Research On Knowledge Discovery Method For Design Knowledge Of Mechanical Structure Symmetry

Structural symmetry is a common phenomenon in mechanical system and plays important roles.Summarizing the design knowledge of mechanical structure symmetry systematically can help to direct the scientific application of structure symmetry in mechanical products design.Massive design knowledge can be mined from design instances of structure symmetry using the knowledge discovery method,nowadays the research on the knowledge discovery method for design knowledge of mechanical structure symmetry still need to be improved.Aiming at the limitations of the research on the knowledge discovery method for design knowledge of mechanical structure symmetry,this article conducted the studies in the design knowledge model,design knowledge discovery and mining method and design knowledge application.The author and other peoples in the research group have collected a lot of mechanical structure symmetry design instances,analysed the existing method,functions,application rule and application method of structure symmetry in design instances.Based on the above research,a design knowledge model for mechanical structure symmetry was built,the model has many advantages,such as multi-dimensions,multi-views,static and dynamic states combination,and can completely show the effects of structure symmetry in achieving the mechanical product design demands.By aiming at the knowledge discovery method of mechanical structure symmetry,the author proposed a multidimensional fuzzy association rules mining method.This method is more general,and can effectively avoid the disadvantage of region boundary division.By using this method,rich practical design knowledge can be mined easy.An improved Eclat algorithm for mining association rules called Eclat_growth was built to adapt the characteristics of the database of mechanical structure symmetry instances.Comparing with the classical Eclat algorithms,the Eclat_growth algorithm has the highest performance in mining association rules from various databases with different mount of transactions,different mount of attributes and different denseness.By applying above research results,the existing computer-aided knowledge discovery platform was improved and the structure symmetry design instances database was mined,the design knowledge and application method in the symmetry degree changes of structure symmetry were proposed.Finally,a summarization for the research of this article was conducted,some prospects of further research on design knowledge model of mechanical structure symmetry,the data mining methods and the knowledge discovery platform were discussed.