Access Type

Open Access Dissertation

Date of Award

1-1-1990

Degree Type

Dissertation

Degree Name

Ph.D.

Department

Industrial and Manufacturing Engineering

First Advisor

Kailash C. Kapur

Abstract

This research is concerned with product and process design optimization by quality engineering based on the work of Dr. Taguchi, with emphasis on the optimization of dynamic systems and tolerance design. Various quality loss functions are presented in this thesis which can be used for quality evaluation. The goal of robust design for dynamic systems is to reduce the deviations of quality characteristics for the real system from an ideal target which can change based on the requirements of the customer. The optimization can be simplified by decomposing the selection of control factors Z and the adjustment of scaling/leveling factors R into a two-step procedure. The first step is selecting levels for factors Z to maximize the signal-to-noise (SN) ratio that is supposed to be independent of the adjustment of factors R. The second step is used to adjust the real system to a desired model. A systematic approach to optimization is provided for dynamic systems. The motivation of the SN ratio is given and the validity of the SN ratio is examined for various systems. However, for the specific models where the use of the SN ratio is questionable, the necessary modification is suggested. In addition, discrete dynamic characteristics are discussed. The objective of tolerance design is to balance quality loss due to variations and cost increase due to control of variations. Based on the variation transmission equation developed in this thesis, the best tolerance levels are specified for components and subsystems. The tolerance design approach is presented for quality characteristics which may deteriorate over time. Also, a method is presented to develop the tolerances for lower-level quality characteristics based on the tolerances for higher-level quality characteristics, to reflect the voice of the customer. Illustrations are given to demonstrate the efficiency of the tolerance design methodology.

Share

COinS