ArticlePDF Available

Abstract and Figures

The growing importance of manufacturing SMEs within the European economy, in terms of Gross Domestic Product and number of jobs, emphasizes the need of proper ICT tools to support their competitiveness. Major ICT players already offer one-does-all Product Lifecycle Management suites, however, these do show consistent shortcomings in terms of SME accessibility, degree of personalization and they often lack of an acceptable level of interoperability. These problems are being addressed by the development of a Virtual Factory Framework (VFF). The approach is based on four pillars: (1) Semantic Shared Data Model, (2) Virtual Factory Manager (VFM), (3) Decoupled Software Tools and (4) Integration of Knowledge. This paper will focus on the Virtual Factory Manager that acts as a server supporting the I/O communications within the framework and its stored knowledge for the Decoupled Software Tools needing to access its repository.
Content may be subject to copyright.
Virtual Factory Manager for Semantic Data Handling
Giorgio Ghielminia, Paolo Pedrazzolia, Diego Roverea, Walter Terkajb, Claudio R. Boëra, Giovanni Dal Masoc,
Ferdinando Milellad, Marco Saccob
a ICIMSI-SUPSI, Galleria 2, 6928 Manno, Switzerland
b ITIA-CNR, Via Bassini 15, 20133 Milano, Italy
c Technology Transfer System S.r.l., Via Pacini 15, 20131 Milano, Italy
d SimX Ltd, Furness House, Salford Quays, Manchester M503XA, UK
The growing importance of manufacturing SMEs within the European economy, in terms of Gross
Domestic Product and number of jobs, emphasizes the need of proper ICT tools to support their
competitiveness. Major ICT players already offer one-does-all Product Lifecycle Management suites,
however, these do show consistent shortcomings in terms of SME accessibility, degree of personalization
and they often lack of an acceptable level of interoperability. These problems are being addressed by the
development of a Virtual Factory Framework (VFF). The approach is based on four pillars: 1) Semantic
Shared Data Model, 2) Virtual Factory Manager (VFM), 3) Decoupled Software Tools, 4) Integration of
Knowledge. This paper will focus on the Virtual Factory Manager that acts as a server supporting the I/O
communications within the framework and its stored knowledge for the decoupled software tools needing
to access its repository.
Virtual Factory, Enterprise Modelling, Reference Model, Interoperability, Semantic Data Model
Market needs and expectations require a continuously rapidly evolving production framework: thus production
systems, from small to large scale and integrated factories, have to be conceived and set-up in shorter and shorter
times (Chryssolouris et al, 2008). Several critical aspects, related to this need of rapid prototyping of factories,
have to be addressed: it is critical to provide sufficient product variety to meet customer requirements, business
needs and technical advancements (Huang et al, 2005), while maintaining economies of scale and scope within
the manufacturing processes (Terkaj et al, 2009). Therefore, the current challenge in manufacturing engineering
consists in the innovative integration of the product, process and factory worlds and the related data, aiming at
synchronizing their lifecycles (Tolio et al, 2010).
The creation of a holistic, integrable, up-gradable, scalable Virtual representation of the Factory can empower
this synchronization, promoting high cost savings in the implementation of new manufacturing facilities or
reconfiguration of existing ones, thanks to the effective virtual representation of buildings, resources, process,
and products: this is shown both by industrial practice and academic scientific research. The entire factory is
simulated as a continuous and consistent digital model, which can be used, without interruption, all the way from
the product idea to the final dismantling of the production plants and buildings (Bracht and Masurat, 2005).
This challenge is being addressed by the development of a Virtual Factory Framework (VFF), within an EU
funded project ( The approach is based on four pillars: 1) Semantic Shared Virtual
Factory Data Model (VFDM), 2) Virtual Factory Manager (VFM), 3) Decoupled Software Tools, based on the
VFDM and that can interact through the VFM, and 4) Integration of Knowledge. VFF objective is to fosters an
integrated virtual environment that supports factory processes along all the phases of its lifecycle.
This paper will focus on the Virtual Factory Manager (VFM), proposing an evolution of the former VFF Pillar II
(Sacco et al, 2010). This evolution finds its justification in the identified weakness of the former second Pillar,
found both in the support to the data consistency check against modifications performed by different modules
and in the integration of the knowledge layer with the pure factory data layer.
Semantic shared
Integration of knowledge
Virtual factory
Decoupled modules
Knowledge Manager
Figure 1 - The Semantic Virtual Factory Framework architecture
A viable solution has been identified in the adoption of ontology as means for data and relationships
representation, promoting knowledge integration in the data-model. This approach introduces a modification in
the overall VFF architecture, where pillar IV (knowledge integration) is no longer seen as foundation of Pillar I
(reference data model), as presented in (Sacco et al, 2010), but rather considered as an additional decoupled
module (Figure 1).
This paper presents the new VFF framework born from this evolution.
As mentioned, an answer to the market requirements previously highlighted has been provided by the
development of a first version of the Virtual Factory Manager (Sacco et al, 2011). That solution proved the
validity of the concept of having an integrated virtual environment supporting the design and management of all
the factory entities, ranging from the single product to the network of companies, along all the stages of the
factory lifecycle. The centralized data management platform based on a common description of the digital
factory demonstrated the capability to improve the integration process between software design tools (existing
and new developed ones) and to provide a shared knowledge base to be used during the factory modelling
phases. Nevertheless, the former approach showed some weaknesses both in the support to the data consistency
check against modifications to parts of the factory instances performed by different modules and in the
integration of the knowledge layer with the pure factory data layer. The result, highlighted by advanced tests,
was a framework affected by some problems of usability.
The main cause of this situation has been identified in the impossibility of the previous implementation of the
VFDM, to represent not only valid data structures, but also their semantics. A viable solution has been identified
in the adoption of ontology as means for data and relationships representation in order to improve the integration
of knowledge among the VFF pillars. This approach introduces some modifications in the whole VFF picture,
affecting the way the components of the architecture interact. Therefore, the result is a tighter cooperation
between the Knowledge Manager and the VFDM pillars.
Figure 1 shows the new architecture of the Semantic Virtual Factory Framework composed by the four pillars of
Semantic Shared Data Model (Pillar I), Semantic VF Manager (Pillar II), Decoupled VF Modules (Pillar III) and
Knowledge Manager (Pillar IV).
The Semantic VFDM establishes a coherent standard extensible set of ontologies for the common integrated
representation of the factory objects and of the factory knowledge domain, basing on the tools of the semantic
web (mainly the Web Ontology Language - OWL). Section 3 is dedicated to a thorough description of the new
Semantic VFDM approach and of the reasons that drove the change.
This common ontology set is governed by the Semantic VFM (Pillar II) that completes the functionalities of
access control, data versioning and selective data query, already implemented by the previous VFM, with a full
support to the semantic data validation. In this way, Decoupled Modules (Pillar III) modifying single parts of the
factory data immediately receive feedback on the consistency of their actions with the overall definition of the
factory instance. Section 4 and Section 6 are respectively dedicated to the analysis of the Semantic VFM and to
the description of the current prototype implementation.
In order to couple with the new features of the VFM and with the new data exchange formats, it has been
necessary to intervene on the internal architecture of the Decoupled Modules and, in particular, of their VF
Connector modules. Section 5 reports on the new structure of Pillar III, while an example of new module
interacting with the Semantic VFM prototype is provided in Section 7. Also the Knowledge Manager (Pillar IV),
that was the only component of the previous architecture already based on the usage of ontologies, has been
affected by the new approach. With the new structure, in fact, it can directly interface the Semantic VFM as a
decoupled module managing a dedicated part of the Semantic Data Model ontology. In this way it has been
removed another weak point of VFF that was represented by the need for adaptation of formats and protocols
between pillars.
The Reference Model (Pillar I) establishes a coherent standard extensible Virtual Factory Data Model (VFDM)
for the common representation of factory objects related to production systems, resources, processes and
products. The common data model can be considered as the shared meta-language providing a common
definition of the data that will be governed by the VFM (Pillar II) and used and updated by the Decoupled
Functional Modules (Pillar III).
According to the original requirements, the VFDM has to be holistic, covering all the relevant fields related to
the Factory domain and exploit existing technical standards to represent the data. Moreover the VFDM has to be
extensible and guarantee the proper granularity, providing at the same time the enablers for data consistency,
data safety, and proprietary data management.
Sacco et al (2011) conceived the VFDM as a set of XSD files (W3C, 2004c) defining the structure of the XML
files that would be stored and managed by the VFM. This solution offers relevant advantages in terms of:
Syntactic validation of the XML files according to the defined XSD files.
Rich expressiveness since several default data types can be further extended and complex constraints
and properties can be modelled.
Possibility to integrate several XSD files within a single project.
However, the XSD technology alone is not suitable for knowledge representation and several flaws can be
No explicit characterization of data with their relations on a semantic level.
Intradocument references are supported but interdocument references (cross-references) are poorly
modelled, thus endangering referential consistency.
Distributed data can be hardly managed.
The integration of different knowledge domains can be cumbersome.
The presented considerations led to evaluate and finally adopt the Semantic Web technologies which offer key
advantages to the whole VFF because they enables to:
Represent a formal semantics.
Efficiently model and manage distributed data.
Ease the interoperability of different applications.
Process data outside the particular environment in which it was created
Exploit generic tools that can infer from and reason about an ontology, thus providing a generic support
that is not customized on the specific domain.
The new semantic VFDM has been designed as an ontology (W3C, 2004a) by adopting the OWL language
(W3C, 2004b). In particular, it defines all the classes, properties and restrictions that can be used to create the
individuals to be stored in the data repository (Pillar II). Given the wide range and heterogeneity of the
knowledge domains to be covered by the VFDM in the scope of VFF, it is necessary to integrate various
knowledge domains as already highlighted by Colledani et al (2008) and Valente et al (2010) in previous related
works. Therefore, the VFDM has been decomposed into macro areas (i.e. bricks), creating a hierarchical
structure of sub-ontologies that have been named Factory, Building, System, Resource, Process, Product,
Strategy, Performance and Management. This architecture allows decomposing the problem, downsizing its
complexity while keeping a holistic approach. These sub-ontologies have been developed by referring to the
state-of-the-art technical standards available in the different domains, and in particular the Industry Foundation
Classes (IFC2x3, 2006), STEP-NC (ISO 14649-10:2004), and ISA-95 (ISA-95).
This section presents the analysis of the requirements for the VFM (Sect. 4.1) and its proposed architecture
(Sect. 4.2).
The main goal of the VFM design and implementation consists in obtaining an open integration platform
representing a common and shared communication layer between already existing and newly developed software
tools to support the factory design and management.
The preliminary architecture of VFM proposed by Sacco et al (2011) was related to a VF Data Model based on
the XSD/XML format. The adoption of an ontology-based representation of the VF Data Model in the VFF
project has led to a re-design of the VFM where Semantic Web technologies have been exploited. In the new
architecture previous basic requirements have been extended to include specific semantic functionalities:
Platform independent interfacing capabilities. The VF modules are software tools developed by
different vendors/organizations, with different programming languages, operating systems and HW
architectures. The VFM has to interface all of them by providing its service in an open and “proper”
Management of concurrent access and data consistency. Several software tools can access and/or
modify partial areas of the factory data at different, and possibly overlapping, times. Therefore, the VFM
is required to ensure that concurrent accesses occur without endangering the data integrity and slowing
down the planning process to unacceptable levels.
Management of evolving Factory Data. The VFM has to provide functionalities for managing the
evolution and revision of the data related to complex entities like production systems, processes and
products. A typical VFDM object is made by several files, depending on the sub-ontologies it refers to,
as described in section 3. Hence, a coherent versioning mechanism must take into consideration the
inter-document references between sub-ontologies.
Data safety must be ensured in case of hardware failures or user errors.
Addition of customized functionalities. Third party developers need an appropriate mechanism to enrich
the set of functionalities provided by the VFM without impacting on its core.
Response time. The interaction between VFM and the VF modules requires the support of
communication mechanisms that are able to provide answers in an appropriate time frame.
A Semantic Web Endpoint which enables stakeholders to query virtual factory models with the required
level of granularity for a more efficient and selective data access.
Most of the above underlies the development of the previous version of VFM. For this reason the architecture of
the new semantic version shares similar features with its predecessor. However the ability to support validation
and queries of semantic data introduces novelty aspects in the overall design of VFM.
The architecture of the semantic VFM was designed to provide support to the required functionalities. Each
solution implemented by the VFM is based on stable and well established technologies in order to obtain an
overall system capable to respond to industrial needs of reliability. The resulting VFM architecture is shown in
Figure 2 as an UML component diagram.
The functionalities of the VFM are exposed as web services that have been identified as a suitable and widely
adopted solution to guarantee platform independent interfacing capabilities. The Application Server provides the
front end for the exposure of VFM functionalities and takes care for the information transport of the VFM.
The Information Exchanging Platform (IEP) is the main component of the VFM and provides VF modules and
plugins with a high level access to the two functional cores of VFM: the Versioning Layer and the Semantic
Layer. It represents the preferred way (even if not the only one) to connect to the VFM, since it provides a
complete set of methods for structured data retrieval and semantic validation, data locking mechanism and
factory version management.
The Versioning Layer contains the VF Data Repository where all the shared data are stored. The evolution of the
factory data is managed by the Versioning System that organizes and updates the set of virtual factory instances.
The Versioning System guarantees the data safety as well, since it allows restoring an older version at any time,
thus preventing data losses due to user errors. Moreover, rollback methods can be used in case of data
inconsistencies due to broken connections or other factors, always ensuring data safety. In particular, the locking
mechanism exposed through the IEP helps to manage the concurrent access of the VF modules.
Plugin Connector
Tomcat mod
Application server (Apache)
Web Service XYZ
Plugin XYZ
Server Application
Web Service XYZ
IEP Plugin
Data I/O
High Level Data I/O
Versioning System
VF Data Repository
Jena (RDF-api, OWL- API, SPARQL query engine
Factory Model Pool
Adaptation Module Module A Module B
High Level Data I/O
Semantic Layer
Versioning Layer
Storage (TDB or SDB)
Figure 2 - Semantic VFM Architecture
The Semantic Layer is implemented by embedding in VFM one of the most common and reliable Semantic Web
Frameworks: Jena (Jena, 2011). Through the IEP users can carry out semantic validations of VFF models using
Jena functionalities directly on the server. Thanks to Jena, the IEP can also provide a VFM SPARQL endpoint.
By starting a Query Session on data extracted from the VF Data Repository it is possible to perform SPARQL
queries (W3C, 2008a). Through queries each module (or plugin) can select and aggregate information and be fed
with exactly the data it needs for its business process. Model modifications can also be executed using the
SPARQL Update language (W3C, 2008a). Modified models can then be serialised in output files in the same
format used by the VF Data Model ontology (RDF/XML).
Currently many software applications, called VF Decoupled Modules, are under development and will interface
the VFM for accessing the information kept in the data model.
Since the VFM is the centre of the data exchange among modules, it has been conceived with openness in mind
to be able to deliver data to the large variety of tool involved in the factory planning process. The decoupled
modules range from completely new developments to integration of existing application, to off-the-shelf
commercial software, characterized by different operating systems and development languages, among them
Windows, Linux and Java, C++, Python.
Since the exposed functionality of the VFM is implemented as a web service (W3C, 2011), all the modules are
required to implement a web service client according to the WSDL file (Booth and Liu, 2007) describing the
published interface. Additionally, to address the issue that web services are intrinsically stateless, the VFM has
implemented a few specific functions to keep track of the state of its clients. Therefore the decoupled modules
need to actively support this mechanism. Finally, the data received from the VFM are in RDF/XML format
(Beckett, 2004) and the utilization of third party libraries for the handling of that format is essential.
The listed common requirements lead to a similar overall module architecture which foresees a few predefined
The following diagram illustrates the generic architecture of a VF decoupled module with the mentioned
components and a section of the VFM in the bottom part of the picture.
Figure 3 - Decoupled VF Module Architecture
5.2.1. VF Connector
Since all the decoupled modules will face common tasks related to the VFM connection, in order to avoid
repeated development efforts among the VFF partners, a specific VF Connector for the most common
development languages (C++, Java and Python) has been implemented.
The VF Connectors take care of the web service client implementation and the connection state mechanism.
5.2.2. RDF/XML Library
Each decoupled module will manage different parts of the data model in different ways. Nevertheless most of the
results coming from the VFM are in form of RDF/XML streams so that the development effort will be reduced
using already existing third party libraries conceived for RDF/XML data manipulation. The following table lists
some of the most used open source libraries.
Redland RDF Libraries
C with Python- Perl- PHP-
Ruby- Interfaces
Protegé API
Sesame OpenRDF
Table 1 - RDF/XML Libraries
Semantic data handling obviously is more complex than the one required for XSD/XML-based models. Most of
the VF Decoupled Modules are not semantic applications (Motta and Sabou, 2006). As such they access the
VFDM semantic representation only to extract and modify “plain” data. Indeed this is one of the few
disadvantages of the proposed semantic approach that can be mitigated only by fully exploiting the related
technology to ensure the full integration of the four Pillars of VFF.
5.2.3. Business Logic
This part of the software is peculiar to each module and will be developed on top of the mentioned components.
Nevertheless it is possible to distinguish two substantially different functionalities; the ad hoc developed
modules will provide a specific logic and expose it through a graphical user interface while the adaptation
modules will provide the required interface for a seamless integration of existing commercial tools.
The Semantic VF Manager has been implemented on the basis of the previous prototype presented in Sacco et al
(2011). Even if this version is a complete rewrite of the software, the main inspiring guidelines that have driven
the first prototype release have not been changed. The choice of development based on an open source and cross
platform architecture is still valid. This allows the deployment of the tools in real industrial scenarios where the
VF Manager should be integrated into existing legacy intranet architectures. Having chosen to adopt
technologies with proven reliability, cross platform compatibility and well known by IT personnel grants a
smooth integration and successful operation inside most of the existing network configuration.
We hereby describe, for each of the main components of the architecture shown in Figure 2, the prototypal
choices of the applied software tools. The Application Server was implemented with Apache HTTP Server, one
of the most deployed and reliable HTTP servers (The Apache Software Foundation, 2011a). The Servlet
Container was developed with Apache Tomcat (The Apache Software Foundation, 2011b), used in numerous
large-scale, mission-critical web applications across a wide range of industries and organizations. Tomcat is
paired with “Tomcat mod” to support integration with Apache HTTP server: this connector redirects the
information received by the Apache Server to Tomcat, and therefore to the plug-ins. The Versioning Layer was
developed by adopting Subversion (Collins-Sussman et al, 2004) that is an open source version control system
widely used in the open source world. Access to the ontology model, i.e. creating, writing and reading models
from OWL, has been implemented on top of the Jena framework (Jena, 2011) which is a proven library that
implements Web Semantic in Java.
Web Service
Web Pages
Web Pages
Jena Subversion Transaction Users
Projects Working Copies Ontology Models Logging
Figure 4 - Architecture of the VF Manager Prototype
The architecture of the VF Manager in Figure 4 highlights the internal division in tree main layers: front end,
managers and utilities.
The front end layer consists of three components that expose the functionalities:
IEP provides a SOAP Web Service to the VF modules.
Administration lets the administrative personnel manage the user and the opened sessions using a web-
based interface.
Users Pages are web pages that a user can access to see his open session and to chat with other active
The managers layer groups most of the business logic of the VF Manager and is composed of the following
Jena handles the ontology model using the Jena framework;
Subversion is responsible for data storage and versioning using the SVNKit library (SVNKit, 2011);
Transaction manages user sessions and commits, coordinating the Jena, Subversion and Users
Users manages the user database, access control and sessions list.
The utilities layer consists of components providing a common infrastructure to handle:
Working copies
Ontology models
All this components have been developed in Java and JSP (Java Server Pages) and deployed as Tomcat web
This prototype is an evolution of the software presented in Sacco et al (2011) and focuses on the implementation
of a wider set of features:
Data versioning
Web access
The previous prototype focused only on versioning and locking. A key feature introduced in this version is
granularity, implemented with the concept of projects to represents the minimal independent unit (i.e. set of files)
that can be locked. A project can have one or more dependency on other projects, allowing reusing of shared
resources. A project can not only declare its dependencies allowing a module to manually retrieve them - but
can also automatically include them to build a complete ontology model. This model can be queried using
SPARQL language (W3C, 2008a). Finally in this prototype we have implemented the ability to access the VF
Manager using an Internet browser: administrators and registered user can access some functionalities of the VF
Manager without needing to use a module application.
Further development of the prototype is targeted at consolidating the robustness of the implemented features and
at improving the functionalities that can be accessed from the web interface. This last development will
eventually enable a user to access the data in the VF Manager without relying to any module, maybe allowing
limited editing capabilities. Another important feature that will be enabled by the web interface is the access
from mobile devices, such as smart phones and tablets.
A VFDM ontology-based model (for brevity VF model) is defined by a number of files organised in folders.
Each folder contains the following files:
one (and only one) file with extension .owl containing a Factory Project or a Factory Library.
binary files (BLOBs) which are instrumental for the correct representation of the model (graphic files,
process data, etc.).
A Factory Library contains “reusable” individuals that can be imported by a Factory Project. For what stated
above Projects can depend on a number of Libraries and no Project depends on other Projects. However,
Libraries can depend on other Libraries. If we see this mutual dependency of individuals from both classes
organised into a hierarchical structure, a typical VF model could be represented as in Figure 5:
Library A
Library C
Library E
Library F
Library B
Library D
Figure 5 Typical hierarchical structure of a VF model
The organisation of a Project and its Libraries into a hierarchical tree can be exploited to provide an efficient
versioning mechanism that promotes concurrent development of VF models. Indeed after the first release of a
Library, the development of a Project (or other Libraries depending on it) can be carried out separately and in
parallel with further developments of the base Library. For example, taking as a reference the structure of Figure
5, once the first version of Library F has been released, Library E can be developed meanwhile other individuals
are added to Library F. In this way the development of both Libraries could be carried out in parallel by different
users. Concurrent development is a crucial aspect of VFF, since a consistent number of stakeholders can
contribute to the development of complex factory models. On the other hand the safe concurrent access strategy
of VFM requires for the user to reserve the right of modifying model data through Transactions initiated on
Project and Library files. Transactions determine a Subversion lock on files preventing other users to perform
different, uncontrolled modifications at the same time. Transactions however can be applied by the user (further
than the Project) to specific Libraries only, leaving the possibility to other stakeholders to write-access the
unlocked ones.
As a Project grows in complexity, the number of Libraries which it depends on could be quite significant and the
corresponding hierarchical structure quite complex. A user can start a Transaction directly on a Library that
he/she needs to modify for further developments of the depending Project. This happens, for example, when new
machines of a Resource Catalogue (the Library) are modelled side-by-side with a factory layout (the Project).
However, in doing so the user might not be aware that other Libraries depending on the one that he/she is going
to modify could also be affected. With reference to Figure 5, if the user needs to modify Library D for the
purpose of enhancing the Project model, Library A might also need to reflect changes in D. Therefore a
Transaction on Library A is also required. It is then crucial for the VFM to guarantee a harmonised versioning of
Project and Libraries related to each other according to their hierarchical tree. The versioning mechanism must
preserve the consistency of version dependencies within the same Project. The Semantic VFM implements this
mechanism through a hierarchical lock and commit procedure. This technique makes sure that any change in a
low-level library is accounted for in higher hierarchical depending sub-models (the Project and/or other
Libraries), so that any possible inconsistent association of Project and Libraries versions is prevented. The
hierarchical lock and commit is implemented in the Transaction Manager (Figure 4) and consists in executing
Subversion locks and commits recursively along hierarchical paths in the Project tree structure, in ascending and
descending order respectively. For example, in Figure 5 a Transaction on Library E would also trigger a
Transaction on Library B and the Project, but not on Library F. The lock path therefore goes from Library E to
Project. The hierarchical commit works in the opposite way. Closing the Transaction on the Project would end
the Transaction on Library B and E as well, since all modelling business is supposed to have come to an end.
However, the user can decide that he/she does not need to modify the Libraries any longer but still needs to work
on the Project. In this case by closing only the Transaction on Library B he/she closes the one on Library E as
well, leaving the Transaction on the Project open.
The hierarchical lock and commit process described here provides an efficient way for versioning groups of
interrelated data within a single IEP call. However it is still not enough to ensure a consistent cross-versioning of
VF models, since an unlocked Library can be modified outside a checked-out Project at the same time by a
different stakeholder.
Each version of a Project can in theory depend on all possible versions of Libraries which it depends on. To have
a more deterministic system and reduce the risk of model inconsistencies, a cross-versioning strategy has been
applied, which takes into account the practical aspects of the day-by-day design and modelling work of virtual
factories. VFM leaves stakeholders in power of deciding if and (above all) when they want to commit a Project
to newer versions of Libraries than those corresponding to the Head version of the Project itself. The procedure
can be performed through a dedicated GUI on a user VFM web page. The Version References page is accessible
from the home page of VFM through the link Project Management (Figure 6).
A Project designer can only update the dependencies to newer version of Libraries. The opposite (make a Head
version of a Project depending on older version of Libraries than the current ones) is considered inconsistent
with the evolution of a VF Model file system (Transactions can only be initiated on Head versions of projects),
therefore not allowed (lower version numbers are “greyed out” and not selectable). Stakeholders could still
“recombine” Project and Libraries with versions of their like by branching. However within the same branched
Project lifeline the principle of “same or greater Libraries versions” dependency still applies.
Figure 6 - Cross-version management on the VFM Web GUI
This section describes two decoupled modules that can interact with each other thanks to the VFM. The former is
named Factory Layout Planner (FLP) (Ceruti et al, 2010), whereas the latter GIOVE Virtual Factory (Viganò et
al, 2009).
8.1 FLP
8.1.1 Functionality
FLP is a client/server application that enables the collaborative development of a factory layout.
The main functionality of the FLP consists in:
3D visual editing of the layout
3D visual editing of the building
running Discrete Events Simulation (DES)
Figure 7 - FLP Main Window
The application is characterized by a two-level architecture with a fat client (Figure 7) dealing with complex 3D
models and real time requirements, and a server which acts as a synchronization manager and as VFM web
8.1.2. Implementation
FLP is an application written in Java. For handling the data received from the VFM in RDF/XML format, the
third party library Jena (Jena, 2011) is used.
8.1.3. Accessed Data
FLP will interact with different areas of the VFDM to exploit its functionality:
the multisite information - the production plants of the enterprise [read-write]
the building data [read-write]
the resources templates - for every resource template (or type) a set of information (icon, VRLM
file(s) for 3D representation, properties and further data) [read-only]
the layout data (which contains for example the instantiated resources, their position and their property
values) [read-write]
the production plans and processes data to feed the DES engine [read-only]
the results of the DES simulation [read-write]
8.2.1 Functionality
GIOVE Virtual Factory (GIOVE-VF) is a virtual reality collaborative environment aimed at supporting the
factory layout design. In particular, GIOVE-VF offer 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 8). The virtual environment can schematically display performance measures that are provided
by simulators and/or monitoring tools. Thanks to its user-friendly interface, GIOVE-VF enables the
collaboration between managers, experts and also workers in an intuitive way.
Figure 8 - GIOVE-VF environment
8.2.2. Implementation
GIOVE-VF is a software tool developed onto the C++ library GIOVE (Graphics and Interaction for OpenGL-
based Virtual Environments), that is a set of libraries and tools designed for the creation of collaborative virtual
environments and the realization of real-time 3D interactive scenes including working digital representations
of objects, products, systems that are being designed and evaluated (Viganò et al, 2011). GIOVE-VF
communicates with the VFM thanks to a C++ VF Connector that has been developed by exploiting the Redland
RDF Libraries.
8.2.3. Accessed Data
GIOVE-VF can access data stored in the VF Data Repository that are related to the definition of:
representation contexts;
production site and buildings;
placement of objects inside a building;
production resources;
part types and process plans;
evaluated performance of production systems.
The prototype of the VFM hosts the VFDM and a test case that can be accessed by all VF modules to check the
correctness of input/output communications. This test case defines data related to the layout of a factory,
importing both resource types and instantiated resources.
The following subsections describe two use cases showing how FLP can manage the data defined in the test
case. These steps would be similar also for a different VF module (e.g. GIOVE-VF).
8.3.1. Caching resources types (read only)
The FLP composes a layout by creating and placing into the 3D window instances of resources from the
resources catalogue. The FLP itself does not modify the resource templates, which consists of resource type
description data and typically very large 3D model files (those data are not subject to frequent changes). Given
those assumptions, for performance purposes, the FLP maintains a local cache of the data by checking from time
to time if a synchronization of the local cache is required.
The steps to accomplish this task consist in
select the current revision of the sub-project Resource Library”
download the ontology file(s) containing the individuals
load the file(s) into a Jena ontology model in order to access the content
retrieve all the 3D files from the “resource catalogue” folder and store them locally
8.3.2. Layout planning (read-write)
The FLP user selects the current version of a layout with the purpose of modifying it. It must signal to the VFM
its intention, so that other users do not modify the same data and the Data Model remains consistent.
Precondition for this use case is that the local catalogue of resources has been successfully synchronized for the
selected project.
The additional steps (compared with use case 7.3.1.) required to accomplish this task consist in:
start a transaction on the selected project
display all the instances of resources defined in the project in a editable 3D view of the layout
apply the modifications to the Jena ontology model and modify the local project file(s)
send back the modified files to the VFM
make the changes permanent and available to all others VFF users by committing the open transaction
8.3.3. Discrete Event Simulation
The FLP provides a set of functionalities developed to perform a discrete-event simulation (DES) through which
the user can assess the productivity related to a designed layout and compare it against a target value. The FLP
models through a sequence of discrete events, the machines operations and combine them according to a pre-
defined production plan. An external PLM, acting as VF module, provides the production plan basing on the
designed factory layout and makes it available to the other VF modules via the VFM. The output of the
simulation enables both the FLP user and other interested VF modules, to analyse the calculated KPIs (i.e. the
production annual volume, the time delays, the machine utilization, etc.) and compare them with the target
values in order to assess if the customer requirements are fulfilled.
This use case want to remark the ability of VF system to guarantee the data interoperability among the involved
VF modules.
The new approach driven by the Semantic VFDM and enabled by the Semantic VFM represents a step forward
in the improvement of the Virtual Factory Framework and in particular towards the target of a fully integrated
architecture of all the Pillars. Data individuals and their Semantics come now from the same coherent source and
can be seen by different perspectives, according to the needs of the accessing clients (Knowledge Manager or
Decoupled Modules).
A first prototype of the Semantic VFM exploring the potentials and the issues related to this new approach has
been presented. In particular, the applied Semantic Web technologies represent the cornerstone to obtain a
framework where the different stakeholders can effectively contribute in a harmonized way to the definition of
the virtual factory along all the phases of its lifecycle.
In the coming months improved versions of the VFM will implement the defined architecture and fulfil the
expected functionality. An increasing number of Decoupled Modules will interface the VFM and fill in the VF
Data Repository with individuals according to the developed Virtual Factory Data Model, thus validating the
new approach.
The research reported in this paper has received funding from the European Union Seventh Framework
Programme (FP7/2007-2013) under grant agreement No: NMP2 2010-228595, Virtual Factory Framework
Beckett, D., “RDF/XML Syntax Specification (Revised)”, W3C, 2004, Retrieved: 15.06.2011, <
Booth, D. and Liu, C.K., “Web Services Description Language (WSDL) Version 2.0 Part 0: Primer”, W3C, 2007,
Retrieved: 15.06.2011, <>
Bracht, U., Masurat, T., "The digital factory between vision and reality", Computers in Industry, Volume 56, Issue 4, 2005,
pp. 325333
Carroll, J.J., Dickinson, I., Dollin, C., Seaborne, A., Wilkinson, K. and Reynolds, D., “Jena: Implementing the Semantic
Web Recommendations”, Proceedings of the 13th international World Wide Web conference, 2003, pp 74-83
Ceruti, I.F., Dal Maso, G., Ghielmini, G., Pedrazzoli, P. and Rovere, D., “Factory Layout Planner”, ICE - 16th International
Conference on Concurrent Enterprising, 2010, Lugano, Switzerland
Chryssolouris, G., Mavrikios, D., Papakostas, N., Mourtzis, D., Michalos, G., Georgoulias, K., "Digital manufacturing:
history, perspectives, and outlook", Proceedings of the Institution of Mechanical Engineers Part B: Journal of
Engineering Manufacture, Volume 223, No. 5, 2008, pp.451-462
Colledani, M., Terkaj, W., Tolio, T. and Tomasella, M. “Development of a Conceptual Reference Framework to manage
manufacturing knowledge related to Products, Processes and Production Systems”, In: Bernard A, Tichkiewitch S (eds),
Methods and Tools for Effective Knowledge Life-Cycle-Management”, Springer, 2008, pp 259-284.
Collins-Sussman, B., Fitzpatrick, B.W. and Pilato, C.M., “Version Control with Subversion”, 1st Edition, O'Reilly Media,
Sebastopol, CA, 2004, p 320
Colombetti, M., “Ingegneria della conoscenza: modelli semantici”, 2010-11 Edition, Facoltà di ingegneria
dell’informazione, Politecnico di Milano, Italy, 2011, p 42
Huang, G.Q., Simpson, T.W. Pine II, B.J., “The power of product platforms in mass customization”, International Journal
of Mass Customisation, Vol. 1, No. 1, 2005, pp 1-13
ISA-95, “ISA-95: the international standard for the integration of enterprise and control systems”,, Retrieved:
15.06.2011, <>
Jena, “Jena A Semantic Web Framework for Java”,, 2011, Retrieved: 15.06.2011,
IFC2x3, “IFC2x3 Release”, buildingSmart, 2006, Retrieved: 15.06.2011, <
Mc Bride, B., An Introduction to RDF and the Jena RDF API”, 2010, Retrieved: 15.06.2011,
Mc Carthy, P., Search RDF data with SPARQL”, IBM, developerWorks, 2005, Retrieved: 15.06.2011,
Motta, E. and Sabou, M., Next Generation Semantic Web Applications”, 1st Asian Semantic Web Conference (ASWC),
2006, Beijing, China
Sacco, M., Dal Maso, G., Milella, F., Pedrazzoli, P., Rovere, D. and Terkaj, W., “Virtual Factory Manager”,
HCI International, 2011, Orlando, USA
Sacco, M., Pedrazzoli, and Terkaj, W., “VFF: Virtual Factory Framework”, ICE - 16th International Conference on
Concurrent Enterprising, 2010, Lugano, Switzerland
SVNKit, “[Sub]Versioning for Java”, TMate Software, 2011, Retrieved: 15.06.2011, <>
The Apache Software Foundation, “Apache HTTP Server Project”, The Apache Software Foundation, 2011, Retrieved:
15.06.2011, <>
The Apache Software Foundation, “Apache Tomcat 6.0”, The Apache Software Foundation, 2011, Retrieved: 15.06.2011,
Terkaj, W., Tolio, T., Valente, A., “Designing Manufacturing Flexibility in Dynamic Production Contexts”, In: Tolio, T.
(ed) Design of Flexible Production Systems. Springer, 2009, pp 1-18
Tolio, T., Ceglarek, D., ElMaraghy, H.A., Fischer, A., Hu, S., Laperrière, L., Newman, S., Váncza, J., “SPECIES -- Co-
evolution of Products, Processes and Production Systems”, Cirp Annals-Manufacturing Technology 59 (2), 2010, pp
Valente, A., Carpanzano, E., Nassehi, A. and Newman, S. T., “A STEP compliant knowledge based schema to support
shop-floor adaptive automation in dynamic manufacturing environments”, Cirp Annals-Manufacturing Technology 59
(1), 2010, pp 441-444
Viganò, G.P., Greci, L., Mottura, S., Sacco, M. "GIOVE Virtual Factory: A New Viewer for a More Immersive Role of the
User During Factory Design". In: Canetta, L., Redaelli, C., Flores, M. (Eds.) "Digital Factory for Human-oriented
Production Systems ". Springer, 2011, pp 201-216
Viganò, G.P., Greci, L., Sacco, M., “GIOVE Virtual Factory: the digital factory for human oriented pr oduction systems”,
Proceedings of the 3rd International CARV Conference, Munich, Germany, 2009, pp. 748-757
W3C, “OWL Web Ontology Language - Use Cases and Requirements”, W3C, 2004a, Retrieved: 15.06.2011,
W3C, OWL Web Ontology Language - Reference”, W3C, 2004b, Retrieved: 15.06.2011, <
W3C, “SPARQL Query Language for RDF”, W3C, 2008a, Retrieved: 15.06.2011, <
W3C, “SPARQL Update”, W3C, 2008b, Retrieved: 15.06.2011, < >
W3C, “Web Services Activity”, W3C, 2011, Retrieved: 15.06.2011, <>
W3C, “XML Schema Part 1: Structures Second Edition”, W3C, 2004c, Retrieved: 15.06.2011,
“ISO 14649-10:2004 Industrial automation systems and integration -- Physical device control -- Data model for
computerized numerical controllers -- Part 10: General process data
... It should provide a virtual representation of buildings, resources, processes, and products. The entire factory should be simulated as a continuous and coherent digital model that can be used throughout the entire product lifecycle -from the product idea to the final disassembly of production plants and buildings [4]. ...
Full-text available
The Augmented Reality (AR) can be used to simulate and improve production processes in a virtual manufacturing system before their physical implementation. These techniques make it possible to connect the real world with the virtual one in real time and in a three-dimensional environment. Their effective application to the implementation of the product development phases would significantly limit its subsequent reworking and modifications. The difficulty, however, is to design and integrate AR with virtual production systems. The paper presents a proposal for a technical procedure for the implementation of the integration of a commercial PLM and a proprietary AR system.
... It should provide a virtual representation of buildings, resources, processes and products. The entire factory should be simulated as a continuous and coherent digital model that can be used throughout the entire product lifecycle -from the product idea to the final disassembly of production plants and buildings [4]. ...
Full-text available
Nowadays, many industrial companies, including small and medium-sized ones, are adopting virtual manufacturing concepts to face global competition and major manufacturing challenges. The aim is to improve quality, shorten delivery time and reduce costs. However, most virtual manufacturing methodologies, tools and software are not integrated well enough to perform the required activities efficiently. Attention is usually focused on local and specific proficiency, thus jeopardising information exchange between departments, parallelism of work and communication along the product lifecycle. In these circumstances, the use of virtual production and digital representation of the production system and its processes becomes even more important to optimise production activities. Manufacturing industries are evolving towards digitisation, networking and globalisation. In the process of rapidly evolving information technology, digital tools and systems are being used in all industries, managing a variety of tasks throughout the product lifecycle with the use of PLM (Product Lifecycle Management) applications. PLM class systems integrate a set of applications supporting product development.
... From a different perspective a virtual factory not only provides an integrated platform for manufacturing systems design and analysis cross different organisations (Tolio, Sacco, Terkaj, & Urgo, 2013), but it can also be an integrated simulation model of the major subsystems in a factory (Ghielmini, G. et al., 2013). The model considers the factory as a whole and provides an advanced decision support capability (Sanjay, Ngai Fong, Khin Maung, & Ming, 2001). ...
... From a different perspective a virtual factory not only provides an integrated platform for manufacturing systems design and analysis cross different organisations (Tolio, Sacco, Terkaj, & Urgo, 2013), but it can also be an integrated simulation model of the major subsystems in a factory (Ghielmini, G. et al., 2013). The model considers the factory as a whole and provides an advanced decision support capability (Sanjay, Ngai Fong, Khin Maung, & Ming, 2001). ...
... En el ambiente se aplica el código de colores definido por el estándar ISO 14726 y ANSI/ASME A13.1 Los diferentes tipos de tubería se ilustran en la Figura 18, en donde se encuentran tubos, codos y elementos en T. Figura 18.Tubería Unity 3D® no acepta elementos directos de SolidWorks®, por este motivo con Blender® se pasa de formato VRLM (lenguaje de modelado de realidad virtual) a 3DS. Este formato crea el elemento dentro de una estructura binaria con identificadores que la plataforma virtual es capaz de interpretar [65]. Elementos como las bombas, motores, condensadores, válvulas entre otros se exportan desde GrabCAD®. ...
Full-text available
Crude refining plants have high energy consumption, particularly in the distillation process. Various authors state that they are the most energy-consuming systems in the petrochemical industry. Currently, work is being carried out on the implementation of controllers that use minimization techniques and digital tools in search of improving the efficiency of the distillation process and the quality of the product. This work presents the design and development of a virtual 3D laboratory for the oil refining process. Initially, the mathematical descriptions of the threads are obtained, together with the computer aided designs, which will be controlled using multivariable control techniques. As a result, the development of the laboratory is shown with the implementation of adaptive and predictive control techniques that guarantee the regulation of temperature and the required concentrations. The main contribution of this work is the interaction that a user can have with this type of process and the possibility of implementing control strategies for itself, which allows it to be used as support in training activities. Las plantas de refinamiento de crudo poseen un consumo de energía elevado particularmente en el proceso de destilación. Diversos autores manifiestan que son los sistemas de más consumo energético en la industria petroquímica. En la actualidad se está trabajando en la implementación de controladores que usan técnicas de minimización y herramientas digitales en busca de mejorar la eficiencia del proceso de destilación y la calidad del producto. En este trabajo se presenta el diseño y desarrollo de un laboratorio virtual 3D, para el proceso de refinado del petróleo. Inicialmente se obtienen las descripciones matemáticas de los subprocesos, junto con los diseños asistidos por computador, los cuales van a ser controlados utilizando técnicas de control multivariable. Como resultado se muestra el desarrollo del laboratorio con la implementación de técnicas de control adaptativo y predictivo que garantizan la regulación de temperatura y las concentraciones requeridas. El principal aporte de este trabajo es la interacción que puede tener un usuario con este tipo de procesos y la posibilidad de implementar estrategias de control para el mismo lo cual permite utilizarlo como apoyo en actividades de formación.
... The IIoT will enable the development of a virtual factory that provides a holistic, scalable, and virtual representation of a manufacturing facility as well as synchronization, dynamic configuration and thus, cost savings of manufacturing facilities [Ghi13]. And the implementation of IIoT in the manufacturing sector will result in a significant reduction in quality inspection, surveillance, and operational expenditures [Fro16a]. ...
Full-text available
The Industrial Internet of Things (IIoT) is the application of emerging Internet of Things (IoT) technologies within industrial domains to further enhance productivity. This worldwide paradigm shift, the digital transformation, can be observed across the entire value chain and spans multiple verticals, such as energy, health care, manufacturing, smart cities, and transportation. Each domain developed highly optimized silos in the past, based on Operational Technology (OT), and will be interconnected with others using Information and Communication Technology (ICT). This chapter provides an overview of current concepts to bring the OT and the ICT worlds together. More specifically, a number of challenges and technologies in this context are described and relevant initiatives are presented. Finally, the outstanding field of application “Industry 4.0” is being highlighted in more detail.
Full-text available
Simulation solutions for manufacturing, as one of the key Industry 4.0 technologies, has received much attention in the literature. In this paper, we show how different simulation solutions have been discussed especially concerning small and medium-sized enterprises (SMEs). This review was performed by systematically searching out literature in the Scopus database. The application areas of simulation solutions in SMEs, simulations tools, and the challenges of adopting these technologies are addressed in this review. Further studies on developing solutions targeted towards SMEs are still needed.
Full-text available
The data interoperation between VF (virtual factory) platform and MES (Manufacturing Execution System) plays an important role in intelligent factory construction. The study focuses on the integration strategy between the VF and the MES by incorporating VF manufacturing assets in two ways, i.e., vertical integration (used for production line performance evaluation) and the horizontal integration (cloud manufacturing based on manufacturing assets services discovery and their composition). The VF platform which integrates the manufacturing assets in two manners is designed as the bottom layer in the entire integration framework. It has been applied to build a four tiers integration model in an intelligent production system construction of a domestic ship manufacturer and verified its feasibility and availability.
The core problem of the application of MES (Manufacturing Execution System) in intelligent manufacturing systems is integration, which solves the problem of the data interoperation between the distributed manufacturing systems. The previous researches on MES integration rarely considered the problem of system data security access. A three-level data security access mechanism based on the independence of the system administrators, security administrators, and security auditors is proposed which integrated into the MES integration framework to guarantee the business and engineering data security access for the related distributed clients. The principle is using the domain to make the logical isolation for different clients and data sources and applying the pre-defined data sharing rules for safe access. In the proposed MES integration framework model, the data interoperation between MES and the engineering software systems is discussed which includes ERP (Enterprise Resource Management), CAPP (Computer Aided Process Planning), DNC (Distribution Numerical Control), WMS (Warehouse Management System), and SCADA (Supervisory Control and Data Acquisition), etc., the implementation method of personalized data display GUI is discussed as well. The study is based on the KMMES developed by Wuhan KM-Software of China, and it has been deployed in over forty companies from the sections of aerospace, automotive, shipbuilding and other industries.
Full-text available
This paper serves as an extended editorial for this inaugural issue of the International Journal of Mass Customisation. Its main objective is to give a brief overview of the recent developments in Mass Customisation (MC) research and practices, although more comprehensive reviews should be referred to existing review articles (Silveira et al., 2001; Simpson, 2005). For this new journal, more position papers have been planned to appear. Materials are mainly drawn from well-cited monographs, conference proceedings and articles published on various MC topics. This new journal aims to serve the MC community as a major forum to exchange substantial ideas and share experiences, complementing newsletters and resources maintained at various websites (e.g.,;;
Full-text available
Manufacturing Flexibility is seen as the main answer for surviving in markets characterized by frequent volume changes and evolutions of the technological requirements of products. However, the competitiveness of a firm can be strongly affected by capital intensive investments in system flexibility. This chapter presents an approach to design new manufacturing system architectures endowed with the right level of flexibility required by the specific production problem. These systems are named Focused Flexibility Manufacturing Systems (FFMSs). The key idea consists in tuning system flexibility on the production problem to cope with uncertainty related to the evolution of product demand. The significance of this topic and its potential impact on the industrial sector in the medium-long run is testified by the interest shown by companies making initial efforts in this field.
Full-text available
The Manufuture strategic research agenda identifies two fronts of intense and growing competitive pressure for European manufacturing: in the high-tech sector, we face other developed countries and, on the other hand, in more traditional sectors, low-wage countries pose a serious threat. One of the response envisioned and strongly promoted is the development of ground-breaking Information and Communication Technologies meant to support new approaches to industrial engineering. Clearly drawing inspiration from this vision, the tool presented in this paper, the Factory Layout Planning, aims to be such kind of technology, enabling the multi-site, multi-level distributed design of factory layout and performance simulation. This paper gives an overview on the main features, in terms of concrete software characteristics of this tool, that can be considered one of the cornerstones for the Next Generation Factory.
Conference Paper
The Manufuture strategic research agenda identifies two fronts of intense and growing competitive pressure for European manufacturing: in the high-tech sector, we face other developed countries and, on the other hand, in more traditional sectors, low-wage countries pose a serious threat. One of the response envisioned and strongly promoted is the development of ground-breaking Information and Communication Technologies meant to support new approaches to industrial engineering. Clearly drawing inspiration from this vision, the tool presented in this paper, the Factory Layout Planning, aims to be such kind of technology, enabling the multi-site, multi-level distributed design of factory layout and performance simulation. This paper gives an overview on the main features, in terms of concrete software characteristics of this tool, that can be considered one of the cornerstones for the Next Generation Factory.
The need to solve a wide range of complex tasks urged companies to adopt new software tools, including Virtual Reality (VR) and Augmented Reality (AR), in their production chains. Major software houses released a lot of tools to help experts to manage different complex tasks, included the design or the modification of the layout of a production plant. Unfortunately, such tools are far from being integrated in a unique tool. Furthermore, the small-to-medium enterprises prefer a more customized and less expensive solution (Consoni et al. J Intell Manuf 17(6):725–735, 2006). In this context, GIOVE virtual factory (GIOVE VF), a tool for the collaborative design and review of a factory layout, was developed by ITIA-CNR (Istituto di Tecnologie Industriali e Automazione, Consiglio Nazionale delle Ricerche, Italy) and integrated with the other DiFac components of the Factory Constructor (Sacco et al. Human computer interaction conference, Beijing (PRC), 25–27 July 2007; Mottura et al. 57th CIRP general assembly, Dresden, Germany, 19–25 August 2007; Smparounis et al. 14th international conference on concurrent enterprising, Lisbon, Portugal, 23–25 June 2008; Dürr et al. 14th international conference on concurrent enterprising, Lisbon, Portugal, 23–25 June 2008; Constantinescu et al. 14th international conference on concurrent enterprising, Lisbon, Portugal, 23–25 June 2008).