Content uploaded by Walter Terkaj
Author content
All content in this area was uploaded by Walter Terkaj on Oct 23, 2017
Content may be subject to copyright.
Semantic Virtual Factory supporting interoperable modelling and
evaluation of production systems
Botond Kádára(2), Walter Terkajb, Marco Saccob
aComputer and Automation Research Institute, Hungarian Academy of Sciences (MTA SZTAKI), Fraunhofer PMI PC, Kende u. 13–17, H-1111 Budapest, Hungary
bIstituto Tecnologie Industriali e Automazione, Consiglio Nazionale delle Ricerche (ITIA CNR) via Bassini 15, 20133, Milano, Italy
Modelling, simulation and evaluation of manufacturing systems are relevant activities that may strongly impact on the competi tiveness of production
enterprises both during the design and the operational phases. This paper addresses the application of a semantic data model for virtual factories to
support the design and the performance evaluation of manufacturing systems, while exploiting the interoperability between various Digital Enterprise
Technology tools. The paper shows how a shared ontology-based framework can be used to generate consistent 3D virtual environments and discrete
event simulation models, demonstrating this way how the proposed solution can provide an interoperable backbone for heterogeneous software tools.
Modelling, Simulation, Virtual Factory
1. Introduction
One of the main challenges in manufacturing today is to design
and operate systems producing a high variety of customized
products as efficiently and quickly as possible, while dealing with
uncertain and highly volatile demands. Managing manufacturing
companies and systems requires both long-term and short-term
decisions, which all deeply influence the performance of these
firms. From strategic point of view the decisions have impact on
longer time horizon (usually more than two years) and involve
major commitment of financial resources [1]. For instance,
strategic decisions may regard the number of plants or facilities
to be built, their size and their location, the variety of products to
be manufactured, the appropriate manufacturing technologies,
and, within a plant, the number and types of production
resources, the characteristics of the transportation and handling
systems and the degree of automation, to mention the most
important aspects only. From tactical and operational points of
view the decision makers should consider changes like expansion,
reduction or reconfiguration of production structures, mid-term
planning, short-term scheduling and optimized control of the
systems in question. On all levels of hierarchy the complexity of
these decisions and their importance from the point of view of the
profitability emphasizes the need to have formal and structured
approaches to support the design, management and performance
evaluation of production systems.
Nowadays managers and engineers usually apply Digital
Enterprise Technologies (DET) as decision support tools in
handling the challenges enumerated above [2]. The concept of
digital enterprise – the mapping of the key data and processes of
an enterprise to digital structures by means of information and
communication technologies – gives a unique opportunity for
planning and controlling the operation of production systems [3].
Although DET provide all the necessary components for
modelling, analysis and evaluation of production systems, in real
scenarios these tools and solutions are concurrently used by
different decision makers with various objectives. Mainstream
commercial Digital Factory (DF) tool providers already offer
integrated solution of their modular products but, on the one
hand, the procurement and deployment of such integrated
solutions are severely expensive and usually without the
warranty to effectively exploit all provided functionalities, on the
other, they requires specific knowledge and expertise frequently
inaccessible even in big enterprises.
This paper introduces a research work aiming at homogenizing
the modelling basis of production system and on this normalized
foundation allowing the smooth interoperation of different DF
tools that can be both commercial and self-developed.
2. Modelling production systems on semantic basis
Production system modelling may use different formalisms and
approaches, depending on the characteristics of the considered
problem and the expected results. Whether the system is a
machine tool, a production line, a distribution network or a
communication system, we can use modelling for formalizing and
gaining knowledge from the system at different life-cycle phases,
for evaluating a certain feature, for making comparison between
several reconfiguration options, for problem detection or for
evaluating and improving the system performance.
Formal, descriptive modelling methods proposed in Enterprise
Engineering (EE) discipline [4] support the analysis and re-
engineering (design) of existing (new) business entities – namely
business processes, application systems, business departments,
industrial plants and in the broadest sense, complete enterprises
or networks of enterprises. Descriptive methods like UML or
IDEFxxx [5] consist of modelling languages, defined in their
syntax and usually in correspondent graphical notations. In the
most complete cases, formal methods provide a reference
architecture where the supported modelling languages are
organized according to pre-defined criteria (e.g. GERAM, ARIS)
[6], [7].
2.1. Standardization efforts
From standardization point of view, several contributions have
faced the problem of developing a holistic and complete data
model for representing manufacturing systems, both considering
Contents lists available at SciVerse ScienceDirect
CIRP Annals Manufacturing Technology
Journal homepage: www.elsevier.com/locate/cirp
tangible (e.g. machine tool, workpiece to be produced, etc.) and
intangible (e.g. process plans, production logics, etc.) aspects.
ANSI/ISA-95 is an international standard for developing an
automated interface between enterprise and control systems. It
aims at providing both consistent terminology and information
models as well as reliable operations models [8]. A different
approach in the modelling of manufacturing process is offered by
the Process Specification Language (PSL) standard. PSL is an
ontology providing a way to formally describe a process and its
characteristic. The ontology has been developed at the National
Institute of Standards and Technology (NIST) and has been
approved as an international standard in the document (ISO
18629) [9].
Partially based on Standard for the Exchange of Product model
data (STEP), the Industry Foundation Classes (IFC) represents an
open specification for Building Information Modelling (BIM) data
that is exchanged and shared among the various participants in
an architecture, engineering and construction project [10].
2.2. The role of ontologies in production systems modelling
Several research efforts were already made to apply semantics
and create ontologies aiming at modelling the components and
the interconnections in manufacturing systems and/or
supporting the simulation model building of such systems. In [11]
the study introduces a component-based modelling and
simulation approach that supports model reuse across multiple
application domains called CODES. The attributes and behaviour
of the components are abstracted as meta-components and are
described using COML (COmponent Markup Language). The
integrated approach is supported by a component-oriented
simulation and modelling ontology called COSMO. Authors in [12]
presents the main issues and challenges of creating a simulation-
based modelling ontology. On the base of the experience coming
from the creation of the Discrete event Modeling Ontology
(DeMO) the authors propose the decomposition of the models in
behavioural and observable parts on the base of Hidden Markov
Models. An overview of the DeMO ontology is also given, focusing
on the semantic modelling of a discrete-event simulation kernel
including both event-driven and process oriented approach.
In [13] the authors explain the role of ontologies in facilitating
simulation modelling and highlighting the importance of
integrating and modularizing different simulation systems.
Special view is taken on the process of simulation model creation
and the authors underline the most relevant project phases
where ontology supports the common understanding. Further
possible roles of ontological modelling are also introduced like
distributed and component-based simulation. Finally the authors
present the Ontology driven Simulation Modelling Framework
(OSMF) solution, which provides a "visual programming
environment" to rapidly compose, build, and maintain
distributed, federated simulations
The research presented in [14] focuses on reusability and
composability aspects in simulation modelling of large-scale
systems independently from the type of simulation.
On the base of our review it was clear that semantic web
technologies, even if promising, were only partially used in
different, well-bordered segments of production systems’
modelling field. Inspired by this fact and the supplementary
knowledge manipulation opportunities provided by ontology-
based modelling, the European FP7 Virtual Factory Framework
(VFF) [15] project aimed at creating a new framework based on
semantic web technologies to manage the data and models
related to the whole factory life.
3. The Virtual Factory Framework
The Virtual Factory Framework (VFF) can be defined as “An
integrated collaborative virtual environment aimed at facilitating
the sharing of resources, manufacturing information and
knowledge, while supporting the design and management of all
the factory entities, from a single product to networks of
companies, along all the phases of the their life-cycles” [16].
As presented in Figure 1, the VFF architecture is based on three
main pillars:
• Virtual Factory Data Model (VFDM), i.e. a coherent, standard,
extensible, and common data model for the representation of
factory objects related to production systems, resources,
processes and products, i.e. the Data & Knowledge.
• Virtual Factory Manager (VFM), i.e. the software application that
manages and provides access to the shared repository
containing data structured according to the VFDM. A prototype
implementation of VFM as web-service is presented in [18].
• Digital Factory tools (or VF modules), i.e. the software tools that
are able to communicate with the VFM to retrieve and send
shared data formalized according to the VFDM. Specific VF
modules (e.g. Factory Image in Figure 1) may access real factory
data to synchronize the real and virtual representations.
Figure 1. The architecture of the Virtual Factory Framework
As an integrated semantic data model VFDM implements and
extends the core aspects of the IFC standard [10] to provide the
base for integration of various digital tools applied in different
planning and operation life-cycles phases of a manufacturing
system. The extensions, partly based on ISA-95 standard, were
developed to represent the characteristics of a manufacturing
system in terms of the products to be manufactured, the
manufacturing process they must undergo and the resources
entitled to operate the different manufacturing operations.
The VFDM is decomposed into a hierarchical structure of
ontologies, thus downsizing the problem and its complexity while
keeping a holistic approach. Indeed, the VFDM was designed as a
set of related ontologies [17] serialized as Web Ontology
Language OWL) files according to Resource Description
Framework (RDF/XML) [19]. The VFDM defines only the so-
called meta data (i.e. the classes, properties and restrictions),
whereas the actual instances (i.e. the individuals) are stored in
the data and knowledge repository (Figure 1). As an overview,
the following main areas are covered by the ontology classes
defined in the VFDM:
• Product, modelling the data related to the product, i.e. the
production goal of the factory.
• Resource, modelling the data related to the production
resources that are used by a system with the final goal of
Virtual
Factory
Data
Model
(VFDM)
Virtual Factory
Manager (VFM)
Data &
Knowledge
1
2
Real Factory Interface
3Digital
Factory tools
ARENA®
Plant Simulation
GIOVE-VF
…..
Factory Image
Legacy
System
Data
Base
Data
Base
transforming the product (or a work in progress). These
resources can be human operators, machines, transportation
and logistics related devices, etc.
• Process, modelling the data regarding the processes that are
adopted by the system to directly (e.g. manufacturing system,
assembly system) or indirectly (e.g. logistic processes,
maintenance processes) transform a product.
• System, modelling the data of a transformation system (e.g.
manufacturing, assembly, transportation and manipulation
systems) that affects a product by means of physical resources
and/or human resources within a process.
• Building, modelling the data related to the physical structure of
the factory and important for 3D layout planning and
visualisation aspects (e.g. walls, columns, floor, power supply
lines, etc.).
Further details about the VFDM with a comprehensive
description of relations between classes and properties are
described in [20].
4. VFF-based integration of Digital Factory tools
This section delves into the problem of integrating a Digital
Factory tool into the VFF, by developing a software layer named
Virtual Factory Connector (VfConn) that takes care of
Input/Output conversions from data stored in the shared
repository to the internal data structures of the Digital Factory
tool, and vice-versa. A proper VfConn may be developed only if
the following fundamental requirements are met:
• The developer knows both the VFDM and the specific data
model adopted by the Digital Factory tool, at least those parts
that are relevant for supporting the data exchange.
• The Digital Factory tool offers a way to access and modify (if
needed) its internal data structures, typically an application
programming interface (API).
Figure 2. Integration of a new Digital Factory tool in VFF
If such requirements are met, then a Digital Factory tool can be
integrated into VFF by realizing the following steps:
1. The classes and properties of the VFDM must be mapped to
the data model adopted by the Digital Factory tool. If
necessary, the VFMD can be extended by adding domain
ontologies that formalize application-specific concepts.
2. Development of the VfConn by adopting the best feasible
option according to the available technologies. Depending on
the language required by the API of the digital tool, different
programming libraries can be exploited to support the
development of the VfConn.
3. Population of the shared repository with data and knowledge
required as input by the newly integrated Digital Factory tool.
Such population can be accomplished by means of Graphical
User Interface (GUI) tools, by transferring data from existing
databases or legacy systems, or by integrating further digital
tools within VFF.
Figure 2. shows how a DF tool can be integrated in VFF thanks
to a specific VfConn, while referring to the common VFDM. The
functionalities of the DF tool can be used during the creation,
modification of data and relations about the model of the system
in question.
The following sub-sections address how two Digital Factory
tools were integrated into VFF and can be employed to
concurrently create and evaluate factory projects while sharing
consistent data. Specifically, a 3D layout design tool and a
commercial discrete-event simulation tool will be presented and
then applied to a common industrial case representing a
production line, as described in [22]. This production line
consists of seven machine tools, characterized by failure modes
that have to realize a part type by executing a process plan
decomposed into five stages that include a drilling, two sequential
milling, a quality control and a grinding operation.
4.1. Production system design and visualization
The production line of the industrial case can be designed and
placed in its building by using 3D software tools, like for instance
GIOVE Virtual Factory (GIOVE-VF) [21]. GIOVE-VF is a 3D virtual
reality collaborative environment aimed at supporting the factory
layout design. In particular, GIOVE-VF offers the user the
possibility to design factories by selecting machines, operators
and other resources from available catalogues and place them in
the 3D scene of the virtual factory (Figure 3). The virtual
environment can schematically display performance measures
that are provided by simulators and/or monitoring tools.
GIOVE-VF has been developed in C++ and can import/export
ontologies serialized in RDF/XML format thanks to a specific
VfConn that was developed using a C++ library named
VfConnectorLibCpp providing functionalities to parse, create and
modify the ontologies by exploiting an internal map between
OWL classes/restrictions and C++ classes/methods. The
VfConnectorLib library is based on the Redland C libraries and
makes use of an in memory RDF storage.
Figure 3. 3D representation of the test-case in the GIOVE-VF module
4.2. Production simulation with a commercial DF tool
Not only self-developed but also commercial DF tools can
interoperate thanks to the VFDM while designing a production
Virtual Factory Connector
in arbitrary programing language
Data and Knowledge Repository
containing VFDM-compliant Factory Projects
Digital Factory (DF) tool
MS04
MtC
Object mapping (optional)
Internal data structure
of the DF tool
DF tool
functionalities
MS01
DS01
MS03
MS04
MS02
GS01
CS01
system. The commercial off-the-shelf simulation package Plant
Simulation by Siemens PLM was also integrated in VFF to support
the performance evaluation of production systems. Figure 4
shows the similar test-case that was designed in the GIOVE 3d
virtual tool. A standard VfConn was designed and implemented in
Java language to easily retrieve and save factory projects from
and to the shared repository. This connector applies SPARQL
Protocol and RDF Query Language (SPARQL) queries for ontology
retrieval and upload. The functionalities and the user interface of
the commercial package are kept in the background and the
model of the production system is entirely defined in the
ontology. As the simulation software does not possess a direct
semantic reader a converter is required to correctly interpret the
input data and their interrelations. This conversion is based on a
specific XML table (represented as XML mapping in Figure 4.)
which was also designed according to the ISA-95 standard.
Figure 4.Representation of the test-case in Plant Simulation
The transformation of the model in the final, native Plant
Simulation format is performed inside Plant Simulation in the
SimTalk programming language. This process frees the user from
manually creating the simulation model that is automatically
generated from the initial VFDM-compliant factory project.
The creation of the simulation model is performed
automatically as well as the initialization of the parameters and
the model runs. In the current configuration the only data
required by the user is the production order in a form of list
indicating the products to be produced. A separate report and
data visualization tool has been developed as well to support the
presentation of simulation results. This is also achieved entirely
through VFDM-based interoperation.
A similar approach can be adopted to integrate also other
commercial DES tools into VFF, e.g. Arena as described in [22].
Conclusion
This paper has presented a framework to enable
interoperability between software tools supporting the design
and performance evaluation of a factory. The interoperability is
based on a common semantic data model for representing virtual
factories that was designed and implemented covering both the
structural and operational aspects of production systems.
custom-tailored or commercial tools can provide the same or
subsequent functionalities required in the life-cycle of a
manufacturing system. Further research should be carried out to
develop more efficient solutions for accessing and managing huge
amount of data in the shared repository and to exploit the
enablers of the Semantic Web approach to perform reasoning and
enrich the knowledge about specific manufacturing contexts.
Acknowledgment
The research reported in this paper has been funded by the
European Union 7th FP (FP7/2007–2013) under the grant
agreement No: NMP2 2010-228595, Virtual Factory Framework
(VFF), the grant agreement No: 262044, VISION Advanced
Infrastructure for Research (VISIONAIR) and National Office for
Research and Technology (NKTH) grant "Digital, real-time
enterprises and networks", OMFB-01638/2009.
References
[1] Terkaj, W.; Tolio, T.; V alente, A. (2009) Designing Manufacturing Flexibility in
Dynamic Production Conte xts, Design of Flexible Production Systems,
Springer, Ch. 1:1–18.
[2] Monostori, L.; Erdős, G.; Kádár, B.; Kis, T.; Kovács, A.; Pfeiffer, A.; Váncza, J.
(2010) Digital enterprise solution for integrated production planning and
control, Computers in Industry 61(2):112-126.
[3] Maropoulos, P.G.: Digital enterprise technology – Defining perspectives and
research priorities. In : Proc. of the 1st CIRP (UK) Sem. on Digital Enterprise
Technology (DET02), September 16-17, 2002, Durh am, UK, Part V: 3-12.
[4] Vernadat, F.B. (1994) Standards a nd prenorms in design, manufacturing and
automation, Handbook of Design, Manufacturing and Automation (Dorf, R.C.;
Kusiak, A. (Ed)), chapter 49:993-1019.
[5] Menzel, C.; Mayer R. J. (1998) The IDEF family of languages, H andbook on
Architectures on Information Systems (Bernus, Mertins, Schmidt ( Ed)), ch.
10:249–262.
[6] GERAM (1998) Generalised Enterprise Reference Architecture a nd
Methodology, IFIP-IFAC Task force. ISO 15704 (ISO TC184/SC5/WG1 N423).
Annex A.
[7] Scheer, A.W. (1999) ARIS- Business Process Modeling, Springer-Verlag,
Berlins.
[8] ISO (2010) ISA-95: the international standard for the i ntegration of
enterprise and control systems. URL: http://www.isa-95.com/, last retrieved:
October, 2012.
[9] National Institute of Standards and Technology. (2008) Process Specification
Language (PSL). URL: http://www.mel.nist.gov/psl/, last retrieved: October,
2012.
[10] buildingSMART (2012). Industry Foundation Classes – IFC2x Edition 4
Release Candidate 2. URL: http://buildingsmar t-
tech.org/ifc/IFC2x4/rc2/html/index.htm, last ret rieved: October, 2012.
[11] Teo, Y. M.; Szabo, C. (2007): CODES: An Integrated Approach to Composable
Modeling and Simulation, Asia Pacific Science and Technology Center SUN
Microsystems Inc.
[12] Miller, J. A.; Baramidze, G. (2005) Simulation and the semantic web,
Proceedings of the 2005 Winter Simulation Conference:2371-2377.
[13] Benjamin, P.; Patki, M.; Mayer, R. (2006) Using Ontologies for S imulation
Modeling, Proc. of the 2006 Winter Simulation Conference: 1151-1159.
[14] Balci, O.; Arthur, J.; Ormsby, W. (2011) Achieving reusability and
composability with a simulation conc eptual model, Journal of Simulation
1(5):157–165.
[15] VFF (2 012): VFF, Holistic, extensible, scalable and standard Virtual Factory
Framework (FP7-NMP-2008-3.4-1, 228595). URL: http://www.vff-
project.eu/, last retrieved: October, 2012.
[16] Sacco, M.; Dal Maso,G,; Milella, F.; Pedrazzoli, P.; Rovere, D.; Terkaj, W. (2011)
Virtual Factory Manager in Lecture Notes in Computer Science, Ed:
Springer:397-406.
[17] W3C (2004): OWL Web Ontology Language Reference, W3C
Recommendation 10 Feb-ruary 20 04. URL: http://www.w3.org/TR/owl -ref/,
last retrieved October, 2012.
[18] Ghielmini, G., Pedrazzoli, P., Rovere, D., Terkaj, W., Dal Maso, G., Milella, F.,
Sacco, M., Boer, C.R. "Virtual Factory Manager of Semantic Data," in
Proceedings of DET2011 7th International Conference o n Digital Enterprise
Technology, Athens, Greece, 2011
[19] W3C. (2004) RDF/XML Syntax Specification (Revised). [Online].
http://www.w3.org/TR/REC-rdfsyntax/ last retrieved December, 2012
[20] Terkaj, W; Pedrielli, G.; Sacco, M. (2012) Virtu al Factory Data Model,
Submitted to 2nd OSEMA (Ontology and Semantic Web for Manufacturing)
Workshop Graz, Austria, July 24, 201 2. [Online] http://ceur-ws.org/Vol-
886/paper_4.pdf.
[21] G.P. Vig anò, L. Greci, S. Mottura, M. Sacco, (2011) GIOVE Virtual Factory: A
New Viewer for a More Immersive Role of the User During Factory Design, in
Digital Factory for Human-oriented Production Systems, L., Re daelli, C.,
Flores, M. Canetta, Ed.: Springer, 201-216.
[22] Terkaj W, Urgo M (2 012) Virtual Factory Data Model to support Performance
Evaluation of Production Systems. Proceedings of OSEMA 2012 Workshop,
7th International Conference on Formal Ontology in Information Systems,
Graz, Austria, 24-27 July, 2012.
