Research On Knowledge Push Technolo- Gy And Application For The Overall Scheme Design Of CNC Machine Tools Driven By Design Process

The overall scheme design is the basis for the design and development of CNC machine tools, but low efficiency of knowledge acquisition in the design process has become a key factor restricting of improving the design quality and shortening the design cycle of it. To solve these problems that the heterogeneous design knowledge of CNC machine tools is difficult to express using an unified model, the current knowledge push methods are departing from design process, and the traditional methods have low accuracy rate and recall rate, this paper has studied knowledge push technology for the overall scheme design of CNC machine tools driven by design process. Combined with major national science and technology project, the CNC machine tools overall scheme design knowledge push system was developed, which has achieved high efficiency design knowledge push. Several examples of knowledge push during overall scheme design of VDL600E and TGK46100 were given, which verified the feasibility of methods described above.The main content of this thesis as follows:The first chapter introduced the research background and significance of this paper, reviewed the research status of product design knowledge representation, product design process modeling and design knowledge push mode. Current problems of CNC machine tools design knowledge push were proposed, which determined the contents of this paper, and the research framework was given.The second chapter presented a method of modeling and evolution for the heterogeneous knowledge of CNC machine tools based on matter-element theory. Analyzed the characteristics and divided the matter-element of CNC machine tools design knowledge. The matter-element of CNC machine tools design knowledge was modeled, which has been described with specific examples. The superiority assessment system was built to assess the superiority of CNC machine tools design knowledge, and based on the superiority to achieve evolution of design knowledge, which has been described with specific examples.The third chapter presented a construction method of overall scheme design knowledge push engine based on Petri-net model. Analyzed the process and built the context scenes of overall scheme design. The workflow model of overall scheme design was built using Petri-net, generating process definition file. On this basis, structured design knowledge needs were extracted based on the context scenes and the execution and routing logic algorithms were designed. This chapter realized knowledge push driven by overall scheme design process.The fourth chapter presented a matching optimization method of CNC machine tools design knowledge based on SVD. Generated descriptors-design knowledge matrix, ameliorated the TF-IDF weight calculation method bye nonlinear function, simultaneously, considered the occurrence of descriptors in knowledge model, the improved weight calculation method has been proposed. Converted descriptors-design knowledge matrix into latent semantic space and reduced the dimension of it. The matching calculation and sorting filtration method of search conditions with design knowledge was given, and extraction algorithm of the user knowledge behavior was designed, by which the user knowledge model was built, thereby achieving secondary filter of knowledge, enabling personalized push. The effectiveness of this method was verified by algorithm analysis experiments.The fifth chapter developed the CNC machine tools overall scheme design knowledge push system according to the theory and methods of this paper. The architecture and function modules of the system were introduced and several examples were taken to verify the feasibility of this system.The sixth chapter summarized the main contents of the paper, and the follow-up studies were discussed.