Title: "DESIGN OF EXPERIMENTS (DoE) METHODOLOGY FOR INDUSTRY"
Author: PABLO MARTÍN TANCO RAINUSSO
Date: 2008-07-17
Director: Dr. Elisabeth Viles Díez
Abstract:
Nowadays, when considering the current state of the business world, experimentation stands out as an important element that contributes to learning and improving products and processes.
Design of Experiments (DoE) is presented as an effective statistical tool to learn and optimize processes and products by maximizing information gained from experimentation. In spite of the numerous advantages of DoE, it is not usually applied in industry. There are many difficulties or barriers that hinder the application of DoE. Among others, we believe that the lack of assistance when it comes to the planning of experiments is a significant barrier. Therefore, this PhD thesis aims to present a simple methodology for the application of DoE in industry.
Our approach presents a framework for the experimentation process which follows the traditional DMAIC steps of “Six Sigma” as a generic problem solving methodology. We aim to present a methodology to make the implementation of DoE in industry easier. We are trying to present a simple and easy methodology from an engineering point of view, without ignoring the complexity of this task and its statistical severity. This methodology is presented as an alternative to existing ones, which tend to be muddled, excessively plain and without the depth required for planning experiments. Those methodological shortcomings are especially troublesome, application-wise, for those who must carry out experiments and are not experts in the field.
Therefore, our methodology includes guidelines and worksheets to aid and support practitioners through every step of the DoE application process.
Firstly, this thesis includes two empirical studies, which intend to characterize experimentation in industry and present a thorough analysis of the existing barriers that companies face in terms of DoE application. Those studies corroborated our initial findings that DoE is hardly used in industry and that companies consider themselves in need of a DoE methodology.
This thesis also includes some experiences that we’ve had with companies, as well as two cases in which our newly developed methodology was applied. These two examples provided us with important feedback and allowed us to improve the quality of our methodology.
