Online Monitoring and Fault Diagnosis of Multivariate-attribute Process Mean UsingNeural Networks and Discriminant Analysis Technique

In some statistical process control applications, the process data are not Normally distributed andcharacterized by the combination of both variable and attributes quality characteristics. Despitedifferent methods which are proposed separately for monitoring multivariate and multi-attributeprocesses, only few methods are available in the literature for monitoring multivariate-attributeprocesses. In this paper, we develop discriminant analysis technique for monitoring the mean vector ofcorrelated multivariate-attribute quality characteristics in the first module. Then in the second module,a novelty approach based on the combination of artificial neural network (ANN) and discriminantanalysis is proposed for detecting different mean shifts. The proposed approach is also able to diagnosequality characteristic(s) responsible for out-of-control signals after detecting different step mean shifts.A numerical example based on simulation is given to evaluate the performance of the proposedmethods for detection and diagnosis purposes. The detecting performance of the second module is alsocompared with the extended T2 control chart and with the extension of an ANN in the literature. Theresults confirm that the proposed method outperforms both methods.

Keywords:

Discriminant Analysis , Multivariate-attribute , NORTA Inverse , Neural Network , Fault Detection , Fault Diagnosis