B. Yuan et Xz. Wang, Multilevel PCA and inductive learning for knowledge extraction from operational data of batch processes, CHEM ENG CO, 185, 2001, pp. 201-221
A new methodology for monitoring batch processes is presented which is base
d on analysis of historical operational data using both principal component
analysis (PCA) and inductive learning. Historical data of batch operations
are analysed according to stages. For each stage, PCA is employed to analy
se the trajectories of each variable over all batch runs and groups the tra
jectories into clusters. The first one or two PCs for all variables at a st
age are then used in further PCA analysis to project the operation of the s
tage onto operational spaces. Production rules are generated to summarise t
he operational routes to produce product recipes, and to describe variables
' contributions to stage-wise state spaces. A method for automatic identifi
cation of stages using wavelet multi-scale analysis is also described. The
methodology is illustrated by reference to a case study of a semi-batch pol
ymerisation reactor.