Double EWMA-Based Polynomial Quality Profiles Monitoring

Abstract

Profile monitoring is useful whenever the quality is described by a relationship between a response and one or more of explanatory variables. It is conventional to refer to this relationship as Quality Profile. Despite the fact that there are numerous examples of the polynomial quality profiles in industrial practice, only few Phase II techniques are available in literature for monitoring these profiles. However, to promote this research context, we extend the applications of the double exponentially weighted moving average (dEWMA) statistic to include phase II of the Mth polynomial quality profiles. Some of existing methods have been modified in order to capture the nature of the dEWMA statistic under the polynomial profile monitoring. Their detection accuracy measured by the average run length performance is evaluated. To expand our comparison, the average run length performance of the multivariate cumulative sum and T2 are also subjected to evaluation as well. The comparative simulations showed the superior performance of the dEWMA-based charting techniques in detecting changes in the parameters of the polynomial quality profiles.