Monitoring Non-Parametric Profiles using Adaptive EWMA Control Chart

Abstract

To monitor the quality of a process in statistical process control (SPC), considering a functional relationship between a dependent variable and one or more independent variables which is denoted as profile monitoring is becoming an increasingly common approach. Most of the studies in the SPC literature have dealt with cases in which the functional relationship has the same form in the in-control (IC) and out-of-control (OC) situation or considered parametric approaches. Non-parametric profiles, which have different relationship in OC conditions is also very common. This paper designs a novel control chart to monitor not only the regression parameters but also the variation of the profile in phase II applications, using an adaptive approach. Adaptive control charts adjust the final statistic with regard to information of previous samples and the proposed method considers the relative distance of the chart statistic to the control limits as a tendency index and provides some outcomes about the process condition. The results of Monte Carlo simulations show the superiority of the proposed monitoring scheme in comparison with the common non-parametric control chart.