Ontology-based Representation of Time Dependent Uncertainty
Information for Parametric Product Data Models
Maximilian Zocholl and Reiner Anderl
Department of Computer Integrated Design (DiK), Technical University Darmstadt,
Otto-Berndt-Str.2, Darmstadt, Germany
Keywords: Uncertainty, Ontology, CAD, Parametric Product Data Model, PLM.
Abstract: The lack of information about uncertain conditions in manufacturing and the behaviour of load carrying
systems still lead to fatal design decisions. Semantic technologies provide the necessary capabilities to link
information from different domains along the product lifecycle and enable engineers to cope with
uncertainty. This paper presents an overview of existing literature in the fields of semantics in CAD and
PLM. We identify future research challenges and present our concept for the integration of uncertainty
information in parametric CAD models.
1 INTRODUCTION
A shorter time to market, an increasing degree of
product customization and massive cost pressure call
for the exploitation of all knowledge available.
Semantic technologies and especially ontologies can
support the federation of domain specific
information in order to make existing knowledge
accessible.
Increasing amounts of sensor data from the
manufacturing and usage phase of load carrying
systems such as landing gears reveal the non-
deterministic nature of single parameters and their
interaction in large systems. Despite of a shared
understanding for the impact of over- and under
sizing components Engineers still work with
nominal dimensions, fixed margins of safety and
tolerance specifications. Uncertainty Data cannot be
processed in current CAD-kernels.
The approach presented in this paper introduces
the possibility to apply semantic technologies for the
representation of processable uncertainty
information in the parametric product model. This
allows Engineers in the design process a better
understanding of the actual conditions during
manufacturing and usage. Hence, product
development converges to real product lifecycle
data.
2 STATE OF THE ART
2.1 Related Literature
Literature in Knowledge-Based Engineering, PLM
and domain specific approaches in Computer Aided
Engineering discuss possible future applications for
Ontologies. In the field of CAD, Ontologies are
often used as exchange format for CAD features
between different CAD Systems (Altidor, et al.,
2011), (Ramya T. Chaparala, 2013), (Ding, 2010),
(Lee, Cheon, & Han, 2005), (Tessier & Wang,
2013), (Abdul-Ghafour, Ghodous, & Shariat, 2012),
(Song & Han, 2010). Zhong et al. investigate the
exchange of semantic assembly information between
computer-aided tolerancing systems (Zhong, Qin,
Huang, Lu, & Chang, 2014). Andersen et al. propose
a domain-specific Ontology for boundary
representation models complying ISO 10303-42
(Andersen & Vasilakis, 2007).
Other approaches integrate CAD-systems with
different IT-systems along the product lifecycle.
Dartigues hands on knowledge stored in CAD data
for process planning (Dartigues, Ghodous,
Gruninger, Pallez, & Sriram, 2007). Young et al.
share knowledge for decision making in the
manufacturing process (Young, Gunendran, Cutting-
Decelle, & Gruninger, 2007). Alani generates
geometric models out of requirement templates
(Alani, 2007). Related to this topic, Kuhn presents
400
Zocholl M. and Anderl R..
Ontology-based Representation of Time Dependent Uncertainty Information for Parametric Product Data Models.
DOI: 10.5220/0005158004000404
In Proceedings of the International Conference on Knowledge Management and Information Sharing (KMIS-2014), pages 400-404
ISBN: 978-989-758-050-5
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
an approach for updating existing templates (Kuhn,
Dusch, Ghodous, & Collet, 2012). Assembly process
generation based on assembly information is
proposed by Zhu (Zhu, Wu, & Fan, 2010).
In contrast, PLM oriented literature supports the
interaction of more than two different systems along
the product lifecycle. General information models
are proposed by Sudarsan and Matsokis (Sudarsan,
Fenves, Sriram, & Wang, 2005), (Matsokis &
Kiritsis, 2011). Evangelou and Karacapilidis develop
an Ontology for multicriteria collaborative decision
making (Evangelou & Karacapilidis, 2005). Since
different domains often use several ontologies, Zhan
presents a mapping methodology in the context of
PLM (Zhan, Jayaram, Kim, & Zhu, 2010). Franke
investigates the automatic generation of ontologies
and a subsequent design rule checking process
(Franke, Klein, Schröder, & Thoben, 2011). Anderl
et al. distinguish the representation, the presentation
and the visualization along the product life cycle
(Anderl, Maurer, Rollmann, & Sprenger, 2013).
Here, the uncertainty information is attached to the
topological elements of the B-rep model.
2.2 Research Challenges
For future research, different challenges can be
identified. In the context of Knowledge-Based
Engineering, Verhagen claims a stronger
standardisation for the development of new
methodologies and a shared framework for the
assertion of solutions in order to facilitate the
exchange and the re-use of research findings
(Verhagen, Bermell-Garcia, van Dijk, & Curran,
2012). Looking more detailed into the existing
literature about CAD- and PLM-Ontologies,
Verhagens observations can be confirmed and
extended as follows.
Extending Ontologies by the representation of
parameters seems a promising approach for the
integration of both B-Rep and Feature-models.
Taking into account current standards can be helpful
for the transfer of research findings into industry.
For CAD applications the upcoming ISO 10303-242
allows new perspectives with its capabilities of
representing parameters, features, modelling history
and semantic Product and Manufacturing
Information (PMI). In the Area of PLM the Product
Lifecycle Support library (PLCSlib) supports
implementations of ISO 10303-239 with a semantic
model for the terms in use.
Attaching Uncertainty Information to parameters
represents the next step in the Collaborative
Research Centre (SFB) 805 in order to make more
information from the manufacturing- and usage-
phase exploitable in the development-phase.
Especially the connection of time dependent
information to parameters allows the avoidance of
unpredictable interactions within assemblies.
The automatable acquisition of information
along the Product Lifecycle for the successive re-use
is maybe the biggest challenge in semantic research
today. Extracting and storing multiple parameters for
one feature for the instantiation of individuals
implies the integration of different applications and
data repositories along the Product Lifecycle.
3 METHODOLOGY
Connecting time dependent information about
uncertainty to the product model requires the
interaction of different technologies. For a common
understanding, all relevant notions and technologies
are resumed briefly.
To these belongs the notion of uncertainty, the
principals for semantic representation and the
parametric product model description.
3.1 Uncertainty in Load Carrying
Structures
Building up on existing definitions of epistemic and
aleatoric uncertainty, the SFB 805 established three
categories of uncertainty. With respect to the quality
and the quantity of all relevant information,
“unknown uncertainty”, “estimated uncertainty” and
“stochastic uncertainty” can be distinguished.
Figure 1: Uncertainty model, compare to (Engelhardt, et
al., 2010).
In case of “unknown uncertainty”, the considered
product- or process properties are not available or
not trustful. For “estimated uncertainty” properties
and interdependencies can be described by intervals,
Ontology-basedRepresentationofTimeDependentUncertaintyInformationforParametricProductDataModels
401
tolerances or nominal values. “Stochastic
uncertainty” implies a low level of uncertainty so
that properties and their interdependencies can be
described by frequency distributions (Engelhardt, et
al., 2010). One possible representation of
uncertainties for data exchange is proposed by
Sprenger (Sprenger, 2013).
3.2 Semantic Representation
As in the semantic web, Resource Description
Framework (RDF) can be used as data model for
web infrastructures between different applications
and knowledge bases. RDF uses a triple syntax of
subject, predicate and object to formalize and
describe relations between resources, such as
information or documents.
All resources can be addressed by a namespace
dependent Unique Resource Identifier (URI). The
Web Ontology Language OWL 2 builds on the
capabilities of RDF and extends them. Restrictions
like “disjoint with” or “same as” can be expressed
by set operations which allow the inference of
implicit information, consistency checking and
classification.
Unlike in RDF, individuals and classes are
supposed to be disjoint in Ontologies. While the
terminological box (t-box) contains the description
of concepts in a domain, the assertional box (a-box)
contains individuals and information about them. In
this context individuals are specific CAD models
such as assemblies, parts, their composing elements,
as well as parameters which are related by
constraints.
3.3 Parametric Product Description
Parametric product descriptions are used in
parametric CAD systems for hybrid CAD models as
well as for generative and accumulative CAD
models. Following Anderl and Mendgen, parameters
are connected by constraints to each other in
parametric product descriptions. Parameters can
descend from different domains such as geometry,
material or technology as shown in Figure 2.
Constraints can be differentiated in geometric
and engineering constraints. Geometric constraints
like parallel or horizontal connect geometrical
parameters to each other. Engineering constraints
define relations between geometrical and non-
geometrical parameters. These constraints are
functional or logical and can also impact on the
topology.
All types of constraints are represented by
equations or predicates. Equations are summarized
in explicit or implicit equation systems. Solving
procedures are sequential for explicit equation
systems and simultaneous for implicit equation
systems. In order to deduce at least one possible
solution, the equation system needs to be
determined.
Figure 2: Parametric product description, compare to
(Anderl & Mendgen, 1996).
When Constraints are expressed with predicates
rule-based solving approaches can be used (Anderl
& Mendgen, 1996).
In case of uncertainty, the equation system is
underdetermined due to missing parameters or it is
overdetermined due to scattered values of distributed
parameters. Thus, neither simultaneous nor
sequential solving methods can be used. However,
rule-based approaches are not restricted by these
conditions so that they make it possible to reason
about the available information and to reduce the
solution space.
4 CONCEPT
The axiom of incomplete information makes
ontologies a suitable tool for the exploration of a
network of constraints. Incomplete or uncertain
information can be used for inferring implicit part
properties. For a given part, information about the
geometry, topology and engineering constraints are
collected on the feature level and the B-Rep-level.
As shown in Figure 3, the given information is
combined with PLM-information on the part level.
To these belong information about manufacturing
constraints and information about usage scenarios.
Subsequent reasoning allows constraint solving or
indicates inconsistent assumptions. The linking
between the CAD model and the corresponding A-
Box-Ontology allows returning inferred knowledge
on the level of features and the B-Rep.
KMIS2014-InternationalConferenceonKnowledgeManagementandInformationSharing
402
Figure 3: Ontology-based constraint solving for parametric
product data.
5 FUTURE WORKS
Future work is supposed to extend the existing
framework for the control of uncertainty in different
directions:
The uncertainty information model has to
be extended by time dependent properties.
Strategies for an automated A-Box
Ontology generation will be defined.
Existing feature- and B-Rep-Ontologies
will be combined and extended for
parametric modelling.
Existing translation algorithms between
feature-Ontologies and features have to be
analysed.
6 CONCLUSIONS
Parametric modelling is a well-known standard in
CAD-systems. Combining this technology with
ontologies opens up the possibility to reason over
incomplete information and uncertainty information.
We provide an overview of current literature about
semantics in CAD and PLM and deduce challenges
for future research and development. Our concept
for the use of Ontologies in parametric modelling
extends our existing research findings on the field of
uncertainty representation and proposes new
capabilities for the integration of PLM knowledge
into CAD. Future research on the properties of
probabilistic first-order and description logic opens
up possible applications of Ontologies in CAD.
ACKNOWLEDGEMENTS
The authors like to thank the German Research
Foundation DFG for funding this research within the
Collaborative Research Center (SFB) 805 “Control
of Uncertainties in Load-Carrying Structures in
Mechanical Engineering”
REFERENCES
Abdul-Ghafour, S., Ghodous, P., & Shariat, B. (2012).
IEEE IRI. Integration of Product Models by Ontology
Development (pp. 548-555). Las Vegas, Nevada, USA:
IEEE.
Alani, L. I. (2007). Template-basierte Erfassung von
Produktanforderungen in einem CAD System. Berlin:
Technische Universität Berlin.
Altidor, J., Wileden, J., McPherson, J., Grosse, I.,
Krishnamurty, S., Cordeiro, F., et al. (2011). A
Programming language Approach to Parametric CAD
Data Exchange. Proceedings of the ASME 2011
International Design Engineering Technical
Conferences & Computers and Information in
Engineering Conferences, pp. 779-791.
Anderl, R., & Mendgen, R. (1996). Modelling with
constraints: theoretical foundation and application.
Computer-Aided Design (Vol.28 No.3), pp. 155-168.
Anderl, R., Maurer, M., Rollmann, T., & Sprenger, A.
(2013). Representation, Presentation and Visualization
of Uncertainty. CIRP Design 2012 - Sustainable
Product Design -, pp. 257-266.
Andersen, O., & Vasilakis, G. (2007). Building an
Ontology of CAD Model Information. In K.-A. L. G.
Heasle, Geometric Modelling, Numerical Simulation,
and Optimization: Applied Mathematics at SINTEF
(pp. 11-40). Springer.
Dartigues, C., Ghodous, P., Gruninger, M., Pallez, D., &
Sriram, R. (2007). CAD/CAPP Integration using
Feature Ontology. Concurrent Engineering: Research
and Applications, pp. 237-249.
Ding, L.-J. S. (2010). Ontology-based Semantic
Interoperability among Heterogeneous CAD Systems.
Information Technology Journal, pp. 1635-1640.
Engelhardt, R., Koenen, J., Enss, G., Sichau, A., Platz, R.,
Kloberdanz, H., et al. (2010). Proceedings of the 1st
International Conference on Modeling and
Management of Engineering Processes. A Model to
Categorize Uncertainty in Load-Carrying Systems.
Springer.
Franke, M., Klein, P., Schröder, L., & Thoben, K.-D.
(2011). Ontological Semantics of Standards and PLM
Repositories in the Product Development Phase.
Global Product Development, pp. 473-482.
Hughes, T., Cottrell, J., & Bazilevs, Y. (2005, Vol. 194).
Isogeometric analysis: CAD, finite elements, NURBS,
exact geometry and mesh refinement. Computer
Methods in Applied Mechanics and Engineering, pp.
4135-4195.
Kuhn, O., Dusch, T., Ghodous, P., & Collet, P. (2012).
Framework for the support of knowledge-based
engineering template update. Computers in Industry
(Vol. 63), pp. 423-432.
Ontology-basedRepresentationofTimeDependentUncertaintyInformationforParametricProductDataModels
403
L. Mosch, A. S. (2011). Proceedings of the ASME 2011
International Design Engineering Technical
Conferences & Computers and Information in
Engineering Conferences. Consideration of
Uncertainty in Virtual Product Design. Washington,
DC, USA: ASME.
Lee, Y., Cheon, S., & Han, S. (2005). Enhancement of
CAD Model Interoperability based on Feature
Ontology. SOTECH, pp. 33-42.
Matsokis, A., & Kiritsis, D. (2011). Ontology applications
in PLM. International Journal of Product Lifecycle
Management (Vol.5, No.1), pp. 84-97.
Ramya T. Chaparala, N. W. (2013). Examining CAD
Interoperability through the Use of Ontologies.
Computer-Aided Design And Applications, pp. 83-96.
Song, I., & Han, S. (2010). Parametric CAD Data
Exchange Using Geometry-Based Neutral Macro File.
Cooperaative Design, Visualization, and
Engineering(Vol. 6240), pp. 145-152.
Sprenger, A. (2013). Kollaboratives, ontologiebasiertes
System zur Integration von unsicherheitsbehafteten
Prozesseigenschaften in die Produktentwicklung.
Shaker.
Sudarsan, R., Fenves, S., Sriram, R., & Wang, F. (2005).
A Product Information Modeling Framework for
Product Lifecycle Management. Computer-Aided
Design (Vol. 37), pp. 1399-1411.
Tessier, S., & Wang, Y. (2013). Ontology-based feature
mapping and verification between CAD systems.
Advanded Engineering Informatics, pp. 76-92.
Verhagen, W., Bermell-Garcia, P., van Dijk, R., & Curran,
R. (2012). A critical review of Knowledge-Based
Engineering: An identification of research challenges.
Advanced Engineering Informatics (Vol.26), pp. 5-15.
Young, R., Gunendran, A., Cutting-Decelle, A., &
Gruninger, M. (2007). Manufacturing knowledge
sharing in PLM: a progression towards the use of
heavy weight ontologies. International Journal of
Production Research (Vol.45, No.7), pp. 1505-1519.
Zhan, P., Jayaram, U., Kim, O., & Zhu, L. (2010).
Knowledge Representation and ontology Mapping
Methods for Product Data in Engineering
Applications. Journal of Computing and Information
Science in Engineering (Vol.10, No.2), pp. 1-11.
Zhong, Y., Qin, Y., Huang, M., Lu, W., & Chang, L.
(2014). Constructing a meta-model for assembly
tolerance types with a description logic based
approach. Computer-Aided Design (Vol.48), pp. 1-16.
Zhu, H., Wu, D., & Fan, X. (2010). Assembly semantics
modeling for assembling process planning in virtual
environment. Assembly Automation (Vol.30, No.3), pp.
257-267.
KMIS2014-InternationalConferenceonKnowledgeManagementandInformationSharing
404
