Monitoramento do controle estatístico do processo utilizando ferramentas estatísticas

Abstract

This study used the data from a tobacco company that monitors its production process through statistical process control. The objective of this study was to apply multivariate statistical tools of Statistical Process Control to monitor the efficiency of the production process of the company under study. This production process consists of four variables monitored during the manufacture of cigarettes: RDT (Resistence to Draw), weight, ventilation and circunference. These data were verified using descriptive statistics, test of normality, univariate R and X-bar charts and correlation matrix. After observing the instability of the system, the Principal Component Analysis (PCA) was applied. PCA is an exploratory multivariate analysis technique that aims at gathering various features of the significant variables in the instability of the production process. EWMA was then applied on the principal components. An unstable process presenting too much variability was observed, with several points outside the bounds of statistical control. This situation presents some problems that may cause inconvenience and loss to the company. This way, the company should review its manufacturing process in order to improve its productivity and quality that in turn may help the company to be more competitive.