Design and Implementation for Ontology Modeling of Design Knowledge Based on UML Class Diagram

TELKOMNIKA, Vol.14, No.3A, September 2016, pp. 326~332
ISSN: 1693-6930, accredited A by DIKTI, Decree No: 58/DIKTI/Kep/2013
DOI: 10.12928/TELKOMNIKA.v14i3A.4399



326

Design and Implementation for Ontology Modeling of
Design Knowledge Based on UML Class Diagram
1

Heng Xu1*, Oren Musicant2
School of Management, Henan University of Technology, Zhengzhou, Henan, 450001, P. R. China
2
Computer Department, Rohm and Haas Company, Aachen 52064, Germany
1
*Corresponding author, e-mail: xhletter@126.com

Abstract
With the research and development of knowledge management in product design, OWL models

and other semantic languages have some problems such as the fuzziness hierarchy, lake of the
representation of visualization model, so the modeling of semantic web maybe become the main
bottleneck. Now it has become important research direction that applying UML which is the concept
modeling language into the modeling of semantic web. This paper proposes the knowledge management
model in product design through the UML class diagram. By using the UML modeling method and objectoriented technology the knowledge-oriented task structure model of product design was established
though the task structure of product design. And the knowledge-oriented component structure model of
product design was established though the component structure of product design and object-oriented
design knowledge of design feature. At last, through the simulation of design knowledge ontology that was
established and tested for the bicycle product, it shows that using this method can build ontology modeling
directly and clearly which also can reduce the complexity of establishment and maintenance of large-scale
ontology model, and lay a good foundation for the future research about developing more mature semantic
soft.
Keywords: design knowledge, ontology modeling, UML, class diagram
Copyright © 2016 Universitas Ahmad Dahlan. All rights reserved.

1. Introduction
With the continuous development of manufacture information engineering, it is an
inevitable trend that realizes the design data management and the design process management
in the domain of product design based on the semantic model. At the same time,
implementation of effective knowledge management mechanism in the manufacturing supply

chain network is the important ways that could train the competitive advantage of the whole
manufacturing supply chain. Therefore, it has the important theoretical and practical value that
establishes the semantic model of knowledge application for supporting the collaborative
product design (CPD) [1]. Product design, especially variant design, is capable of making rapid
response to the variations of market demand by reusing the design experiences and knowledge
accumulated in years of design practice [2]. Since product design is a kind of knowledge
intensive activities involving multiple knowledge fields, complicated product design tasks are
generally divided into a series of relatively simple subtasks or subcomponents through
decomposition and integration. This kind of characteristics of product design has laid a good
foundation for building the ontology model for CPD based on structural relationship between
task and component.
2. Knowledge-Oriented Structure Model of Product Design Task
Being different with common information, design knowledge presents more complex
expression ways. This attributes to that the complex correlations in the structure, behavior, and
function of design knowledge also are also in a dynamic evolution. Such correlations would be
damaged when the static and concentrated methods were used for organization and storage.
Moreover, the knowledge in the database is hard to be queried, transformed, and updated due
to the information deficiency concerning the background and usage process of the knowledge.
Therefore, the product design task-oriented knowledge organization, as well as the


Received March 16, 2016; Revised July 18, 2016; Accepted August 1, 2016

327



ISSN: 1693-6930

characteristics and attribute information of knowledge unit recorded would facilitate the query
and utilization on the knowledge in perspective of design task [3].
The structural model of the knowledge-oriented product design task defines the basic
functional structure of the acquisition, description, management, and usage on the information
and knowledge of the tasks in design knowledge management. Such task structure model
further provides standard description framework for the information and knowledge of task.
Figure (1) shows the UML static structure of this model [4].
State Class

Identification Class

Attribute Class


Operation Class

-s_id : string
-s_name : string
-s_attribute : string
-s_operation : string
-s_condition : string

-id_id : string
-id_description : string
-id_CreatTime : Date
-id_CreatDepartment : string
-id_AssignDepartment : string
-id_TaskOptional : bool = 1

-at_id : string
-at_name : string
-at_type : string
-at_value : string

*

-o_id : string
-o_name : string
-o_parameter : string
-o_return_value : string

1

1

*

1

1
*
Association Class

Task Class

1

-t_id : string
-t_name : string
-t_role : string
-t_component : string
-t_rosource : string
-t_operation : string
-t_parent : string

*

1
*
1
Knowledge Class::Rule
*
Knowledge Class

1

Interface Class
-in_id : string
-in_name : string
-in_role : string
-in_information : string
+getComponent_of_Task() : string
+getRole_of_Task() : string

-as_id : string
-as_name : string
-as_role : string
-as_attribute : string
-as_operation : string
-as_restriction : string

-k_id : string
-k_name : string
-k_role : string
-k_attribute : string
-k_value

-k_TaskAttribute : string
-k_ComponentAttribute : string

Knowledge Class::Method

Knowledge Class::Fact

Figure 1. Knowledge-oriented structure model of product design task
3. Object-Oriented Design Feature Knowledge
Features are the information set that constitute the expressions concerning the shape,
parameters, attributes, and properties of product. It mainly includes shape, accuracy, assembly,
material, and performance analysis etc. in content. In accordance with the requirements in
different design phases, features are categorized into demand features, design features,
manufacturing features, and evaluation features etc. In conceptual design phase, design
features are the carrier of the design information. As a functional unit, they describe the product
structure plan using the abstract information sets composed by certain functional semantics and
shapes. In case of the design features being close to the thought way of human, the hierarchical
information of the abstract product can be effectively matched with the detail structural
expression. Moreover, the descriptions on the features are not merely pure result expression.
Instead, they include design intentions and the factors considered in the design process [5]. Sun

et al. pointed out that features were the geometric shapes or entities that were combined with
certain functions and engineering semantics in the whole life cycle [6].
As the early embryos of the parts and components composing the product, design
features can be used to describe the basic composition of the design product. In conceptual
design phase, the structures that are designed for function purpose are mainly analyzed in
regard of principle rather than manufacture. Moreover, the design features are extracted to
establish corresponding design features database in actual application domain. Design features
show the following features: (1) they are associated with the geometric description of parts; (2)
they show certain engineering semantics; (3) they can be identified and conversed; (4) they
show disparate forms and connotation in different engineering activities; (5) they are capable of
covering all the requirements in their application domains [7].
Brunetti proposed a feature-based integrated product model, which involved the
function model, working principle model, assembly model, component model, feature model,
TELKOMNIKA Vol. 14, No. 3A, September 2016 : 326 – 332



ISSN: 1693-6930

TELKOMNIKA


328

generalization model [8]. However, this model failed to solve the connection and mapping
relationship between models effectively; Jiang et al. redefined the connotation of demand
features, design features, manufacturing features etc. [9]. Moreover, he put forward the concept
of neutral features and the component structure modeling method based on features and
multiple plans; Xia et al. constructed the generalized feature information model by packaging
the design knowledge and generalized features [10]. Using this model, the design process was
expressed through constructive solid geometry; Wu et al. stated that the relationship between
user demand and original understanding could be established by employing design features as
functional units [11]. Moreover, the design information could be brought to downstream for
application. On these bases, she proposed the design feature-based conceptual design
process.
In this study, the design feature model was described using object-oriented technology.
In this mode, related knowledge, such as the attributes, dynamic behavior, domain knowledge,
and process method of the object, were sealed in a unified object structure to more conveniently
abstract the features of design feature and the design knowledge used to comprehensively
describing a feature. Figure 2 shows the UML static structure of this model.
In this model:

(1) Correlations are classified as the static relationship between objects. It is used to
represent the information among objects, such as the position relationship, assembly
relationship, and tolerance matching relationship.
(2) Attributes refer to various attributes of the design features. It includes two aspects,
namely, designing attributes and geometrical attributes.
(3) Rules are design rules of the features. They are expressible using traditional {if...
then...} production rule set and are represented specifically using a series of design process and
design regulations. Design rule is a kind knowledge itself.
(4) Methods are the methods operated to the feature object.
-Attribute

Attribute

Method

-InheritedAttribute
*

DesignFeatu
reKnowledge

*
*

-Attribute

Rule

-m_id
-r_id
-m_name
-r_name
-m_MessageMode
-r_precondition if…
-m_LocalVariable
-r_conclusion then…
-m_inheritance
-r_inheritance
-InheritedAttribute
1

*
1

AssociationRelationship
Attribute::GeometricAttribute

Attribute::DesignAttribute

-ga_id
-ga_name
-ga_type
-ga_value
-ga_inheritance

-da_id
-da_name
-da_type
-da_value
-da_inheritance

-ar_id
-ar_name
-ar_precondition
-ar_conclusion
-ar_inheritance
-ar_instance

Method::MethodProcess
-mp_OperatingFunction
-mp_ConditionalStatement
-mp_JudgmentStatement
-mp_ControlStatement
-mp_LoopStatement

Figure 2. Object-oriented UML view of design feature knowledge
4. Knowledge-Oriented Structural Model of Product Design Component
The collaborative product development is based on the unified structure model of
product design established. The support system of the CPD under distributed environment also
calls for the support of complete product design knowledge. On the basis of the object-oriented
design feature knowledge and the decomposition structure of product design component, the
knowledge-oriented structural model of the component of product design was built. A complete
component structure model of product design is a set of the product design structure and also a

Design and Implementation for Ontology Modeling of Design Knowledge Based on … (Heng Xu)

329



ISSN: 1693-6930

reflection of the organization approaches, expression approaches, and correlations of the
knowledge related to design feature.
4.1. Top-Level Package: Product
Product is composed of components and parts, components are assembled from parts,
while parts are provided with various design features. Therefore, the top-level package of
product is defined using an object-oriented method which are constituted by product,
component, and part. Component is a subclass of the product, while part is a sub-class of
component.
4.2. Top-Level Package: Design Feature
This package is a detailed definition of a series of sub-feature model and corresponding
relationships in. It is composed of three subclasses, namely, design feature, feature definition,
and feature version. Design features provide a definition on the only global identification feature
ID of features; Feature definition is used as supplementary information to define the feature
code. Moreover, the hierarchical structural relationship between features is defined by the ID of
parent features; Feature version is used to define the different design data with the same
feature formed in different phases obtained from different designers.
4.3. Top-Level Package: Design Feature Knowledge
It is formed by the entity objects of design features, such as the dimension, structure,
surface roughness, and tolerance of form and position of features. Since this top-level package
inherits the object-oriented static structure of the knowledge of design features, the design
knowledge correlated with the object in this class are expressed by the knowledge structure of
the design features in Figure 3.
Using components and features as basic objects, this model supports the collaborative
development support system of products under distributed environments obtaining the product
structure information required. Meanwhile, it provides high level knowledge expressions that are
in agreement with the human thoughts and reflects the design demand and design features.
Thus it lays an important basis for building the ontology model of CPD.

SubClass
*
Top-level
package::Prod
uct

*
*

Top-level
package::Desi
gnFeature

Product

Component

Part

DesignFeature

-c_id
-c_name
-c_GroupCode
-c_FunctionalRequirement
-c_ComponentVariable
-c_OuterDimension
-c_Assembly/Component
-c_ComponentVersion
-c_ProductionUnit
-p_id
-c_parent_id

-pa_id
-u_user_id
-t_id
-pa_type
-pa_name
-pa_GroupCode
-pa_FunctionalRequirement
-pa_OuterDimension
-pa_JobDimension
-pa_Assembly/Component
-pa_PartVersion
-pa_ManagementInformation
-pa_DesignInformation
-pa_FeatureNumber
-pa_ProductionUnit
-c_id
-pa_parent_id

-df_id
-df_relation
-df_InstanceStatus
-df_GeometricalRelationship

1..*

 hasFeature

*
*

Top-level
package::Desi
gnFeatureKn
owledgeClass

*

1

1
1
*

1
1..*

DesignFeature::Definition
-df_code
-df_name
-df_FeatureCode
-df_type
-df_parent_id

inherit
inherit

inherit

*
1

*

inherit
*

DesignFeature::GeometricalTolerance
DesignFeature::Roughness
-GeometricElement_id
-Roughness_id
-Roughness_type

1..*

-p_id
-p_name
-p_GroupCode
-p_FunctionalRequirement
-p_ProductQuality
-p_OuterDimension
-p_Assembly/Component
-p_ProductVersion
-p_ProductionUnit
*

*

*

SubClass

-GeometricElement_id
-gt_id
-gt_type
-gt_value
-gt_grade

1

1

DesignFeature::Dimension

1
DesignFeature::Version

DesignFeature::Structure
-df_version_id
-GeometricElement_id
-GeometricElement_name

*

-df_id
-df_version_id
-df_version_status
-df_LocatingSurface
-df_LocatingPoint

*

-df_version_id
-df_dimension_id
-df_dimension_type
-df_dimension_value
-AccuracyGrade
-UpperDeviation
-LowerDeviation

Figure 3. Knowledge-oriented structure model of product design component

TELKOMNIKA Vol. 14, No. 3A, September 2016 : 326 – 332

TELKOMNIKA

ISSN: 1693-6930



330

5. Results and Analysis
This paper selects the bicycle which is relatively simple manufacturing product for
modeling object. Because of the intensifying competition in bicycle industry, the product
differentiation and innovation of parts is a very urgent task. It is considered that in the bicycle
industry combined with consumers, the development for the consumer demand and interact with
the consumers in the process of design can improve the product level and enhance the core
competitiveness. The bicycle ontology model of CPD established in this paper is based on the
national standard of People's Republic of China "Classification, Name and Main Terms of
Bicycle Parts" (GB/T 3564-93) and "General Technical Conditions of Electric Bicycle" (GB
17761-1999), and the industry standard "Method of Bicycle Naming and Model Compilation"
(QB 17714-93).
5.1. Logical Structure of Design Knowledge Ontology of Bicycle
According to the above three standards, the ontology structure of design knowledge for
bicycle based on the structure of task and component is shown in Figure 4.

Figure 4. Logical structure of design knowledge ontology of bicycle (Part)
5.2. Query Test of Design Knowledge Ontology of Bicycle
On the basis of design knowledge ontology of bicycle, we define the following query
rules (see, Figure 5).

Design and Implementation for Ontology Modeling of Design Knowledge Based on … (Heng Xu)

331



ISSN: 1693-6930

Figure 5. Code of a knowledge entry query
The search result is shown in Figure 6. There is one record matched the query rules
exactly. It is an instance of class “electric bicycle". To view the detailed information and
knowledge entries contained by this instance is easy by clicking on the instance name. The
search result shows that the knowledge ontology model for CPD based on task and component
is reasonable and perfect on semantic rules, and is able to provide the full support for the
knowledge retrieval, also could lay a good foundation for knowledge reasoning [12-14].

Figure 6. A design knowledge retrieval page in Protégé
6. Conclusions
The knowledge-oriented task structure model of product design was established though
the task structure of product design; the knowledge-oriented components structure model of
product design was established though the components structure of product design and objectoriented design knowledge of design feature. According to the above analysis, this paper
defines the hierarchical structure of knowledge ontology of CPD and the main relationships
among the classes of CPD. At last, this paper selects the bicycle for modeling object and has
carried on the simulation test for knowledge ontology model for CPD by using Protégé. The
results shows that the ontology model for CPD based on structural relationships between tasks
and components is accurate and effective on the syntax rules, also is reasonable and perfect on
semantic rules.

TELKOMNIKA Vol. 14, No. 3A, September 2016 : 326 – 332

TELKOMNIKA

ISSN: 1693-6930



332

Acknowledgements
This work is partially supported by the National Nature Science Foundation of China
(71001034) and the Grant of Youth Backbone Teachers of Institutions of Higher Education in
Henan Province (2012GGJS-090) and the Doctorial Foundation of Henan University of
Technology (2009BS017) and the Grant of Science and Technology Innovative Talents of
Henan University of Technology (2011CXRC21) and the Science Research Project of Education
Department of Henan Provincial Government (2010B630008) and the Fundamental Research
Grants for the Henan Provincial Colleges and Universities (2014YWQQ23) and the Key
Research Institute of Humanities and Social Sciences in Henan and the Program for Scientific
and Technological Innovation Talents in Colleges and Universities of Henan Province.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]

Xu H. Knowledge representation and mapping for collaborative in manufacturing supply chain based
on OWL and UML. Information Technology Journal. 2013; 12(21): 6295-6301.
Reglil WC, Cicirello VA. Managing digital libraries for computer-aided design. Computer-Aided
Design. 2000; 32: 119-132.
Hai S, Zuhua J, Jun L. The product design knowledge chain management based on knowledge map
in network environment. Journal of Shanghai Jiaotong University. 2007; 41(7): 1061-1065.
Zhixue W, Xin J, Qingchao D, Fengke F. UML-based domain rule modeling method. Journal of
Nanjing University of Science and Technology. 2012; 36(6): 932-938.
Zhang Zhiyong, Liu Jie, Zhang Xinhui. UML Modeling for Dynamic Logistics System Based on DEVS.
TELKOMNIKA Indonesian Journal of Electrical Engineering. 2013; 11(4): 2168-2173
Shu S, Wentian G, Xin J. The study of feature conversion based on design feature. Modern
Manufacturing Engineering. 2007; 6: 57-59.
Fakhar F. Semantic Constraints Satisfaction Based Improved Quality of Ontology Alignment. Bulletin
of Electrical Engineering and Informatics. 2013; 2(3): 182-189.
Peipei Z, Hua T, Xiaofeng G. The study of parametic finite element modeling in CAD/CAE integration.
Modern Manufacturing Engineering. 2006; 9: 62-63.
Sree MK, Subrahmanyam K. Retail Web System Upgrading with Strategic Customer Using Threshold
Policy. Journal of Telematics and Informatics. 2015; 3(1): 28-32.
Shan J, Yuxin W. Feature based multi scheme modeling for part structure. Modular Machine Tool &
Automatic Manufacturing Technique. 2004; 10: 46-48.
Dadkhah M, Obeidat MM, Jazi MD, Sutikno T, Riyadi MA. How Can We Identify Hijacked journals?
Bulletin of Electrical Engineering and Informatics. 2015; 4(2): 83-87.
Yu X, Rong M, Junbo Z, Qing T, Zhiyong C. Design process saving and reconstruction based on
knowledge. Journal of Computer Aided Design & Computer Graphics. 2005; 17(12): 2702-2708.
Xiaojuan W, Yujin H. The application of design feature in conceptual design. Machinery Design &
Manufacture. 2007; 9: 52-54.
Jia L, Geming T, Ming L, Fengkui Z. Ontology based semantics reasoning of UML class diagram.
Computer Applications and Software. 2011; 28(4): 212-214.

Design and Implementation for Ontology Modeling of Design Knowledge Based on … (Heng Xu)