Summary: | 碩士 === 元智大學 === 工業工程與管理學系 === 98 === In most statistical process control (SPC) applications, it is assumed that the quality of a process or product can be adequately represented by the distribution of a univariate quality characteristic. However in many practical situations, the quality of a process or product is better characterized and summarized by a relationship between a response variable and one or more explanatory variable. Such relationship can be represented by a curve known as profile. However, surprisingly few works have been done in developing statistical process control methodology for monitoring profile data.
In recent years, profile monitoring has become a popular and fertile field of research in statistical process control in both Phases I and II. In this thesis, we propose the use of parametric and nonparametric approaches to monitor the reflow process data. Each profile is modeled using the sum of sine functions; the estimated parameter vector is monitored by using the Hotelling statistics. Due to the very restrictive assumptions in nonlinear estimation, a complementary approach is developed using metrics that measure deviation between observed profiles and a reference profile. The statistical performance of these methods is evaluated through the reflow process data using the Average Run Length (ARL) criteria. A comparison between the control chart and a combination of the metrics by means of the ARL is investigated.
|