Integration of Multivariate Control Charts and Neural Networks to Determine the Faults of Quality Characteristic(s) in a Multivariate Process

Abstract

Because of advanced technology, there are many aspects of quality characteristics in a product. Typical univariate statistical process control ( SPC) charts may not be suitable for monitoring processes that have multiple quality characteristics. As a consequence, the multivariate control charts are developed to simultaneously monitor multiple quality characteristics of a process. Like the function of a univariate SPC chart, the process is hypothesized to be out of control when a signal is triggered by a multivariate SPC chart. The problem is that, it is difficult to interpret the signal for a multivariate SPC chart due to the multiple quality characteristics of a process That is, which quality characteristic(s) is (are) attributed to this out-of-control signal. If the characteristic(s) that is (are) at fault can be quickly and correctly determined, the corresponding remedial actions can be taken to tune the process in time. Therefore, this identification is a very important issue for industry processes.