|
Abstract:
|
Much of multivariate process control has focused on monitoring product/process output characteristics. However, the output variables from a process are influenced by the levels of the input or regulatory variables and uncontrollable factors. Thus, traditional approaches may not identify which of the causal factors, i.e., process regulatory variables or uncontrollable noise factors or both, are possible reasons for out-of-control signals. Further, when the effect of the process regulatory variables and the noise factors cancel each other, the impact on the output response variables may not be detectable. Here, we consider the impact of both process regulatory variables and noise factors through an adjusted measure. The performance of the proposed method is determined, through simulation, by the average run length of first detection of out-of-control conditions.
|
