EWMA monitoring schemes for MCV in short production runs with an application to the steel sleeve manufacturing process

Abstract

In industries, multi-variety and small batch processes are extremely common due to the fact that the production process is moving towards flexible manufacturing to meet the increasing demand of consumers for individualized products. In recent years, monitoring techniques based on the multivariate coefficients of variation (MCV) have been widely developed owing to their applicability in assessing the relative variability of multivariate processes. It is worth noting that few of these studies have targeted monitoring the MCV in a short production run context. In this paper, two new one-sided MCV monitoring schemes are proposed by adopting the exponentially weighted moving average (EWMA) scheme in a finite horizon production. Based on the Markov chain approach, the performance measures, including the truncated average run length (TARL) and truncated standard deviation of the run length (TSDRL), of the EWMA MCV schemes are derived. Moreover, the superiority of the proposed monitoring schemes is illustrated by comparing their performance with that of the conventional Shewhart MCV monitoring schemes in a short production run context. The results show that increasing the batches enhances the advantages of the EWMA monitoring schemes over the Shewhart schemes. Finally, the superiority of the proposed EWMA monitoring scheme is illustrated by a real steel sleeve manufacturing case study.