سعيد اديب فر

When a process is under statistical control, in many cases, the performance of the process is eva‎luated through process capability indices. Extensive studies have been conducted on eva‎luating process capability when the quality characteristic is univariate or multivariate. However, these methods are not effective in cases where the quality characteristic is expressed as a profile. A profile represents a relationship between independent variables and the response variable, which can be expressed in simple, multiple, multivariate, and more complex relationships. On the other hand, many processes, whether manufacturing or service-oriented, consist of several consecutive stages, known as multistage processes. In these processes, the output of each stage is considered as the input for the next stage, which is referred to as the cascade property. Due to the presence of the cascade property in multistage processes, the use of conventional statistical quality control methods may lead to incorrect results. The aim of this study is to examine process capability for multivariate linear profiles in multistage processes. In this study, process capability is initially calculated using a suitable multivariate index. To investigate the capability of intermediate stages, a cause-selecting approach is used to account for the cascade property. The basis of this approach is the analysis of residuals in multistage processes, as the cascade property does not affect the residuals, and allows to investigate the specific process capability of the process in each stage. In addition to providing a process capability index for multivariate linear profiles in multistage processes, a method for eva‎luating the capability of profile parameters is also presented in this dissertation. The performance of the proposed method is examined through numerical example. Since the proposed method not only eva‎luates the capability of quality characteristics alongside the capability of profile parameters, but also considers the in-control values of the parameters in assessing their capability, this approach offers a relatively more accurate performance of the process compared to other developed methods.

قابليت فرايند , فرايندهاي چندمرحله‌اي , خاصيت آبشاري , پروفايل خطي چندمتغيره

Process Capability , Multi-stage Process , Cascade Property , Multivariate linear Profile

Saeed Adibfar

Dr. Rassoul Noorossana