Taguchi Methods – iPaGa

Taguchi Methods is a statistical method developed by Genichi Taguchi to improve the quality of manufacturing, engineering, biotechnology, marketing, and edvertising products.

Philosophy

• Quality should be designed in the product, not checked in the product
• The best quality can be achieved by minimizing deviation from the target. The product must be designed to withstand uncontrollable environmental factors
• The cost of a quality should be measured as a function of deviation from the standard and losses should be measured as a whole system

The Taguchi Methods involve reducing the variation of the process through the robust design of the experiment. The main goal of this method is to produce high quality products at very low costs. Taguchi developed a method for designing experiments to investigate how much influence different parameters have on the mean (average) and variance of the process performance characteristics that determine how well the process functions. This experimental design introduced by Taguchi involves orthogonal arrays to organize the parameters that affect the process and the levels that need to be varied. Taguchi Methods does not test all possible combinations but only tests a few. This test will produce a collection of important data that can determine which factors have the most effect on product quality with a minimum of experimentation so as to save time and money.

The general steps in Taguchi Methods are as follows:

1. Determine the purpose of the process or more specifically the target value for measuring the performance of a process.
2. Determine the design parameters that have an effect on the process
3. Create orthogonal arrays for parameter design indicating the number and condition of each experiment
4. Linking indicated experiments to a completed array to collect data on the effects of performance measures
5. Complete data analysis to determine the effect of different parameters on performance measurement.

Taguchi methods are best used when there is an intermediate number of variables (3 to 50), little interaction between variables, and when only a few variables make a significant contribution.

Incoming search terms: