Application of multivariate statistical process control in sand preparation process

Based on the analysis of characteristics of sand preparation process, an improved PCA modeling method has been proposed, and a multivariate statistical process (MSPC) model to detect individual abnormity sample in modeling data has been set up according to the rules of the variation squared prediction error (SPE) statistics. The rules of the variation SPE statistics and T2 statistics and variance increasing of process variables can manifest on the SPE chart and T2 chart, which have been used to monitor the variations of the real process.