A Study of Experiment Design with Shainin and Taguchi Methods: A Case Study of Aluminum Riveting Process

Abstract

The aim of this thesis is to study and provide the implementation of Taguchi method and Shainin method by using a case of Aluminum Riveting Process. Then, comparative analysis Taguchi, Shainin and classical factorial experiment designs was carried out. Taguchi method uses experiment of Orthogonal Array (L16) and Shainin method operates to 4 processes. 2⁵Factorial Design is used as A Classical Experiment designs. Taguchi method can give the same result as the Classical factorial experiment that provides high confidence within less number of trials. The number of trials of Shainin method is less than those of Taguchi method and Classical factorial experiment. It is valuable to save cost and time of experiment. Applications of Taguchi and Shainin method are useful for the factor studying when a number of trials cannot be performed on many experiments, especially in case of the screening experiment. However, approaches of  Taguchi and Shainin method require more knowledge in the a case study process in order to determine what is factor and level in the experiments than that of Classical design.