Research On Analogy-driven Product Innovative Design Based On Structure Mapping-based Function Model

Under economic globalization,by continuous product innovation can enterprises remain invincible in increasingly fierce market competition.Conceptual design is the critical stage for product innovation.Designing innovative product needs guidance from good methods.As a cognitive skill,analogical reasoning help designers identify solution knowledge of cross-domain similar problems as design inspiration,thereby being important for innovative concepts generation.This gives the birth of analogical reasoning-based design,i.e.,analogical design.Design community has extensively researched main parts of analogical design,including knowledge representation,retrieval and transfer.Although research results seem to be zfruitful,the involved knowledge representation models are not really suitable for analogical design task for not fully considering structure mapping mechanism of analogical reasoning,so that the design analogy retrieval and analogical transfer supported by them are implemented in a relatively single way and not expressive and flexible enough.In order to make full use of analogical reasoning in innovation design,a knowledge representation model which can overcome the above shortcomings is needed for constructing analogical design method/tool.To address this need,this thesis,considering the existing function models and relevant cognitive laws,attempts to design a knowledge representation model close to structure mapping mechanism of analogical reasoning,then studies analogy retrieval and transferring,and finally gets an analogical design method/tool to support mechanical product innovation design.The main content is as following:1.A classification-,clustering-,and benchmarking analysis-based framework,for identifying appropriate function modelling methods to support target engineering tasks.is constructed systematically grasping characteristics of the function models available.2.Aimed at the engineering task of analogical design of mechanical products,a knowledge representation model close to structure mapping mechanism of analogical reasoning,i.e.,structure mapping-based function model(SM-FM),is designed applying the function modelling method identification framework.3.Establishing bridge from analogical knowledge representation model to analogy retrieval.Ontology is used as the knowledge representation strategy for machine understanding.The SM-FM method-based domain ontology model and ontology knowledge base are built utilizing Protégé 5.5.0 and ontology web language.4.A technology supporting functional relation similarity-or functional causality relation similarity-based design analogy retrieval is researched.Function or functional causality relation analogy query is expanded by the SM-FM ontology-based semantic similarity evaluation.Analogy retrieval model is constructed based on vector space model.The elements of query and document vectors are weighted using term frequency-inverse document frequency.When computing the cosine distance between query and document vectors for knowledge matching and ranking,the technology adopts three-scale analytical hierarchy process to weight different parts of cosine similarity respectively decided by metafunction,function and flow terms,so that analogies can be retrieved according to the type of analogy query.5.Following the rational unified process and incremental model and using SQL Server 2016,“Protégé 5.5.0+Jena 3.6.0+Jess 7.1” and Eclipse 4.18.0/Java respectively as database,ontology and programming platforms,an analogical reasoning-based computer-aided conceptual design(AR-CACD)system prototype is constructed in Windows 10.Aided by it,an iterative analogical design process model is constructed,and applied into the innovation design of pipeline detection robot Pipe Guard.