Classical Designed Experiments (DOE) were discussed in a previous article in this series. In the last article we discussed descriptive statistics that describe the center of an output or response (mean, median and mode) and the amount of variability in the output or response (range, standard deviation and variance).
Taguchi Designed Experiments are an alternative to classical DOE. This approach was developed by Genichi Taguchi to improve the quality of manufactured goods. Classical DOE focuses primarily on the mean of the response variable, while Taguchi focuses on both the mean and the amount of variability in the response. Taguchi methods are best applied at the design phase of a product or a process. We will discuss Taguchi methods in the context of product design for the rest of this article.
The philosophy of Taguchi can be summarized as follows:
- Quality should be designed into a product, beginning with parameter design. We determine which parameters (or factors in the terminology of DOE) in the design of the product or in the manufacturing process most affect the response of the product, and then design to meet a specified level of quality.
- Quality is best achieved by minimizing the deviation from the target value that provides the highest level of customer satisfaction. The way to do so is to put the mean value of the response at the target and minimize variability around the mean.
- The product should be designed so that it is immune to uncontrollable environmental factors.
- The cost of quality (i.e. the cost of not having good quality) should be measured in terms of deviation of the response from the target value.
Taguchi designs allow you to design a product (and the related manufacturing process) that performs more consistently in the operating environment. Taguchi designs recognize that not all factors that cause variability can be controlled in practice. These uncontrollable factors are called noise factors.
Taguchi designs attempt to identify controllable factors that minimize the effect of the noise factors on the response. During experimentation, you manipulate the noise factors to force variability to occur and then find optimal control factor settings that make the product or process robust. By robust, we mean that the response is resistant to variation caused by the noise factors.
A well-known example of Taguchi designs is from the Ina Tile Company of Japan in the 1950s. The company was manufacturing many tiles outside of the specified dimensions. A quality team discovered that the temperature in the kiln used to bake the tiles varied, causing non-uniform tile dimensions. They could not eliminate the temperature variation because building a new kiln was too costly. Thus, temperature was a noise factor. Using Taguchi designed experiments, the team found that by increasing the clay’s lime content, a control factor, the tiles became more resistant, or robust, to the temperature variation in the kiln, letting them manufacture more uniform tiles.
Another example that I use in our Master Black Belt course deals with the design of golf balls. The response variable is the distance traveled when the ball is struck. The goal is to maximize the mean distance traveled while minimizing the amount of variation around the mean. Control factors include the number of dimples on the surface of the ball, the thickness of the covering on the ball, the material used for the core of the ball, and diameter of the core. Noise factors include thing like ambient temperature, the type of club used to strike the ball, and the amount of energy imparted by each swing.
Taguchi designs have some advantages when compared to Classical DOE. We can design so that the product performs at the target value that provides the highest level of customer satisfaction, rather than designing to perform within a specification range. The experimental designs that Taguchi utilizes are straightforward and easy to apply. Taguchi designs are good for screening a large number of factors to efficiently and cost effectively narrow down many factors to the significant few.
Taguchi designs also have some disadvantages and criticisms. The results from the experiments are relative – the parameters are ranked based on their impact on the response. Taguchi designs have limited or in some cases no ability to evaluate interactions between factors. Taguchi designs work much better for initial design than for improving quality in existing designs.
Your comments or questions about this article are welcome, as are suggestions for future articles. Feel free to contact me by email at firstname.lastname@example.org.
About the author: Mr. Roger C. Ellis is an industrial engineer by training and profession. He is a Six Sigma Master Black Belt with over 45 years of business experience in a wide range of fields. Mr. Ellis develops and instructs Six Sigma professional certification courses for Key Performance LLC. For a more detailed biography, please refer to www.keyperformance.com.