogy. The success of addressing some of those more recent needs for different kinds of
bioscience ontologies has, in turn, triggered some serious attempts to define a related
ontology hierarchy using a hub-and-spoke approach. That approach is described below.
One of the seminal developments behind this hierarchical approach was the Basic
Formal Ontology (BFO), which was soon recognized as one of the Top-Level Ontolo-
gies (TLOs), and it provides a basis for other ontologies. In fact, BFO has been used
successfully to build several bio-related ontologies. BFO is the first TLO that has been
standardized recently by ISO and IEC [1, 4]. Table 1 shows some of the information
terms and relational expressions formally standardized (that is, with axioms) in BFO.
Currently there are a total of 84 terms and relationships in BFO, attesting to the com-
pactness of this TLO. A directed, ontology graph can be created using the terms as the
nodes and the relational expressions as the directed arcs.
At the root of BFO is the informational term ‘entity.’ Fig. 1 shows how this term is
further refined and specialized using the ‘is a’ relational expression. TLO was followed
by the creation of Mid-Level Ontologies (MLOs) built on TLO. Two of the recently
developed MLOs are the Common Core Ontologies (CCO) and the Industrial Ontologies
Foundry Core (IOF Core) [2], both built on BFO as the TLO. Some of the informational
 Towards Ontologizing a Digital Twin Framework for Manufacturing 319

Table 1. Examples of terms and relational expressions standardized in BFO [1].

Terms Relational expressions

entity is a
continuant continuant part of
occurrent occurrent part of
spatial region located in
temporal region exists at
material entity
process
quality

Fig. 1. A BFO hierarchy of terms using the is a relationship [1].

terms and relational expressions that are formally standardized by IOF Core are shown
in Table 2. Currently, there are 57 terms and 38 relational expressions in IOF Core,
supporting a wide range of industrial (including manufacturing) needs.
Several domain- and application-level ontologies have been built from BFO and IOF
Core, using a hierarchical, hub-and-spoke approach. A Supply Chain Reference Ontol-
ogy (SCRO) is one such domain-level ontology. It defines the terms and relationships for
320 M. Drobnjakovic et al.

the structure (members and their roles, functions, capabilities, relations, and resources)
and operations (processes and flow of material and information) of supply chains. An
application-level ontology called the Supply Chain Traceability (SCT) ontology has been
created from SCRO. SCT is an ontology for supply chain traceability within a specific,
manufacturing supply chain in agriculture.

Table 2. Examples of terms and relational expressions standardized in IOF Core [2].

Terms Relational expressions

design specification acts on behalf of at some
time
material artifact satisfies requirement
engineered system is available to at some time
equipment role has output
product production
process

3 A Digital Twin Framework for Manufacturing

Around the same time as the ontology standards were released, another ISO series of
standards (called the ISO 23247 series) was also released [3, 5]. This series defines a
digital twin framework specifically for the manufacturing sector. These newer standards
provided both a ‘fit-for-purpose’ digital representation of each Observable Manufactur-
ing Element (OME) and the temporal synchronization needed between the OME and
its digital twin representation. An OME, in this context, is a physical manufacturing
element. Figure 2 shows the interconnected layers of the standardized framework for
building digital twins for manufacturing. It should be noted that the ISO 23247 series
provides only a framework, a structure if you will, standard. The series currently does
not include a standard for modeling the data/information that resides on top of that
structure/framework. Such a standard is needed. But first, more about the OMEs.
Observable Manufacturing Elements (OMEs) are classified under eight types, each
of which can contain up to seven attributes, as shown in Table 3. It should be noted that
each of the OME types in Table 3 is an ‘entity’ from an ontological perspective. Also,
each of the attributes is an entity. In addition, Fig. 2 shows layers that contain ‘device
communication entity’, ‘digital twin entity’, ‘user entity’, and ‘cross-system entity.’ Each
of these entities are further broken down into a few sub-entities and functional entities [3,
5]. A comparison of Tables 1, 2, and 3 shows the striking similarity among the terms and
relationships defined in the ISO/IEC 21383 series and the IOF Core ontology standards,
and those defined in the ISO 23247 series of standards on digital twin framework for
manufacturing.
As the popularity of digital twins grows and this growth drives the building of several
digital twins in manufacturing, an important need for interoperability among them has
 Towards Ontologizing a Digital Twin Framework for Manufacturing 321

User Entity

Cross-system Entity
Digital Twin Entity

Device Communication Entity

Observable Manufacturing Elements

Fig. 2. A standardized digital twin framework for manufacturing [3].

Table 3. Standardized OME types and attributes [3].

Types Attributes
personnel identifier (mandatory)
equipment characteristics
material schedule
process status
facility location
environment report
product relationship
supporting document

emerged. An attractive approach to satisfy this need is to use the recently developed top-
level and mid-level ontology standards to ontologize the digital twin framework standard
for manufacturing. The attractiveness stems from the fact that ontologizing the digital
twin could potentially provide (1) logical rigor and thus unambiguity of term definitions,
(2) explicit logical connections that permit higher connectivity between various data
sources associated with a digital twin, and (3) enable more effective knowledge reuse
and sharing. This has been recognized by the broader research community resulting
an increasing amount of literature pertaining to utilizing ontologies for digital twin
representation and data integration [6–9].
322 M. Drobnjakovic et al.

4 A Domain-Level Ontology for Digital Twin Framework

for Manufacturing
As described in Sect. 3, the newly released ISO 23247 series provides a structure for
creating, managing, and using digital twins for manufacturing applications. In addition
to the structure, the series (1) defines the terms, relationships, components, and processes
necessary for developing a digital twin and (2) provides guidelines for their implemen-
tation. The intent is that users of this series of standards can be assured that digital twins
are consistent, accurate, shareable, and reusable across different organizations and sys-
tems. However, at present, there is no formal ontological support for making such an
assurance for the ISO 23247 series.
As proposed below, this situation can be remedied by expanding the hierarchy of
ontology standards, as described in Sect. 2, ‘downward’. The first downward expansion
includes a specific, domain-level (in this case, the physical, manufacturing domain)
ontology. The proposed domain-level ontology for manufacturing will include a set of
manufacturing-related terms, relationships, and definitions, all of which will be based on
the structure and the meaning of the existing digital twin framework. This new ontology
could provide a common understanding of the data and information used to create digital
twins across the entire manufacturing domain. It could also (1) greatly improve the
efficiency and effectiveness of a digital twin development, and (2) ensure that digital
twins can be easily integrated and shared between different systems, organizations, and
sectors.
The terms and relationships that are already defined in the ISO 23247 series can
serve as the starting point for developing such a domain-level ontology. For example,
Table 3 lists the terms for eight OMEs, called entities, and seven attributes. As a concrete
example, consider the entity ‘personnel’ from OME list. According to ISO 23247-1 “Per-
sonnel in manufacturing generally include those employees who are engaged directly or
indirectly in manufacturing.” It is a textual definition that clearly indicates that ‘person-
nel’ is a person who is also an employee with a role in manufacturing. More generally,
‘personnel’ could be a contractor employed in a different organization. Since the same
term can have two different meanings, formalizing it is important. Such terms can be
made more formal using the existing, standardized top-level and mid-level ontologies,
as shown in Fig. 3.
In Fig. 3, the top-level is the BFO and the mid-level is the IOF Core. Additionally,
Fig. 3 uses color coding to indicate where the various terms originated. Other terms
related to OMEs (the types and their attributes) in Table 3 can be ontologized in a similar
fashion. Each digital twin can be related to its entity’s ‘information content’, which is
already available in IOF Core as shown in Fig. 4. The terms for various components of
the digital twin are already defined in the ISO 23247 series and are formally related in
Fig. 4 using the indicated relational expressions that are already defined in the BFO/IFO
Core standards.
Another important concept in the digital twin framework is the bidirectional rela-
tionship between the OME entities and the digital twin entities. Such bidirectional rela-
tionships are critical to maintain the necessary synchronization between both. This syn-
chronization is implemented as the ‘device communication’ entity layer shown in Fig. 2.
In this layer, ‘data collection’ from the OME is performed by sensors, and the OME is
 Towards Ontologizing a Digital Twin Framework for Manufacturing 323

Fig. 3. Introducing the personnel (manufacturing employee) OME into BFO and IOF Core
hierarchy.

‘controlled’ using actuators and controllers. Figure 5 shows how the collected data can
be used to temporally link the OME to its digital twin. Figure 6 shows how the OME
control can be temporally linked to the OME and its digital twin.
The preliminary examples provided in these figures should be viewed only as tenta-
tive first steps towards the goal of building the bottom ontology layers of the proposed
standardized digital twin framework for manufacturing. More work is needed to (1) cover
all the relevant informational terms and relational expressions found in the ISO 23247
series [3] and (2) get their ontologies approved by appropriate standards development
organizations and industrial consortia.
To accomplish those two needs, this paper proposes a joint project between ISO/TC
184/SC 4/WG 15 and OAGi IOF Core team. This project will help develop a new
standard that is more accurate, consistent, and interoperable than the existing ones. This
new standard will pave the way for creating, managing, and using digital twins in the
manufacturing industry. In such a joint project, the ISO team will provide the guidelines
for creating and using digital twins, while the OAGi team will provide the foundation
for representing industrial concepts and relationships. This combination would then
enable manufacturers to create digital twins that are based on a common vocabulary and
framework. This will make it easier for them to exchange information and collaborate
324 M. Drobnjakovic et al.

Fig. 4. Introducing digital twin and its components into BFO and IOF Core hierarchy.

with other organizations. The proposed joint project could also develop software tools
and platforms that support the implementation of digital twins.

Fig. 5. Data collection is connected to OME and its digital twin using BFO/IOF Core relations.
 Towards Ontologizing a Digital Twin Framework for Manufacturing 325

Fig. 6. Control is connected to OME and its digital twin using BFO/IFO Core relations. See Fig. 5
for Legend.

5 An Application-Level Ontology Example – Biomanufacturing

This section will use the ontologies from previous sections to build a biomanufacturing-
specific application-level ontology. Biomanufacturing encompasses production pro-
cesses that utilize living organisms or cell-derived macromolecules (e.g., enzymes) to
produce various products (e.g., bioethanol and bioplastics) [10]. Biomanufacturing can
be split into three distinct areas: bio-industrial, bio-medical, and agri-food. Compared to
traditional manufacturing, biomanufacturing relies more on renewable resources (e.g.,
biomass, sunlight, CO2 ) [11].
Regardless of these potential benefits, market competitiveness within biomanufac-
turing, as well as with traditional manufacturing processes, mandates the development
of production methods that can achieve (1) high yields with minimal manufacturing
costs and (2) cost-effective process development. One of the enabling technologies for
developing these methods is the concept of digital twins. As described in Sect. 3, a
digital twin transcends the capabilities of traditional models by establishing bidirec-
tional communication with the physical entity it represents [12]. In biomanufacturing,
this bidirectionality permits the control of the manufacturing processes in real time and
enables updates based on the physical world’s feedback.
During their research and development, scientists and engineers can utilize digital
twins to gain a dynamic understanding of a process. Such an understanding can help
lower the needed experimentation and thus reduce the developmental cost and time
[13]. Digital twins also permit the real-time optimization of a process, which is essen-
tial for increasing efficiency and maintaining productivity as the bioprocess progresses.
Furthermore, digital twins have many process-associated predictive capabilities, such
as predicting the equipment failure time [14], key process attributes (e.g., yield, con-
centration) [15], batch end times [12], and process parameters [16]. These predictive
capabilities can inform real-time or offline decision making in all stages of a bioprocess
326 M. Drobnjakovic et al.

lifecycle [17]. In other words, the dynamic nature of a digital twin and its bidirectional
connection with the physical world facilitates an optimized, cost-effective process by
(1) providing critical insights to various experts over the entire manufacturing lifecycle
and (2) reliably controlling the parameters of a process over prolonged periods.
To gain the full benefit of applying digital twins in biomanufacturing, an application-
level ontology for biomanufacturing is desirable. For example, terms such as equipment
and process, which are defined in that manufacturing domain ontology, can be fur-
ther specialized to biomanufacturing terms such as bioreactor, chromatography column,
and batch process. It should be noted that for the development and formalization of
such biomanufacturing terms, many constructs from other, existing application-level
ontologies and standards related to biomanufacturing could also be reused (e.g., ISA-88,
Allotrope ontology, OBO Foundry ontologies).

Fig. 7. Example of data collection from a bioreactor to update its digital twin.

As an example of application-level ontology for biomanufacturing, consider a biore-

actor (an essential biomanufacturing equipment) and its digital twin. Figure 7 represents
how the data collected from sensors attached to a bioreactor are used to update the dig-
ital twin of the bioreactor. The other direction represents how simulation results from
a digital twin of the bioreactor are used to control the parameters of actuators of the
bioreactor, as shown in Fig. 8. Data represented using these models enable, for example,
comparisons and improvements across predictive models used in the same or similar
processes.
 Towards Ontologizing a Digital Twin Framework for Manufacturing 327

Fig. 8. Example of a bioreactor parameter control based on digital twin simulation results. See
Fig. 7 for Legend.

6 Concluding Remarks

This paper presented recent advances in the standardization of two important technolo-
gies – one is a hierarchy of ontologies to support digital manufacturing and the other
is a framework for deploying digital twins in manufacturing. Time is now ripe to com-
bine these two advanced developments in standards, and to undertake a research and
development project to ontologize the standardized digital twin framework for man-
ufacturing. This domain-level ontology can then be used to develop application-level
ontology for any specific manufacturing sector. This paper focused on one emerging and
important manufacturing sector – namely, biomanufacturing. Future work can entail
applying the outlined principles to other manufacturing sectors, such as additive man-
ufacturing. Another area of future interest is to utilize the combination of ontologies
and digital twins to enhance the understanding of various process parameters on process
sustainability. Some initial steps taken in this direction are presented in more detail in
[18].
Towards the goal of ontologizing the digital twin framework for manufacturing, this
paper has included some preliminary ideas and approaches for an R&D project. This
project can be jointly managed and executed by the appropriate standards development
organization (SDO) and industrial consortia, both of which have been described in this
paper. In addition, IFIP WG 5.7 would be an appropriate community to contribute to this
effort. By leveraging their expertise and resources, IFIP WG 5.7 can play three crucial
roles: (1) providing technical contributions, (2) performing case studies, and (3) helping
industrial adoption. By executing these roles, IFIP WG 5.7 can deliver value to the entire
manufacturing community and, more specifically, to the biomanufacturing industry.
328 M. Drobnjakovic et al.

Once completed, the new and future ontology-based standards will bring benefits to
the bigger manufacturing community. These benefits include advanced knowledge and
expertise in the field of digital twins, access to new resources and technologies, improved
impacts and outcomes, and opportunities for further collaboration.

References
1. ISO/IEC 21838-1&2: Information technology – Top-level ontologies (TLO), Parts 1 and 2,
International Organization for Standardization, Geneva, Switzerland (2021)
2. Drobnjakovic, M., Kulvatunyou, B., Ameri, F., Will, C., Smith, B., Jones, A.: The industrial
ontologies foundry (IOF) core ontology. In: Proceedings of the FOMI 2022: 12th International
Workshop on Formal Ontologies meet Industry, Tarbes, France (2022)
3. ISO 23247-1. ISO 23247-1 Through 4: Automation Systems and Integration - Digital
Twin Framework for Manufacturing – Parts 1 Through 4, International Organization for
Standardization, Geneva, Switzerland (2021)
4. Drobnjakovic, M., Kulvatunyou, B., Frechette, S., Srinivasan, V.: Recent developments in
ontology standards and their applicability to biomanufacturing, IDETC/CIE2023-116866. In:
Proceedings of the ASME 2023 International Design Engineering Technical Conferences and
Computers and Information in Engineering Conference, Boston, MA, USA, 20–23 August
2023 (2023)
5. Shao, G., Frechette, S., Srinivasan, V.: An analysis of the new ISO 23247 series of standards
on digital twin framework for manufacturing, MSEC 2023-101127. In: Proceedings of the
ASME 2023 Manufacturing Science and Engineering Conference, New Brunswick, NJ, USA,
12–16 June 2023 (2023)
6. Barros, C.D., Salles, R., Ogasawara, E.S., Guizzardi, G., Porto, F.A.: Requirements for an
ontology of digital twins. In: International Conference on Semantic Systems (2021)
7. Dai, S., Zhao, G., Yu, Y., Zheng, P., Bao, Q., Wang, W.: Ontology-based information modeling
method for digital twin creation of as-fabricated machining parts. Robot. Comput.-Integr.
Manuf. 72, 102173 (2021)
8. Singh, S., et al.: Data management for developing digital twin ontology model. Proc. Inst.
Mech. Eng. Part B: J. Eng. Manuf. 235(14), 2323–2337 (2021)
9. Steinmetz, C., Rettberg, A., Ribeiro, F.G.C., Schroeder, G., Pereira, C.E.: Internet of things
ontology for digital twin in cyber physical systems. In: 2018 VIII Brazilian Symposium on
Computing Systems Engineering (SBESC), pp. 154–159. IEEE (2018)
10. Stikane, A., Dace, E., Stalidzans, E.: Closing the loop in bioproduction: spent microbial
biomass as a resource within circular bioeconomy. New Biotechnol. 70, 109–115 (2022).
https://doi.org/10.1016/j.nbt.2022.06.001
11. Schilling, C., Weiss, S.: A roadmap for industry to harness biotechnology for a more circular
economy. New Biotechnol. 60, 9–11 (2021). https://doi.org/10.1016/j.nbt.2020.08.005
12. Udugama, I.A., Lopez, P.C., Gargalo, C.L., Li, X., Bayer, C., Gernaey, K.V.: Digital Twin in
biomanufacturing: challenges and opportunities towards its implementation. Syst. Microbiol.
Biomanuf. 1(3), 257–274 (2021). https://doi.org/10.1007/s43393-021-00024-0
13. Sokolov, M., von Stosch, M., Narayanan, H., Feidl, F., Butté, A.: Hybrid modeling—A key
enabler towards realizing Digital Twins in biopharma? Curr. Opin. Chem. Eng. 34, 100715
(2021). https://doi.org/10.1016/j.coche.2021.100715
14. Chen, Y., Yang, O., Sampat, C., Bhalode, P., Ramachandran, R., Ierapetritou, M.: Digital Twins
in pharmaceutical and Biopharmaceutical Manufacturing: a literature review. Processes 8(9),
1088 (2020). https://doi.org/10.3390/pr8091088
 Towards Ontologizing a Digital Twin Framework for Manufacturing 329

15. Hengelbrock, A., et al.: Digital twin for HIV-Gag VLP production in Hek293 cells. Processes
10(5), 866 (2022). https://doi.org/10.3390/pr10050866
16. Helgers, H., Schmidt, A., Strube, J.: Towards autonomous process control—Digital twin for
Cho cell-based antibody manufacturing using a dynamic metabolic model. Processes 10(2),
316 (2022). https://doi.org/10.3390/pr10020316
17. Park, S.-Y., Park, C.-H., Choi, D.-H., Hong, J.K., Lee, D.-Y.: Bioprocess Digital Twins of
mammalian cell culture for advanced biomanufacturing. Curr. Opin. Chem. Eng. 33, 100702
(2021). https://doi.org/10.1016/j.coche.2021.100702
18. Nikolov, A., Drobnjakovic, M., Kulvatunyou, B.: Produce it sustainably: life cycle assess-
ment of a biomanufacturing process through the ontology lens. In: APMS Conference 2023,
Trondheim, Norway (2023).

View publication stats