Research On Association Rule Mining Method For Surface Defect Data Of Cold Rolled Strip

In recent years,industrial data has been widely used in the various links of the industrial chain,and industrial data mining is a technology which is used to analyze the relationship between the data.The cold rolling strip surface defect data processing method is mainly through visual identification,The cold rolling strip surface defect data processing method is mainly through visual identification,extraction of defect features and classification of defects pushed to decision makers,There is little research on the relationship between mining defects and defects,defects and causes,causes and causes.In present period,the treatment methods of enterprises are generally based on artificial analysis.This method is not only inefficient,but also the analysis result is easy to make mistakes.Aiming at these problems,In this dissertation,industrial data mining technology is applied in the analysis of cold steel strip surface defect data relationship,the association rules mining model of cold rolling strip surface defect data is established and the relevant system is designed by using Visual C++.In this dissertation,the main research work is as follows:(1)This dissertation describes the application of large industrial data in the field and made an in-depth analysis of it.In view of the fact that the product defect data can not be effectively used in cold rolled strip products.In this dissertation,we first establish a data mining algorithm based on historical statistics and expert experience data to provide theoretical support for the association rules mining algorithm.Then,the information source of iron and steel production process is analyzed,and the theoretical model is applied to the data mining.(2)The fuzzy frequent itemsets mining algorithms mainly is the improvement of the classical Apriori algorithm.But,this algorithm has inherent defects of multiple scans of the data set,and it can not solve the data mining problems of complex cold rolled strip product defects.So this dissertation proposes the weighted data of surface defects of fuzzy hierarchy association rules mining algorithm.(3)Aiming at the problem that the association rules of cold rolled strip surface defect data can not be organized effectively after data mining,In this dissertation,a fuzzy classification method of product defect data association rules based on item attribute difference is proposed.Based on the defect data,the fuzzy classification tree is used to get the distance between rules,and the clustering analysis is carried out.The similarity of the larger association rules are divided into a class,it is convenient to decision makers to analyze the results of data mining.(4)Combined with the research of association rules mining algorithm and association rules classification algorithm,the dissertation designs the association rules mining system for the cold rolling strip surface defect data.The system is mainly divided into six modules: user permissions,data import,online evaluation,association rules mining,association rules classification,results visualization.