Research On Process Knowledge Discovering And Reusing Based On Data Mining

As the key of process design,process route designing has influence on manufacturing cost and quality,product lifecycle,the designing level process route is a performance of competitive power of enterprise.Facing with the problem,called rich process data but lacking knowledge,and the eager demand that improving the level of process designing for enterprise,it is an effective for process knowledge discovery and reusing to deal with these problems.This paper uses data mining to discovering the knowledge hiding behind process data based on process route data as research object from a procedure perspective,and achieves the process designing with assisting of technology of reusing process knowledge by machine learning and intelligent optimization algorithm.First the paper proposes a method that obtaining typical process route by clustering partition algorithm.It is called GKMBCNN that means Global K-Medoids Based on Competitive Neural Network by taking full advantages of high robustness and global convergence of Global K-Medoids,of course parallel computing ability of Competitive Neural Network,which has been proven the effectiveness on big data clustering problem by testing UCI database.Aiming at obtaining typical process route,this paper carried out a process route encoding scheme including procedure and parameter encoding,and a multilevel similarity scheme forit,and also a method that describes the text sequence data such as process data.It was proved that it is effective on obtaining typical process route based on GKMBCNN in the results.From the procedure perspective,it makes use of Apriori algorithm to discovering rule of selecting benchmarks and sorting process sequence,relationship between process method and the demand of processing and so on,which is called process decision knowledge.All the content of mining has been explained in a specific case based on axis process data.At last,a method has been carried out that it is a way to automatically generating and optimizing process route considering reusing process knowledge.This paper analyzes and selects 8 main factors influences the process chain of character surface to modeling for the process chain of character surface based on SVM.It predicts the reasonable process chain for surface according to this model,and establishes the constraint matrix in accordance with machining element model and the process rule.Then it uses the BBEA(Block Based on Evolutionary Algorithm)to solve the problem of automatically generating process route for component focusing on minimize changing numbers of machining resources,and has been illustrated by generating process route rod bolt for rod bolt.