Cause Analysis of Chronic Defects in Aluminum Rolling Process

Modelling by Bayesian Network

Abstract

The purpose of this study is to develop approach for cause analysis of chronic defects in aluminum rolling process which is combination of data-based method and knowledge-based method. In aluminum rolling process, defects can be classified into two types. One is a sporadic defect, which can be controlled by monitoring changing points, and the other is a chronic defect, which is difficult to be controlled. To combine a data-based method and a knowledge-based method, Bayesian networks are used to represent chronic defect mechanism.