Page 302 - Six Sigma Advanced Tools for Black Belts and Master Black Belts

OTE/SPH
OTE/SPH
Char Count= 0
August 31, 2006
JWBK119-18
3:6
Training and Learning 287
of experiments in industry: practitioners, although not particularly attracted by ab-
stract arguments brandished by statisticians, will also be wary of embracing Taguchi
methods when they see counterexamples that demonstrate the problems of Taguchi
recipes, as well as open and direct condemnations of the Taguchi approach. Thus the
controversy, the scale of which has never been seen before in the quality profession,
could prove counterproductive if managers and engineers are to be perplexed into
inaction when faced with suggestions to get into experimental design, the source of
the most powerful techniques for fundamental quality improvements. Therefore it
would be useful to consider strategies by which industrial personnel may possibly
be led out of the woods and inducted into the community of experimental design
beneficiaries.

18.7.1 Motivation
As already pointed out, experimental design was historically developed for the under-
standing of complex natural phenomena via empirical means. Taguchi extended its
applications from a mere understanding of existing natural phenomena to the creation
process of human-made physical mechanisms. Taguchi methods offer their users a
structured approach to meeting performance objectives set for a manufacturing pro-
cess or manufactured product in the face of variations induced by raw materials, basic
components, and the environment, expressing the degree of success of the effort in a
language that management can understand, that is, in monetary terms via a special
quality definition and the loss function concept.
To industrial personnel it may be prudent not to speak of experimental design
per se since it tends to invite subconscious resistance by those who have a mental
block against probability and statistics. Instead, robust design or quality engineering
motives should be presented to them as a capability for performance enhancement,
in terms of problem solving or design optimization, that does not require a search
for new technology or additional capital investment. For organizations that are not
unfamiliar with mathematical ideas, however, the subject of experimental design can
be quite logically introduced as further tools to be acquired for fundamental gains
and breakthroughs after SQC or statistical process control has been stretched to the
limits of its usefulness. In this respect the summary statements shown in Table 18.4,
in addition to Table 18.1, would be useful for motivation as well as appreciation
purposes.

18.7.2 Applications
Since statistical quality engineering is a relatively new subject, most people in industry
have to learn it from short training courses designed for working personnel. Sugges-
tions for effective design and implementation of training programs have been offered
59
elsewhere. An important point is that to make an application-oriented course useful,
course contents should be designed in accordance with the 80--20 principle, reflecting
the fact that as much as 80 % of all problems encountered in practice are likely to be
adequately handled by 20 % of all available techniques. It would therefore be unpro-
ductive to pursue a complete range of topics, a practice more for academic courses in
colleges and universities than industrial training programs. In this way some of the