Intelligent software sensors and process prediction for glass container forming processes based on multivariate statistical process control techniques

Glass container forming processes have attracted more attention over the past years due to the problem of lacking process information and correlation for key variables within the processes. In this paper an approach to develop process modeling and intelligent software sensing is presented for application based on multivariate statistical process control methods. The intelligent software sensors are able to provide real time estimation of key variables, and Partial Least Squares (PLS) techniques have allowed for forward prediction of final product quality variables. An application of software sensors used for container forming blank temperature is presented along with PLS being applied to predict the wall and base dimensions of glass container products. Initial results show that these methods are very promising in providing a significant improvement within this area which is usually unmonitored and is susceptible to long time delays between forming and quality inspection.