Research On The Key Technologies Of Semantic Knowledge Service For Product Design

Product design is an knowledge-based innovation process. In the product design field,75%of the design types are adaptive design. In a new product design process, more than60%of the work has to reference past design knowledge and experience. Therefore, reusing theaccumulated knowledge is regarded as a better way to improve the efficiency and quality ofproduct design. However, the current research work of knowledge management and serviceis focus on traditional knowledge modeling, searching and pushing, while the researchwork is not involved semantic knowledge service. This situation directly results in seriousproblems during knowledge reusing process. It hard for designers to gain knowledge whichthey really needed for the project, caused the low design knowledge reusing rate and thelong design cycle. To solve the above problems and achieve the goal of improvingknowledge reusing level and alleviating design cycle, several techniques are studiedincluding the method of product design domain ontology construction, semantic of productdesign knowledge analysis and extraction, semantic representation model construction ofproduct design knowledge, the method of knowledge service based on semantic model. Thefollowing shows the detail of each chapter.(1)Research on multi-level multi-dimensional ontology modeling for semanticexpression of product design knowledge. Aiming at the problem of the existing ontologybuilding methods can’t meet the demand of product design knowledge semantic acquisitionand representation, this paper propose a framework of ontology modeling for semanticexpression of product design knowledge. This framework includes four layers, that ismeta-meta-data layer, meta-ontology layer, ontology layer and ontology instance layer, andfour dimensions, that is product design objects dimension, product design organizationdimension, product design process dimension, design knowledge object dimension. Basedon the above content, this paper proposes a method of bottom-up ontology constructionwhich is under the constraint of multi-level and multi-dimension. This method extractsdomain concepts and abstract semantic relation based on the product design instance. Thedomain ontology which is built by this method can truly reflect the actual semanticenvironment in the field.(2)Research on semantic analysis and extraction of product design knowledge under the restriction of DKSRO. Firstly, the characteristics of product design knowledge isanalyzed and the difficulty in semantic analysis and extraction of product design knowledgeis summarized, then a framework of semantic analysis and extraction of product designknowledge under the restriction of DKSRO is proposed. The conditional random fieldmodel (CRFs) is used to identify the instance name in the product design knowledge andannotate semantic dimension to special lexical. The semantic dimension comes from themeta-ontology layer in DKSRO. On this basis, semantic dependency analysis of productdesign knowledge is implemented based on the semantic relevance and axiom in DKSRO.Semantic information of product design knowledge will be stored in the form ofdependence analysis table after semantic dependency analysis, providing the foundation forsemantic model generation.(3)Research on two layers semantic model of product design knowledge. Aiming atthe existing problems in product design knowledge semantic expression model, a twolayers semantic model based on DKSRO is proposed. In this model, product designknowledge semantic unit model (PDKSUM) is associated with DKSRO ontology layer.Product design knowledge semantic unit model includes three parts, the basic model, thecontext model, and the rules model. According to the corresponding transformation rule andbased on semantic dependency analysis table, basic model will be generated. The contextmodel is a semantic space which is generated through related concepts semantic similaritymatching and semantic extension in DKSRO. Rule model generation is based on SWRLcollection of DKSRO.(4)Research on semantic knowledge service with SWRL reasoning and directed graphtheory. Aiming at the problems of low knowledge utilization efficiency in present productdesign knowledge management system due to it cannot handle semantic information inknowledge retrieval and pushing, a knowledge service method with SWRL reasoning anddirected graph theory is proposed based on the product design semantic model this paperproposed. Firstly, through the service demand model and PDKSUM matching determine thescope of the candidate semantic model unit. Then the candidate set of semantic unit will beconstituted to a connected directed graph through PDKSUM rules model reasoning. At last,the knowledge service solution is generated using the topology sequence of directed graph. (5)Research on product design knowledge service system based on semantic model.Combining with the results of the above key technologies, a framework of product designknowledge service system based on semantic model is designed. In this paper, the processand the function module of the system is introduced. Product design knowledge servicesystem is developed based on Java and Ext technologies. Finally, a military vehicle walkingsystem design process of is applied in this system.