Principal component analysis applied to the statistical control of multivariate processes

Abstract

Objective: Propose the analysis and monitoring of a chemical process sustained in the theoretical principles of a factorial method cataloged as principal component analysis (PCA), whose ultimate objective is to represent the original variables of the system. Methodology: Methodologically the data was analyzed in a more compact dimensional space, under the hypothesis of multivariate normality. A control chart based on the error prediction squares is constructed in the later phase to monitor the behavior of the successor variables once the ACP has been applied. Results: The results of the univariate control charts constructed from the t-scores of the individual components, in which the behavior of the variables contributing to the generation of alarm signals is exhibited, can be used as the basis for an auxiliary analysis based on the results provided by the univariate control charts. Conclusion: The results obtained corroborate that the three retained components explain a majority fraction of the original variability of the data cloud and that the control chart constructed from the sub-dimensions registers the existence of outliers or extreme values.