Research On Construction Methods Of Knowledge Base And Intelligent Retrieval Platform In Modeling For CNC Machine Tools

CNC machine tools are the core equipment of advanced manufacturing.The functional design with technology as the main body is sufficient,by comparison,the research on the modeling design based on the user's kansei needs is obviously insufficient.In this paper,the theory of kansei engineering and artificial intelligence is comprehensively used to study the intelligent modeling design of CNC machine tool.The product modeling is quantified according to the mapping relationship between modeling feature elements and modeling styles;the ontology model and modeling instance database are built based on modeling domain knowledge;based on the database,the modeling intelligence retrieval platform is constructed.The platform solves the problem that the user's kansei needs are difficult to quantify,and provides an objective and quantitative theoretical basis for modeling design.Meanwhile,the designer can quickly and accurately obtain the proper CNC machine tool shape that meets the user's kansei needs,as an example for the user-centered modeling design,which is very significant for shortening the design cycle and reducing resource waste.Firstly,the modeling of CNC machine tools is quantified using theories of kansei engineering and mathematical modeling.The CNC machine tool samples and perceptual vocabularies are collected fully using literature search and research methods,then the representative samples and vocabularies are selected.Vocabulary MDS model is established by cluster experiments,and the modeling styles of CNC machine tools are obtained by cluster analysis of vocabulary.The modeling feature elements of machine tools are extracted by analytic hierarchy process,then the level of each feature elements is obtained using morphological analysis.Finally,the mathematical models between modeling styles and feature elements are established using joint analysis and BP neural network.Above all,the quantification of CNC machine tool modeling is achieved in this part,as an important data foundation for database development.The ontology model of CNC machine tools and modeling instance database are built and the related semantic measurement methods are studied based on the theory of ontology modeling and database development.The modeling domain knowledge of CNC machine tools is studied,and the hierarchical structure of domain knowledge is determined,then the glossary is formed as the content basis of ontology modeling and database development.By the seven-step modeling method,the class set,property set and relationship set of CNCO are defined,and the ontology model is implemented by Protégé.Based on domain knowledge,the table and data dictionary of the database are determined,then the database function is designed,and the database is built by MySQL.The connection between modeling instance database and CNCO is implemented by developing JDBC interface,through OWL based The connection method in detail.Meanwhile,the semantic metrics of open vocabulary are studied,and the WordNet based semantic mapping mechanism is determined,and the toolkit is implemented in Java.The ontology model and modeling instance database are important data foundations for the retrieval platform.Taking the modeling instance database as the underlying database,an intelligent modeling retrieval platform for CNC machine tools is built.The function positioning of the platform is studied,and the content and method of input information are determined according to the user's knowledge of the domain knowledge.The information retrieval mechanism of the platform is designed,and the open vocabulary is processed by the Sem reasoning model,then the modeling comprehensive scores are obtained by using the modeling comprehensive scoring model,as the standard of output samples.Based on the above information,the architecture of the intelligent retrieval platform is designed,including the underlying knowledge base and application system.The architecture is achieved in Java and the GUI is achieved by Jigloo,then the platform and the modeling instance database are connected.The superiority of the platform is studied.In detail,the platform error is divided into retrieval error and semantic measurement error.The error is measured and evaluated by using the precision,recall and RMSE indicators.The platform can realize the intelligent modeling retrieval of CNC machine tools based on user's kansei needs well.