Multivariate quality control of batch processes using STATIS

Batch process control and monitoring: a Dual STATIS and Parallel Coordinates (DS-PC) approach

“…; I K observations (individuals) and J variables; each one of these represents different studies, scenarios or batches. Raw tables must be preprocessed using several methods of centering, scaling, and normalizing data as shown in Tucker (1964), Westerhuis, Kourti, and MacGregor (1998), Naes, Brockhoff, and Tomic (2010), Abdi et al (2012) and Filho and Luna (2015).…”

“…Statistical projection methods, such as principal component analysis (PCA) and partial least squares (PLS) ideally provide, at some degree of confidence, a representation of the correlation structure within data (Lewis, 2014). As for our concern, approaches to batch monitoring are mostly grounded on multiway partial least squares (MPLS) and multiway principal component analysis (MPCA) (Filho & Luna, 2015;Kourti, 2003). Additional issues of the multivariate version refer to its limitation with dynamic variables' behavior of batch processes and real-time monitoring of ongoing batches (Filho & Luna, 2015).…”

“…As for our concern, approaches to batch monitoring are mostly grounded on multiway partial least squares (MPLS) and multiway principal component analysis (MPCA) (Filho & Luna, 2015;Kourti, 2003). Additional issues of the multivariate version refer to its limitation with dynamic variables' behavior of batch processes and real-time monitoring of ongoing batches (Filho & Luna, 2015).…”

“…Its pioneer use in quality control was orientated toward off-line monitoring (Scepi, 2002). In recent contributions, STATIS has been used in off-line and on-line monitoring taking into account the use of IS and CO CCs for batches of fixed time duration and its dual version for variable duration (Filho & Luna, 2015;Niang et al, 2013;Niang, Fogliatto, & Saporta, 2009). These STATIS-based approaches have carried out batch and time-wise analysis, including bivariate boxplots, based on convex hull peeling and B-spline smoothing for CCs' construction, under nonparametric conditions.…”

“…These STATIS-based approaches have carried out batch and time-wise analysis, including bivariate boxplots, based on convex hull peeling and B-spline smoothing for CCs' construction, under nonparametric conditions. However, none complementary analysis was conducted to examine behavior within batches per variable (Filho & Luna, 2015;Niang et al, 2009).…”