Track: Quality Engineering, Control and Management

Abstract

Profile monitoring is usually performed by establishing control charts. In most of the cases, the in-control values of the profile parameters are assumed to be known in Phase II, whereas it is not valid in many practical situations. In this article, we investigate the effect of parameters estimation from in-control Phase I samples on the in-control and out-of-control performance of two Phase II control charts for monitoring multivariate multiple linear profiles designated as  and . The out-of-control performance of the methods is evaluated by using corrected limits to consider the variability due to parameters estimation. The performance of the monitoring approaches is compared in terms of statistical properties of ARL distribution including AARL, SDARL and CVARL in order to consider practitioner-to-practitioner variability trough a Monte Carlo simulation algorithm named as ARLS. The results showed that parameters estimation severely effects on the performance of the monitoring schemes.