Page 560 - Design for Six Sigma a Roadmap for Product Development

518 Chapter Fourteen

Control Factors
A, B, C, …

Output y
System
Input
Quality
Characteristics

Noise Factors Smaller-the-better
U, V, W, … Larger-the-better
Nominal-the-best
Figure 14.13 Taguchi robust parameter design
process.

After noise factors are identified, they will be assigned in the robust
parameter design experiment, the paradigm of which is illustrated in
Fig. 14.13.
The actual assignments of control factors and noise factors, the exper-
imental layout and data collection follows the inner-array/outer-array
template illustrated in Table 14.4.
Control factors are assigned to an orthogonal array called inner
array, where each experimental run corresponds to a unique control
factor level combination. At each inner array experimental run, we can
vary noise factors according to the layout of another orthogonal array
assigned to noise factors, called the outer array. Each run in the outer
array corresponds to a unique combination of noise factor levels. For
instance, if the outer array has N 2 runs, then for each inner-array
experimental run, we will obtain N 2 . Experimental observations in
which each observation corresponds to a different noise factor level

TABLE 14.4 Taguchi’s Inner Array–Outer Array Layout
Outer array
Experiment no. Noise
1 2 … N 2 factors
Inner array 1 1 2 U
Experiment Control Factors 1 2 2 S/N
no. A B C … F 1 2 1 W
1 1 1 1 1 y 11 y 12 y 1N 2 1
2 2 1 2 1 y 21 y 22 Y 2N 2 2

,
N 1 2 2 2 2 y N 1 1 y N 1 2 y N 1 N 2 N 1