Tuesday

Taguchi Method: Faster Race Testing


Dr. Genichi Taguchi

Taguchi Method: Faster Race Testing
By Don Terrill © - RacingSecrets.com

Taguchi was raised in the textile town of Takamachi where he initially studied textile engineering with the intention of entering the family kimono business. However, with the escallation of World War II, in 1942, he was drafted into the Astronomical Department of the Navigation Institute of the Imperial Japanese Navy.

After the war, in 1948, he joined the Ministry of Public Health and Welfare where he came under the influence of eminent statistician Matosaburo Masuyama who kindled his interest in design of experiments. He also worked at the Institute of Statistical Mathematics, during this time, and supported experimental work on the production of penicillin at Morinaga Pharmaceuticals, a Morinaga Seika company.

In 1950, he joined the Electrical Communications Laboratory (ECL) of the Nippon Telephone and Telegraph Company just as statistical quality control was beginning to become popular in Japan under the influence of W. Edwards Deming and the Japanese Union of Scientists and Engineers. ECL was engaged in a rivalry with Bell Labs to develop cross bar and telephone switching systems and Taguchi spent his twelve years there in developing methods for enhancing quality and reliability. Even at this point, he was beginning to consult widely in Japanese industry, with Toyota being an early adopter of his ideas.

During the 1950s, he collaborated widely and in 1954-1955 was visiting professor at the Indian Statistical Institute where he worked with R. A. Fisher and Walter A. Shewhart.

On completing his doctorate from Kyushu University in 1962, he left ECL, though he maintained a consulting relationship. In the same year he visited Princeton University under the sponsorship of John Tukey who arranged a spell at Bell Labs, his old ECL rivals. In 1964 he became professor of engineering at Aoyama Gakuin University, Tokyo. In 1966 he began a collaboration with Yuin Wu who later emigrated to the USA and, in 1980, invited Taguchi to lecture. During his visit there, Taguchi himself financed a return to Bell Labs where his initial teaching had made little enduring impact. This second visit began a collaboration with Madhav Phadke and a growing enthusiasm with his methodology in Bell Labs and elsewhere, including Ford Motor Company, Xerox and ITT.

Since 1982, Genichi Taguchi has been an advisor to the Japanese Standards Institute and executive director of the American Supplier Institute, an international consulting organisation.

Taguchi's reaction to the classical design of experiments methodology of R. A. Fisher was that it was perfectly adapted in seeking to improve the mean outcome of a process.

Design Of Experiments:

The first statistician to consider a methodology for the design of experiments was Sir Ronald A. Fisher. He described how to test the hypothesis that a certain lady could distinguish by flavor alone whether the milk or the tea was first placed in the cup. While this sounds like a frivolous application, it allowed him to illustrate the most important ideas of experimental design.

Design of experiments was built on the foundation of the analysis of variance, a collection of models in which the observed variance is partitioned into components due to different factors which are estimated and/or tested.

Developments of the theory of linear models have encompassed and surpassed the cases that concerned early writers. Today, the theory rests on advanced topics in abstract algebra and combinatorics.

Ok, that's some background, now how do we use it:

Let's say you want to find the best roller camshaft for an application, say a 350" engine for short track stock car racing. The four major components of a camshaft are intake duration, exhuast duration, lobe separation and the lobe family (lobe design). So, let's select a high, middle and low value for each, for example:

Intake Duration: 255, 260, 265
Exhaust Duration: 260, 265, 270
Lobe Separation: 106, 108, 110
Lobe Family: -5, SC, RT (Old Comp Cams Lobes)

To test every possible combinations you'd have to purchase 81 different cams. Through Taguchi's Method that can be reduced to just 9.



You would then dyno test each camshaft and assign an average HP number to each - average over the engine's operating RPM range.

From this information you will be able to calculate, with high probability, the best cam combination and odds are it won't be one of the nine above.

Sign up for my hardcore racing newsletter at RacingSecrets.com