An Ontology-Based Manufacturing Design System

In this study, an ontology-based system is proposed to solve problems raised in the manufacturing design by expanding traditional development activity with Knowledge Management (KM), a Product Knowledge Model (PKM) and the Intelligent Application System (IAS). The KM helps to management the knowledge in the design process, while the PKM supports to locate proper information and the IAS is responsible for applying the product knowledge among different application systems throughout the product life cycle. The PKM is encoded in OWL to realize semantic match and enhance the performance of organization capability and knowledge sharing. The routine design assistance is developed to reuse the product knowledge based on configuration method generating function satisfied solution rapidly by reasoning the configuration rules represented in SWRL. The information retrieval theory is involved to support manufacturing knowledge sharing. A prototype system of binging machine design is developed to verify the proposed approach, using the semantic web technology, for seamless sharing domain-specific design knowledge among multidisciplinary organizations and intelligent supporting the manufacturing design.

INTRODUCTION

To meet market challenges, manufacturers should now improve their efficiency in terms of product development and resource utilization. This means a dramatic reduction in the overhead cost of redundant engineering work, eliminating repetitive design tasks, reallocating technical resources to new product development (Kulon et al., 2006). However, many repetitive engineering tasks are not automated, which creates a bottleneck that slows down the design cycle. If the input specifications change at any time, the entire process must be repeated. The effect of the bottleneck is multiplied by the number of times the design loop is repeated. Although, the software tools such as CAD and CAE can help reducing time and costs involved, they go only part of the way toward providing a completely flexible tool for adapting to changes in a product design and provide no means of capturing the engineering expertise behind the design, which are knowledge about manufacturing design constituting one of the most valuable assets of a modern enterprise, normally known implicitly to the participating designers, relying heavily on the personal experience background of each designer (Brandt et al., 2007). Moreover, multiple designers are frequently involved in manufacturing design and designers often use their own terms definitions and tools to represent a product design (Kim et al., 2006).

Thus, a new approach should be explored as: (1) to provide a product knowledge model-comprehensive representation of the contents of knowledge assets, such as artificial sources, domain terms and multiple experts, etc., (2) to enable the capture and archival of work processes in order to provide information about the circumstances in which the individual knowledge items have been created, (3) to establish the protocol to manage the knowledge and (4) to develop the intelligent system to reuse the knowledge.

As an explicit specification of a conceptualization, ontology plays a significant role in these aspects: sharing information, integrating different applications, implementing interoperability and reusing knowledge. As a matter of fact, some research effort has been attempted to employ ontology-based techniques in manufacturing or in the production domain for integrating engineering applications. Ahmed et al. (2007) developed an ontology of engineering design for knowledge sharing among engineers to assist engineers in indexing, searching and retrieving design knowledge. Cho et al. (2006), realized that seamless integration of current digital part libraries was impeded by semantic heterogeneity. They developed meta-ontology for part libraries to provide an integrated view of multiple part libraries. Based on meta-ontology, domain-specific part ontologies were described. Efforts to add information and/or knowledge to the geometric data from the KBE tool to enable the exchange of product model were discussed by Zhu and Deng (2002). It is shown that in a KBE environment information/knowledge stored in the product model can be extracted and made accessible. In Nawijn et al. (2006), an ontological approach was proposed to enable the exchange of features between applications. It modeled participating ontologies and creates a common intermediate ontology. Rules were manually specified to enable the mapping of concepts from one domain to another. Masaharu et al. (2004) proposed a formal set of ontologies for classifying analysis modeling knowledge, based on the concept that engineering analysis models are knowledge-based abstractions of physical systems. A product modeling language was developed (Patil et al., 2005). It defined products as object sets and relations. Above, research efforts mainly focus on certain aspects of product information modeling, without involving the whole process of the manufacturing design.

In this study, integrating with the earlier studies (Sun et al., 2007; Gao et al., 2007; Sun et al., 2008a) (the studies addressed on the second and the third issues listed above), we concentrates on the left issues, product knowledge modeling and the intelligent system developing. An Ontology-based Manufacturing Design System (OMDS) is explored to present the PKM, realizing semantic match and enhancing the performance of organization capability and knowledge sharing. Ontologies provide a source of shared and precisely defined terms that offer a means to structurally represent, to improve their accessibility to automated processes. Designers are no longer merely exchanging specific geometric data, but rather more knowledge about design and the manufacturing design process, including specifications, design rules, constraints and rationale. The IASs are developed to reuse, retrieve and integrate the manufacturing design knowledge; one is routine design assistance based on configuration method generating function satisfied solution rapidly by reasoning the configuration constraints rules represented in SWRL; the other is general manufacturing design assistance based on knowledge sharing and retrieval theories. A prototype system of binging machine design is developed to verify the proposed approach, using the semantic web technology, for seamless sharing domain-specific design knowledge among multidisciplinary organizations and intelligent supporting the manufacturing design.

THE FRAMEWORK OF OMDS

A proposed OMDS was illustrated in earlier study (Sun et al., 2008a), shown in Fig. 1. The centre of the model is the central Design Repository (DR), which manages all the engineering knowledge. Standards books, domain ontology, design rules encoded in the system by the knowledge engineer etc., are all integrated into the DR and are located by Product knowledge Model (PKM). The role of the engineer who interacts with the model is to provide the intelligent system with the input specifications such as component geometry and make the important design decisions. The design engineer can easily modify the design either by changing the answers to some of the key design questions or by modifying the input specifications. Once a design has been generated, the artifact can be exported to a FEA allowing the designer to test several performance parameters in order to improve the design. The design loop can be repeated efficiently as many times as necessary without incurring the expensive trial-and-error procedure and experimental work. With the product model, generating a new design can be done much faster in a matter of hours by a single design engineer, as opposed to several days or weeks with traditional methods. The main components of OMDS are as follows.

User interface: The user interface is developed in order to provide Web-based access to the data files or databases. This interface enables users to dynamically select data they want to visualize, as well as how they wish to visualize them.

Fig. 1:	The system framework

Design repository: DR includes rule base, product case base, ontology base, engineering analysis template base, manufacturing information base and etc. All of these multiple resources are organized and indexed through Product Knowledge Model. The rule base stores experts’ design experiences, product configuration rules, variant rules and constraints. The case base collects cases of product, component and part. The ontology base stores all ontologies of PKM. The manufacturing information base has dynamic manufacturing resources.

Knowledge management: Knowledge management includes knowledge acquisition, knowledge presentation and knowledge examination. There are many proceeding method according to the type of knowledge. In this study, the knowledge includes ontology and rules particularly. Knowledge acquisition is the process of extracting domain knowledge from human sources or unstructured and semi-structured data held in organizational intranets. We proposed a product configuration rules acquirement method based on rough set theory (Gao et al., 2007), resolving the bottleneck in knowledge acquisition. The acquisition of ontology concept will be illustrated at length in next section. Once knowledge has been acquired from human sources, literature or automatically extracted by acquisition tool, knowledge representation in a proper language is required. The knowledge base from experienced analysts and engineers is constructed in SWRL, a rule language based on OWL. We use Protégé as ontology editing tools which support the creation, maintenance and population of ontologies. It is a free, open source platform and offers direct communication with several rule engines, such as Jena, Racer and Jess. Furthermore, Protégé can be extended by way of a plug-in architecture and a Java-based Application Programming Interface (API) for building knowledge-based tools and applications (Zhao and Ziu, 2008). Regarding the issues of redundancy and circularity in maintenance of rule base, the detection algorithm was established based on directed hypergraph (Sun et al., 2008b). Compared with other existed algorithm based on graphs, new algorithm is more efficiency since its graph is concise and according matrixes are small-scale.

Intelligent application systems: Most importantly, the IASs need be applied to reuse, retrieve and transfer the design knowledge. They are generally combined with commercial application software, knowledge engines and domain databases. As regards routine design, which is a relatively regular design mode and product structure, a configuration system is developed to generate satisfied solution rapidly and reuse the knowledge through reasoning the configuration constraints rules. As regards non-routine deign, which needs more knowledge support in conceptual design, detailed design, engineering analysis and etc., a general manufacturing design assistance is proposed based on knowledge sharing and retrieval theories to support designer making decisions effectively.

Workflow management: Finally, information access tools are required to allow end users to exploit the knowledge at the right. An effective approach supporting knowledge supply is proposed to establish the relation between design process and product knowledge in (Sun et al., 2007). The method of artificial neural network is applied to extend traditional workflow technology.

THE PRODUCT KNOWLEDGE MODEL

The PKM is a map of product knowledge, which can help user to locate proper information and support reuse of product data among different application systems throughout the product life cycle, by utilizing ontology to add semantics to manufacturing design information for semantic searching.

Definition of ontology: Ontology can be simply defined as a formal, explicit specification of a shared conceptualization. A formal ontology means the complex semantics of concepts and the relations among concepts, their properties, attributes, values, constraints and rules. More importantly, ontologies as expressed in an ontology representation language such as Web Ontology Language (OWL) can add values to taxonomies or thesauri through deeper semantics.

Ontology design process: The ontology design process follows the widely accepted five-stage methodological approach (Kim et al., 2004): (1) identification of purpose and scope, (2) knowledge acquisition and conceptualization, (3) integration and reuse of other ontologies, (4) formal specification and (5) evaluation and documentation. This study uses OWL to represent the ontological framework for manufacturing design.

Ontology of PKM for manufacturing design: The PKM should facilitate the location of appropriate contact persons for business tasks requiring specific knowledge, experiences, or skills in manufacturing design. To achieve this objective, staff skill management (organization competency), product semantic, development process and information management appear to be the most important concepts. Definition of PKM is as follows:

Fig. 2:	PKM using UML notation

It contains the classes shown in Fig. 2. The classes are organized in four different ontologies, Information Resource Ontology (IRO), Management Activity Ontology (MAO), Product Semantic Ontology (PSO), as well as Organization Competence Ontology (OCO) in order to realize semantic match for knowledge search and enhance the performance of organization capability and knowledge sharing. The IRO is used to describe the manufacturing design information just like metadata. The OCO is integrated based on competence ontology (Gauthier et al., 2005). The competence ontology is designed to identify employees with particular skills as well as to determine the kinds of skills required to conduct tasks. The MAO describes the work processes to be performed by a designer or a design team confronted with a certain task. The PSO is based on the NIST Core Product Model (CPM) and its extensions, the Open Assembly Model (OAM), the Design-Analysis Integration model (DAIM) and the Product Family Evolution Model (PFEM) (Sudarsan et al., 2005). The PKM can support the full range of PLM information needs and help designers to construct and extend the design knowledge base within which domain concepts, design objects, dependency information of design objects are well structured and consistently maintained. With the help of PKM, the knowledge of manufacturing design is associated with the context and can be effectively located by scenarios reuse.

Ontologies in OWL: For reasons of interoperability, the Web Ontology Language OWL, a W3C standard for the Semantic Web, would be the first choice for the representation of ontologies. A class defines a group of individuals that belong together because they share certain properties. Classes can be organized in a specialization hierarchy using sub class of. Properties are determined based on whether they relate individuals to individuals (ObjectProperties) or individuals to datatypes (DatatypeProperties). For example, the encoding of the class FEM in OWL RDF/XML is shown in Fig. 3a. Figure 3b shows the part of RDF/XML files that encodes the object properties hasForms, modeling, as well as the data property author. An individual of the class Artifact, namely Planet carrier is specified in Fig. 3c. The ontology is developed using Protégé 3.3. Figure 4 shows class hierarchies modeled for the PKM, which are created using OWL-Viz plug-in the Protégé.

INTELLIGENT APPLICATION SYSTEM

The role of a knowledge-based design support system is to extend geometric-based representation and reasoning to knowledge-based representation and inference in order to provide wholesome solutions to a wide range of design problems.

Fig. 3:	The examples of OWL source codes of PKM, (a) encoding classes in OWL, RDF/XML, (b) encoding properties in OWL RDF/XML and (c) encoding individuals in OWL RDF/XML

The competence of a knowledge-based design support system architecture needs to be evaluated based on the following criteria (Tang, 1997); It should provide an efficient mechanism to transform an initial design requirement description to a design specification and allow designers to vary data, design requirements, problem solving strategies or evaluation criteria, to obtain alternative design solutions; It should provide explicit explanations and justifications for any chosen aspects of the current status of a design, not only in terms of how something has been derived, but also in terms of why something is not happening as expected, locating areas of difficulty and suggesting strategies for solutions contribute to effective decision making in design; It also should have easy access to visualization tools, solid modeling tools and other analytic systems and enable semantic interoperability between heterogeneous design software. Therefore, we develop the IASs to achieve these criteria.

Configuration system-an intelligent tool to reuse the product knowledge in manufacturing design: configuration ontology: The configuration ontology identifies concepts and relations common to all the product configuration domains, which extend the set of modeling concepts presented by Felfernig et al. (2001). Based on this, we build the configuration ontology as the association class between product version and component version that defines the actual configuration of component versions in each of the product versions in Product Family Evolution Model (Fig. 4). It provides general concept and a common semantic foundation for modeling configuration knowledge and can be reused in different application domains. Based on the general configuration ontology, we have established the paper currency binder domain configuration knowledge and rules in EXPRESS (Gao, 2003). EXPRESS is a powerful modeling language covering data types, structural aspects and complex constraints, widely used to construct a family of robust and time tested standard application protocols, which had been implemented in most CAX and PDM systems. But the lack of a formal semantic model for EXPRESS schema and the complexity of EXPRESS itself impose challenges on the serialization of product instance and data exchange (Zhao et al., 2008). In this study, we present an ontology based methodology for encoding knowledge and rules of specific product domains (Paper Currency Binder in this paper) in OWL and SWRL. Since the encoding knowledge in OWL is illustrated in previous Section, the representation method of configuration rules in SWRL is detailed in the following.

Configuration rules: To represent rule knowledge, W3C has developed SWRL rule language (Ivan, 2007), which is tightly integrated with OWL since the predicates in SWRL rules may be OWL-based classes or properties.

Fig. 4:	The class hierarchies of the PKM

SWRL combines the OWL DL and OWL Lite sublanguages with the Unary/Binary Datalog RuleML sublanguages and has the advantages of close association with OWL. In common with many other rule languages, SWRL rules are written as antecedent-consequent pairs. In SWRL terminology, the antecedent is referred to as the rule body and the consequent is referred to as the head. The head and body consist of a conjunction of one or more atoms. Atoms are the form C(x), P(x, y), has_type (x, y), where C is an OWL description, P is an OWL property, x and y are variables, OWL individuals or OWL data values. Variables are indicated using the standard convention of prefixing them with a question mark (e.g., ?x). For example, a configuration rule expressing that The Sliding_I port of the component type SR_I must be connected with the Bearing_ I port of the component type TB_I would then be Fig. 5.

To model the above constraints, SWRL rule editor (O’Connor, 2007), a plug-in integrated in Protégé for modeling rules, is utilized in our research to facilitate this modeling process. These configuration constraints in the form of SWRL rules can be encoded and saved in knowledge bases in XML files according to SWRL/XML syntax. Details about SWRL/XML can be referred to Horrocks et al. (2004).

Fig. 5:	The example of SWRL

Fig. 6:	The architecture of the developed product configuration system

The configurator: The role of the configurator is transforming the customer requirement to the product solution through reasoning the configuration knowledge/rules. Since, the customer requirement, configuration knowledge and rules are formulized by ontology in OWL/SWRL, an inference engine supporting OWL/SWRL reasoning is needed.

The SWRL specification does not impose restrictions on how reasoning should be performed with SWRL rules. Thus, investigators are free to use a variety of rule engines to reason with the SWRL rules stored in OWL knowledge base. They are also free to implement their own editing facilities to create SWRL rules (Zhao et al., 2008). Mei et al. (2005) provided an implementation for mapping OWL and SWRL semantics to JESS facts and rules by means of eXtensible Style sheet Language Transformation (XSLT) (World Wide Web Consortium, 1999). The proposed tools OWL2Jess and SWRL2Jess based on XSLT can be used to fill the gap between the two ontology languages and Jess. The domain knowledge, that is, ontologies and rules are transformed to Jess facts and rules by OWL2Jess and SWRL2Jess processing the XML based (World Wide Web Consortium, 2000) syntax of OWL/SWRL. Meanwhile, pre-defined Jess rules are used to encode the OWL/SWRL Semantics. The Jess rule engine is run against the resulting rule base to get implicit and explicit information.

Figure 6 illustrates the architecture of the developed product configurator o n the above mentioned tools. Protégé is employed to edit configuration ontologies and rules, OWL2Jess and SWRL2Jess are used to transform OWL ontologies, SWRL rules to Jess facts and rules and the Jess rule engine is used in the configuring processes. The process is finished until no rules are in the rule base and then the configuration results are derived and displayed. The Jena/Jess/Protégé APIs are employed to support the querying, editing or other operations from the outside Java-based applications and Web services.

The general manufacturing design assistance-an intelligent tool to retrieve and share the product knowledge in manufacturing design knowledge retrieval model: Manufacturing design is a knowledge-intensive process that encompasses conceptual design, detailed design, engineering analysis, assembly design, process design and performance evaluation. Each of these tasks involves various areas of knowledge and experience. The sharing of such knowledge and experience is critical to increasing the capacity for developing products and to increasing their quality. It is also critical to reducing the duration and cost of the development cycle (Chen et al., 2005). Therefore, we propose the general manufacturing design assistance based on knowledge sharing and retrieval theories.

In documents the effect of semantic annotation is to allow the computer interpretation of the documents from particular viewpoints (supported by the ontologies that map the concepts and relationships in the relevant domain). It is proposed that the same effects may be achieved for manufacturing design knowledge by using the same approach and even some of the same technologies. Inspired by the observation in text documents retrieval (Sun et al., 2008c), this study develops a general manufacturing design assistance to intelligent support the designer make decision during the product development by retrieving the reference design knowledge automatically. We propose a knowledge-aided manufacturing design assistance based on the knowledge retrieval model (KRM).

Definition 1. A KRM is formulated as:

where, K is A set of knowledge items, annotated by PKM.

Q is A set of user requirements, annotated by PKM; Q = {KC, UP, UQ, HS}, where KC is the knowledge context of design scenario, UP is designer’s preference, UQ is user query and HS is a hybrid semantic set of above three sets.

R-a real number, R: DxQ→ÿR. It is a cartesian product of two sets mapping to the real number set R, The value of R is the relevance between user requirements Q and knowledge items K.

The process of knowledge retrieval is presented in Fig. 7. The system obtains the user’s knowledge requirements, including user preferences, scenario and query, constructs the KRM and then selects the appropriate algorithm to compute the relevance R between user requirements Q and knowledge items K according to composition of KRM. The ranked design knowledge is listed to the designer to support him making decision during the development process. The calculation formula of R is different according to the components of Q, as presented in Table 1.

R_UQ presents the semantic relevance between user query and knowledge items, calculating with ontology-based Vector Space Model method (Liu et al., 2006). R_UP presents the semantic relevance between designer’s preference and knowledge items, calculating with the method proposed by Da-Pereira et al. (2006). R_KC presents the semantic relevance between current design scenario and the context of knowledge items, calculating with a novel scenario-based r etrieval method. Scenario-based retrieval method involves reviewing and understanding archived knowledge of similar problems and associated the information to support product development.

Table 1:	The calculation of R

Fig. 7:	The process of knowledge retrieval

We represent the context of reference knowledge and design scenario as a set of concepts of PKM and make use of WordNet which is a lexico-semantic network of words. The knowledge in DR has already presented by PKM, which can be also represented as:

where, KC = {MAO, PSO, OCO}.

Therefore, we formulize the design scenario with the KC. The design scenario can be acquire and recognize by the workflow system and further annotated and formulized by ontology concept and rules defined in PKM (Pan et al., 2005). After formulization, the design scenario of Q is presented as KC:

where, c_i is the leaf concept of KC, selected as the components of the current design scenario concept vector.

The context of knowledge item is presented as:

where, c’_i is the leaf concept of KC, selected as the components of the knowledge item concept vector.

According to user’s requirement for knowledge scenario, weights are assigned to nodes in ontology. Node weight, signed w_i, presents the required degree of i in knowledge scenario. Node weight, from 0 to 1, can be reckoned by following means (Bao et al., 2007):

So, the R_KC value can be calculated as follows:

(1)

where, concept_sim(c_i,c_i’) is the semantic similarity of concepts. The calculation method will be illustrated next.

Semantic similarity: The similarity of the designer’s knowledge requirement and knowledge resources is computed by semantic meaning of their knowledge contexts. The contexts both described as a set of concepts selecting leaves in the KC ontology are mapped by computing the semantic similarities of synsets in the WordNet. Many approaches have been presented in the past to measure the semantic distance between synsets in the WordNet. Alexander et al. (2005) evaluated five of these measures, all of which use WordNet as their central resource, by comparing their performance in detecting and correcting real-word spelling errors. An information-content-based measure proposed by Jiang and Conrath (1997) was found superior. It combines link-distance and information theoretic methods and also takes into account the type of relationship between synsets, depth of the synsets in the wordnet and density of the region around the synsets. Based on these, a weight is attached to the edge between a child node c and its parent node p:

(2)

where, d(p) is the depth of the node p in the hierarchy, E(p) is the number of edges in the child links, the average density in the whole hierarchy and R(c,p) is the link relation/type factor such as hypernymy/hyponymy, meronymy/holonymy, etc.

The parameters α (α≥ 0) and β (0≤β≤1 ) control the degree of how much the node depth and density factors contribute to the edge weighting calculation. Node density in different parts of the hierarchy is different and greater density around a node (e.g., plant/flora section of WordNet) indicates more closeness. Information Content (IC) values for synsets can be calculated as follows:

(3)

where, IC(c) is the information content value of synset c and P(c) is the probability of encountering an instance of synset c. P(c) can be calculated by finding the relative frequency of synset c in a sense-tagged corpus in which each word is tagged with a WordNet synset identifier. The overall distance between two nodes c_i and c_j would thus be the summation of edge weights (calculated as shown in Eq. 2) along the shortest path linking two nodes (Given this as the measure of semantic distance from a node to its immediate parent, the semantic distance between an arbitrary pair of nodes was taken, as per common practice, to be the sum of the distances along the shortest path that connects the nodes):

(4)

where, path(c_i, c_j) is the set of all the nodes in the shortest path from c_i to c_j. One of the elements of the set is L(c_i, c_j), which denotes the lowest super-ordinate of c_i and c_j . The node L(c_i, c_j) is removed from path (c_i, c_j) in Eq. 4, because it has no parent in the set. In the special case when only link strength is considered in the weighting scheme of Eq. 2, i.e., α = 0, β = 1 and R(c,p) = 1, expanding the sum in the right-hand side of Eq. 4, plugging in the expression for parent-child distance from Eq. 2 and performing necessary eliminations results in the following final formula for the semantic distance between concepts c_i and c_j:

(5)

Imagine a special multidimensional semantic space where every node (concept) in the space lies on a specific axis and has a mass (based on its information content or informativeness). The semantic distance between any such two nodes is the difference of their semantic mass if they are on the same axis, or the addition of the two distances calculated from each node to a common node where two axes meet if the two original nodes are on different axes. It is easy to prove that the proposed distance measure also satisfies the properties of a metric. Our measure of similarity between two concepts is as follows:

(6)

IMPLEMENTATION AND APPLICATION OF OMDS

Prototype system design: J2EE is utilized to construct the prototype. The Web-enabled prototype system supports semantic interoperations of multi-disciplinary design applications and uses a multi-tier architecture consisting of a set of client-side interfacing software tools. Figure 8 is the functional architecture showing the major server functional modules and the client interfacing tools. The prototype system consists of four tiers to provide online design services.

Client tier: This consists of interfacing tools and add-on toolkits for CAD and non-CAD applications. Thin web clients are used at this tier with a small amount of client-side code to assemble service requests and receive responses. Most of the web pages consist of information entered directly using JSP. Visualization of the component geometry on the client’s machine is achieved using the Virtual Reality Markup Language (VRML) (Wagner et al., 1997).

Web tier: The Model-View-Controller (MVC) pattern is used at this tier. The controller servlets direct online service requests from clients to relevant model beans for processing.

Fig. 8:	System architecture

The server responses are passed to View for transforming JSP pages to be shown on client browsers.

Domain logic tier: This tier uses Java beans to provide domain services for knowledge management, including knowledge acquisition, knowledge presentation and knowledge examination. The intelligent design engines are also provided in this tier, including configuration engine, knowledge retrieval engine and workflow engine.

Data source tier: This is used to store, retrieve and manage the design objects, OWL files and instance files in a server database and file repositories. OWL ontology files are employed to design tables and their attributes in the RDBMS files. To better understand data structures, the Resource Description Framework (RDF) model is first constructed. Based on the RDF model, annotated web resources are stored in RDF triple tables (Kim et al., 2004).

A case study: The feasibility and effectiveness of the above concepts and methods are empirically validated by a case study implementation at a company that produces Paper Currency Binder in Anshan, China. The current scenario is to design a Binging mechanism of a novel Paper Currency Binder machine. Since it is a new design, the general manufacturing design assistance will be helpful to support the decision making tasks during development.

The browser-based user interface for accessing the design knowledge is shown in Fig. 9a. The tree menu on the left allows the selection of several expandable submenus, which in turn update the main viewing window to guide the user interactively through the design process. Conception design is as shown in Fig. 9a. After a designer logins, his authority will be checked by the system and then be served with proper service. Then he searches the designing requirement table to confirm the function and essential conditions of the products that will be developed. He decomposes the total function to basic function element and establishes the function model of the product, under the support of PKM. By searching the principle depository, every feasible solution of each function element is found and constituted all rational schemes. After evaluating them, the designer will gain an optimum solution finally. The main task of structure design is to realize the product structure of the principle scheme clearly and completely and then get the assembly sketch draft of the product and its key parts. The system can offer the criterion of the structure design immediately which has directive significance to the structure design, such as consider the rational design of the craft, rational design advantageous in standardization, rational design easy to assemble, rational design easy to retrieve, etc. The user interface of structure design is as shown in Fig. 9b. The system offers the criterion of product design and experts’ knowledge and provides multiple functions such as user-friendly interface, the definition and extraction of characteristics, information model management for different parts, etc. If the designer wants to check all the parts which have the specified characteristics, what he needs to do is just to click the corresponding link or enter the keyword; the system will automatically search the DR and list all the parts with icon which match the users semantic. If one of the results is considered to be suitable, the designer just needs to click it and the detailed design data such as assembly drawing, parts drawing, analysis documents, BOM files and etc. (refers to Fig. 9c), is available for him immediately, which makes the agile design process possible. Beyond that, the user can interactively check the geometry, engineering and metadata information on the web directly. Now, it is available to modify and update the corresponding components according to the engineering modification note. It is possible to use web communication tools or email to change ideas at the designing phase.

CONCLUSION

This study presents an ontology-based approach to solve problems raised in the manufacturing design. The key feature of the application is that it integrates the whole design process within the PKM to support locating, reusing, integrating and transferring the design knowledge. The routine design assistance is developed to reuse the product knowledge based on configuration method generating function satisfied solution rapidly by reasoning the configuration constraints rules represented in SWRL. Inspired by the observation in text documents retrieval, the paper develops a general manufacturing design assistance to intelligent support the designer make decision during the product development by retrieving the reference design knowledge automatically. A prototype system of Binging Machine design is developed to verify the proposed approach, using the semantic web technology, for seamless sharing domain-specific design knowledge among multidisciplinary organizations and intelligent supporting the manufacturing design. The authors’ future work is to develop more diverse domain-specific engineering ontologies to get a more extensive multidisciplinary integration.

Fig. 9:	The interfaces of Prototype System (a) aided conception design, (b) aided structure design and (c) aided detailed design

ACKNOWLEDGMENTS

This research is sponsored by the National Natural Science Foundation, China (No. 50475156) and Program for New Century Excellent Talents in University, China (No. NCET-05-0285).