The Defect Analysis For The Surfacial Quality Of The Cooling System Based On Data Mining Technology

Recently, Steel and Iron enterprises are working hard to find a way to reduce cost,produce high value added products and improve their quality. Cold-rolled picklingproducts play an important part in the automotive industry, Machinery industry andlight home appliances industry. However, the manufacturing of cold-rolled picklingproducts are still on the initial stage in Steel and Iron enterprises with a high ratio ofexternal failure. To meet the needs of market as well as enterprise development, it isimperative to improve the quality of cold-rolled pickling products and marketcompetitiveness. This topic is based on the quality defects of these products from acertain Steel and Iron enterprise so as to analyze the causes of the quality defects andthen finding a way to improve the external quality and reduce the cost.The manufacturing process of cold-rolled pickling products are widely involved,steelmaking, hot rolling and pickling included. Therefore trying to improve the qualityof products through the key process has its limitations. Moreover there may be differentfactors affecting the quality of products at different conditions. The main content of theresearch in this paper is to analyze quality defects of cold-rolled pickling productsfrom a certain Steel and Iron enterprise by means of data mining technology as well asdata mining models such as Binary Classification Tree, Logistic Regression andMulti-section Decision Tree provided by SAS Enterprise Miner tool. Then the mostsuitable model Multi-section Decision Tree are selected by scoring tool provided bySAS Enterprise Miner as well as the actual situation while manufacturing and applied toanalyze external quality defects of cold-rolled pickling products. Based on the analysis,the factors affecting external quality defects of cold-rolled pickling products areconcluded as follows: Average pickling speed, the width of the export of materials aswell as the thickness of the export of materials. This paper sums up the generationformula of unqualified products by data mining all these factors. At the same time theresults can be used as guidance to manufacturing process to achieve the goals ofimproving the external quality of cold-rolled pickling products. This paper appliedSAS Enterprise Miner to analyze external quality defects of cold-rolled picklingproducts and it turns out to have good effect.