Control Charts for Linear Profile Monitoring with Within-Profile Correlation Using Tweedie Process Model

Abstract

For some manufacturing processes the quality of a product can be characterized by a functional relationship between the response variable and explanatory variables which has been referred to as a profile Recently profile monitoring is used to maintain the stability of the product quality over time by various industries In this research we focus on the single-variate linear profile monitoring Firstly the Tweedie exponential dispersion process (Tweedie ED) model is used to describe the correlated relationship among within-profile data Then a multivariate exponentially weighted moving average (MEWMA) control chart is constructed to detect the process change in Phase Ⅱ study In the simulation studies different combinations of slope and dispersion parameters are considered in our Tweedie ED model Moreover the in-control and out-of-control average run lengths (ARLs) are used as a criteria to evaluation the performance of MEWMA and T^2 MMR control charts Based on the simulation results we find that our proposed MEWMA control chart has a better detecting performance than the T^2 and the MMR control charts in Phase Ⅱ study since the Tweedie ED model is more suitable to fit linear profile data if the within-profile data are correlated Finally two numerical examples are given to demonstrate the usefulness of our proposed control charts in practical applications

Date of Award	2020
Original language	English
Supervisor	Chung-I Li (Supervisor)