Journal of Quality Engineering and Management

Abstract Control charts are powerful tools used for monitoring process variations. In statistical process control, there are many situations in which qualitative (attribute) characteristics of a product or process are monitored simultaneously. Although an out-of-control signal in control charts indicates the presence of a process change, the exact time at which the change occurs is often unknown. Identifying the precise change-point helps process engineers determine the causes of variation and improve the process.
In this study, statistical process control techniques are examined for a process in which qualitative attribute characteristics of the conforming/nonconforming type are measured in a bivariate form. Using the maximum likelihood method, we propose an estimator for determining the change-point in such processes. It is assumed that the two qualitative attribute characteristics in a product are correlated. Simulation results show that the proposed method performs well in detecting the change-point in the process mean vector.





 




 

Abstract Nowadays, many products are the outputs of multi-stage processes. In such processes, the stages are often interdependent, meaning that the quality of the product in a particular stage depends not only on the quality in the current stage but also on the quality of the product in the previous stages. This phenomenon is referred to as the cascading property of multi-stage processes. The existence of this property and the lack of attention to it can lead to errors in interpreting the control charts used in the stages. Therefore, in the literature on multi-stage process monitoring, methods have been proposed to reduce or eliminate this problem. On the other hand, in some cases, the quality of a product is described by the relationship between a response variable and one or more independent variables, known as a profile, which can be the result and output of a multi-stage process. Since fewer studies have been conducted on monitoring profiles resulting from multi-stage processes, this paper investigates the effect of the cascading property on the monitoring of simple linear profiles in a two-stage process, based on the average run length criterion using simulation, and also examines the effect of inter-stage dependency on the estimation of the profile parameters of the second stage.