Content area
Full Text
Keywords
Taguchi methods, Statistical process control, Design of experiments
Abstract
The Taguchi method (Tm) is a powerful problem solving technique for improving process performance, yield and productivity. It reduces scrap rates, rework costs and manufacturing costs due to excessive variability in processes. However, its application by industrial engineers in the UK is limited, in part due to the inadequate statistical education of engineers. This paper presents a simple experiment which can be used in the classroom to teach engineers the basics of the technique and illustrates simple analytical and graphical tools which promote rapid understanding of the results of the experiment.
Introduction
Dr Genichi Taguchi is a Japanese quality management consultant who has developed and promoted a philosophy and methodology for continuous quality improvement in products and processes. Within this philosophy, Taguchi shows how the statistical design of experiments (SDOE or DOE) can help industrial engineers design and manufacture products that are both of high quality and low cost. His approach is primarily focused on eliminating the causes of poor quality and on making product performance insensitive to variation. DOE is a powerful statistical technique for determining the optimal factor settings of a process and thereby achieving improved process performance, reduced process variability and improved manufacturability of products and processes.
Taguchi (1986) advocates the use of orthogonal array designs to assign the factors chosen for the experiment. The most commonly used orthogonal array designs are L8 (i.e. eight experimental trials), L16 and L18. The power of the Taguchi method is that it integrates statistical methods into the engineering process. Bendell et al. (1989) and Rowlands et al. (2000) report success of the Taguchi method in the automotive, plastics, semiconductors, metal fabrication and foundry industries. However, Antony (1996) suggests that the application of the Taguchi method in the UK manufacturing and service industries is limited and often applied incorrectly. Moreover, a typical remark is "I can do the text book and class room examples, but I am not comfortable and confident in applying the concepts and principles of DOE in my work area".
According to Antony et al. (Antony et al. 1996a; 1998a; 1998b; 1999; Antony, 1998) the following issues are key to this lack of or improper application of experimental design...