A Simple and Practical Method for Monitoring Dimensional Quality in a Multivariate Process

摘要

In this paper we propose a simple and practical method for monitoring a multivariate process. This method is used for monitoring the dimensional quality of a door panel in the automotive industry. In order to detect out-of-control signals, we use a T{sup}2 control chart. When the T{sup}2 statistic shows a point out of control, we recommend using a principal components contribution chart and a variables contribution chart. These two charts will show which variables and/or correlations are responsible for the signal. Besides, the principal components contribution chart and the interpretation of the principal components provide information about the dimensional structure of the panel, so that this structure confirms the out-of-control variables and correlations and used to take corrective actions in the process.