Research On Transfer Mechanism And Evolution Motivation For Empirical Engineering Knowledge

With the explosion of technological reform and knowledge creation among exisiting concepts,priciples,methodologies and technologies in various engineering domains,engineering experience shifts frequently between the technology background and application scenarios,and long evolutional history of engineering knolwdge is more impossible for the individual to cognize.Therefore,in order to improve the performance of knowledge management and creation in the knowledge-intensive enterprises,it's urgent to offer an efficient and timely management for the enterprises to organize their engineering experience,and offer a beneficial and powerful support for the engineers to facilitate their innovations.In the related researches of the management of empirical engineering knowledge(EEK),it is difficult to accurately characterize the dynamic contents in EEK,effectively promote the transfer of EEK between multiple situations,and comprehensively grasp the evolutional laws in the whole life cycle of EEK.To rectify this,based on the field of computer-aided innovative R&D of mechanical products,this dissertation separately focuses on the “individual and short-term” process of EEK transfer and the “industrial and long-term” process of EEK evolution.By recognizing features,presenting processes and analyzing motivations,what,how and why EEK changes are answered in short-and-long terms.This dissertation also develops and applies a prototype system for an actual situation of knowledge creation,in order to verify the proposed theories and methods.The contents of this dissertation include the following four aspects:(1)Formal representation and sementical association of EEK from a dynamic aspectsAnalyzing the natural language features of the texts of EEK,this dissertation proposes an EEK representation method with seven attributes of engineering problem,problem context,problem solution,effectiveness,contributor,time,and feature association.Applying semantic similarity calculation and two-parameter time windows,a dynamic organization method is proposed,and the relevance among EEKs is evaluated.(2)Mechanism construction on EEK transfer under technological paradigm shiftUsing eye-tracking technology,this dissertation designs a two-phase experiment for the transfer of EEK,and investigates three factors influencing the performance of transfer,namely the mastery of the prior paradigm,the self-directed learning willingness,and the distributions of visual attention on the learning materials.Based on the four key cognitive operators in “Concept-Knowledge” theory,an operational mechanism for the EEK transfer under technological paradigm shift is constructed.(3)Long-term evolution courses excavation and motivation analysis of EEK accumulated by groupsApplying fuzzy evaluation,clustering algorithm,and topic elicitation,EEK networks are established to present the structure of the engineering domain in a certain period of time,and a modeling method is proposed to portray the evolutional courses among topics in the different networks.Utilizing feature extrcation,event correlation,and abductive reasoning,key phenomena in EEK evolution are recognized.A framework for analyzing the evolutional motivation is proposed,based on the linkage and possible casual relations of the milestones in the domain.(4)Empirical research on EEK creation supported by the mechanisms of EEK transfer and evolutionBased on the database and corpus in the field of computer-aided design,all the methods proposed in this dissertation are implemented to construct a prototype system for EEK management,and the specific application for promoting the transfer of EEK are designed.A comparative experiment is carried out with a case of high-precision 3D surface modeling of a ship propeller,which verifies the effectiveness and superiority of the prototype system in farcilitating novel EEK creation.