On-line Analysis Out-of-Control Signals for Multivariate Control Chart Using Neural Network

Abstract

It is common in industrial process to monitor several correlated quality variables simultaneously. Most of multivariate quality control charts
are effective in detecting out-of-control signals based upon an overall statistics in multivariate manufacturing processes. The main problem of such
charts is that they can detect an out-of-control event but do not directly determine which variable or group of variables has caused the out-of-control
signal and what is the magnitude of out of control. This study presents an artificial neural network-based model to supplement the multivariate χ²
chart. This method consists of two modules. In the first module using a general-neural network, type of unnatural pattern can be recognized. Then
by using two special-neural networks for shift mean and trend, it can be recognized magnitude of mean shift and slope of trend for each variable
simultaneously. The performance of the proposed approach has been evaluated using a simulated example. The results confirm that the proposed
method provides an excellent rate of classification and the output generated by trained network is strongly correlated with the corresponding actual
target value for each quality characteristic.