STATISTICAL PROCESS CONTROL FOR HIGH QUALITY & COMPLEX PROCESSES

Abstract

This dissertation is a study on Statistical Process Control (SPC) techniques for high quality and complex processes. It deals with problems in implementing SPC in high quality or complex manufacturing processes. The objective is to update the theory and practice of SPC for its use in modern manufacturing environments. Chapter 1 is an introduction to the background and purpose of this research. It presents a review of some important concepts and methods of SPC and the statistical background of control charts. It also decribes the problems studied in this dissertation and defines the scope of their study. Following the introductory chapter, the dissertation is divided into four parts. Part I presents a literature review of recent approaches to SPC for high quality processes. It provides a summary of new techniques designated for high quality processes and conditions of their applications, advantages and drawbacks. The objective is to give chart users an application guide. Part II contains three chapters on Cumulative Count of Conforming (CCC) control charts. Chapter 3 presents a Shewhart-like CCC control chart suitable for high yield processes and automatic manufacturing environments. The sensitivity study shows that the chart is effective in monitoring processes with very low fraction defective and is able to detect process improvement. Chapter 4 is a study on supplementary run rules for CCC control charts. The objective is to further improve the chart’s sensitivity to small process shifts. It presents an algorithm to evaluate the performance of the supplementary run rules using a Markov-chain approach. It also provides some guidance in implementing run rules based on the result from a sensitivity study. Chapter 5 deals with the economic design of CCC control charts for maintaining the current control of fraction nonconforming of a process. A process model is proposed to obtain an appropriate loss function. An algorithm to search for the optimal settings of the sampling and control parameters is derived. Numerical illustrations of the method, some properties of the optimal economic design and a sensitivity analysis of the model are provided. Part III contains two chapters on quality control for high yield processes subject to random shocks. This type of processes are common in a high-yield production environment and the conventional Shewhart control charts are not efficient for its monitoring and controlling. Chapter 6 presents a charting technique for monitoring this type of processes. The proposed technique is able to detect process improvement, easy for decision making, and more concise and informative than other methods available before. In addition, it could provide diagnostic information which is highly useful in practice. A two-component mixed Poisson distribution useful in modeling and analyzing of certain quality and reliability data is studied in Chapter 7. Its statistical properties, physical interpretations, and some practical issues are discussed. The model generalizes the modified Poisson distribution presented in Chapter 6. The approach to using the model and its application are highlighted by means of some actual data. Part IV contains two chapters on quality control for complex processes. Chapter 8 is about implementing SPC in a complex production process. Because such a process could consist of many sub-processes and the human and economic resources are limited, the management has to decide which sub-processes are to be given higher priorities. A preliminary selection method and an Analytic Hierarchy Process (AHP) approach are then proposed. The approach can be used in management decision making in planning for SPC implementation. Chapter 9 is a discussion on multivariate quality control. It presents a brief review of current approaches to multivariate quality control and highlights the problem of quality control for multi-type nonconformities. Several approaches to the problem are provided and potential scope for future research is pointed out.