Multivariate statistical process control (MSPC) is applied to an electrolysis process. The process produces extremely pure copper, and to monitor its quality the levels of eight metal impurities were recorded twice a day. These quality data are analysed adopting an (1) 'intuitive' univariate approach, and (2) with multivariate techniques. It is demonstrated that the univariate analysis gives confusing results with regards to outlier detection, while the multivariate approach identifies two types of outliers. Moreover, it is shown how the results from the multivariate principal component analysis (PCA) method can be displayed graphically in multivariate control charts, Multivariate Shewhart, cumulative sum (CUSUM) and exponentially weighted moving average (EWMA) control charts are used and compared. Also, an informationally powerful control chart, the simultaneous scores monitoring and residual tracking (SMART) chart, is introduced and used. (C) 1998 Elsevier Science B.V. All rights reserved.

Multivariate time series analysis is applied to understand and model the dynamics of an electrolytic process manufacturing copper. Here, eight metal impurities were measured, twice daily, over a period of one year, to characterize the quality of the copper. In the data analysis, these eight variables were summarized by means of principal component analysis (PCA). Two principal component (PC) scores were sufficient to well summarize the eight measured variables (R2=0.67). Subsequently, the dynamics of these PC-scores (latent variables) were investigated using multivariate time series analysis, i.e., partial least squares (PLS) modelling of the lagged latent variables. Stochastic models of the auto-regressive moving average (ARMA) family were appropriate for both PC-scores. Hence, the dynamics of both scores make the exponentially weighted moving average (EWMA) control chart suitable for process monitoring.

In this tutorial article, an introduction to the basics of mixture design is provided. The discussion on mixture design is made from the experimenter's point of view, rather than from an elaborate theoretical perspective. Much emphasis is placed on `how to think' when defining a mixture problem, and analyzing the resulting data. Also, a working strategy for mixture experimentation is proposed. This strategy relies heavily on important contributions from the chemometrics research field, notably the impact of the chemometric `modelling' philosophy and its multivariate data analytical tools. In order to illustrate the introduced working strategy, two-mixture applications are outlined. The first example deals with tablet manufacturing and, partly, reflects the classical approach to mixture design based on regular mixture regions. The second example concerns the preparation of a good bubble mixture from which children may produce long-lasting bubbles. This application represents a contemporary chemometrics approach to mixture design when (1) the experimental region is irregular and (2) process and mixture factors are varied within the same experimental protocol.

Immense amounts of data are collected into today's modern process monitoring systems. There are, however, few methods that have the capability to grasp the essentials in these, usually heavily correlated, data. The multivariate statistical techniques, principal components (PC) modelling and modelling by projection to latent structures (PLS) are two methods that have a great potential for process monitoring and forecasting in these situations

Treize systemes chromatographiques en couche mince sont utilises pour caracteriser seize monosaccharides (aldoses et aminoglycosides). Utilisation de l'analyse a composantes principales

Previously derived principal property scales for the 20 coded amino acids have been extended to six noncoded amino acids, namely ornithine, norvaline, norleucine, O-methylated threonine, α-amino butyric acid, and citrulline. This has been done first, for the 20 coded amino acids, by developing PLS models for the relations between the previously derived scales (z1, z2, and z3) and newly measured thin-layer chromatography data, proton nuclear magnetic resonance data, plus literature data describing amino acid "bulk" (van der Waals volume and molecular mass). Thereafter, z values for the noncoded amino acids were calculated by inserting the same measured data for these compounds into the PLS models. As a validation, the calculated z values were then used to describe the structural variation of a set of 60 oxytocin peptide analogues for which literature data on biological measurements existed. The models explain 71–81% of the variance of the biological data and are highly significant according to cross-validation. Prediction errors of about twice the estimated biological measurement error are obtained for the 11 peptide analogues containing noncoded amino acids.

In multivariate data analysis such as principal components analysis (PCA) and projections to latent structures (PLS), it is essential that the training set systems (objects) are selected to provide data with substantial information for model parametrization, and to represent properly any future situations where the multilvariate model is used for predictions. In the framework of multivariate projections (PCA, SIMCA and PLS), elementary concepts of statistical design (fractional factorials and composite designs) can be used with the latent variables (PC or PLS scores) as design variables. The plan of action thus becomes: (1) problem formulation (specify aim and model, make a conceptual division of the investigated system into subsystems); (2) collection of multivariate data for each type of subsystems; (3) estimation of the practical dimensionality of the data for each type of subsystems by PC or PLS analysis; (4) use of the PC or PLS scores (t) as design variables in the combination of subsystems to systems in the training set; (5) measurement of responses (Y); (6) analysis of data by PCA or PLS; (7) interpretation of results with possible feedback to steps 1, 2 or 3. The procedures are illustrated by two problems: a structure/activity relationship for a family of peptides, and optimization of an organic synthesis with respect to system variables (solvent, substrate, co-reactant_) and process variables (temperature, reactant concentrations).