Author

Lee

Date of Award

8-2003

Degree Name

Master of Science

Department

Industrial and Entrepreneurial Engineering and Engineering Management

First Advisor

Dr. Leonard Lamberson

Second Advisor

Dr. Abdolazim Houshyar

Third Advisor

Dr. Steven Butt

Access Setting

Masters Thesis-Open Access

Abstract

The concept of parameter design developed by Taguchi to achieve robust engineering is widely promoted in industry as a means to design a product whose output is largely immune to noise in the system. Taguchi has used the signal-to-noise (S/N) ratio to summarize measures over the noise factor settings at the parameter design and to allow control of the experiment.

The objectives of this investigation are: 1) to infer distributions of S/N ratios (lower-the-better S/N, higher-the-better S/N, and nominal-the-best SIN), and 2) to determine the effectiveness of the S/N ratio as a means for identifying changes in the mean and /or the variance in designed experiments based on orthogonal array.

Results show that gamma distribution was best fit distribution for the lower-the- better S/N and the nominal-the-best S/N, while beta distribution was the best fit distribution for the higher-the-better S/N. It was also found that a change in the mean of a factor leads to a negative change for the lower-the-better S/N and a positive change for the higher-the-better S/N and the nominal-the-best S/N; a change in the variance of a factor leads to a negative change for the nominal-the-best S/N only.

Share

COinS