Elsevier, Chinese Journal of Chemical Engineering, 6(20), p. 1121-1127
DOI: 10.1016/s1004-9541(12)60596-5
Full text: Download
Multi-way principal component analysis (MPCA) is the most widely utilized multivariate statistical process control method for batch processes. Previous research on MPCA has commonly agreed that it is not a suitable method for multiphase batch process analysis. In this paper, abundant phase information is revealed by way of partitioning MPCA model, and a new phase identification method based on global dynamic information is proposed. The application to injection molding shows that it is a feasible and effective method for multiphase batch process knowledge understanding, phase division and process monitoring.