Improving Statistical Process Monitoring of Quality Characteristic with Polynomial Profile in Phase II

Abstract

Profile monitoring is a new research subject in Statistical Process Control which has been recently considered by many researchers. Profile describes the relationship between a response variable and one or more independent variables. This relationship is often modeled using regression which can be simple linear, multiple linear, polynomial or sometimes nonlinear. To monitor polynomial profiles, some methods have been developed, the best of which is the orthogonal polynomial approach. One of the disadvantages of this method is the large number of control charts which are simultaneously used. In this paper, a new approach has been proposed based on orthogonal polynomial approach, in which only two control charts are used to monitor a kth-order polynomial profile. Simulation findings and average run length (ARL) curve analysis imply better performance of the proposed approach compared to the existing approach. Also, using the proposed approach is much easier in practice.

Keywords

• Average run length
• Exponentially weighted moving average control chart
• Orthogonal polynomial profile
• Statistical Process Control