Page 514 - Design for Six Sigma a Roadmap for Product Development
P. 514

Taguchi’s Orthogonal Array Experiment  473


           2. For each factor, A,B,…; if the number of levels are n A ,n B ,…, for each
              factor, the degree of freedom   number of levels  1; for example,
              the degree of freedom for factor A   n A   1.
           3. For any two-factor interaction, for example, AB interaction, the
              degree of freedom   (n A   1)(n B   1).

             Example 13.1 In an experiment, there is 1 two-level factor, A, and 6 three-
             level factors, B,C,D,E,F,G, and 1 two-factor interaction, AB. Then, the total
             degree of freedom is as follows:

                   Factors  Degree of freedom
               Overall mean  1
               A            2   1   1
               B,C,D,E,F,G  6 
 (3   1)   12
               AB           (2   1)(3   1)   2
               Total DOF    16


           13.2.2 Experimental design
           Taguchi experimental design follows a three-step procedure:

           ■ Step 1: Find the total degree of freedom (DOF).
           ■ Step 2: Select a standard orthogonal array using the following two
             rules:
             Rule 1: The number of runs in the orthogonal array   total DOF.
             Rule 2: The selected orthogonal array should be able to accommodate
               the factor level combinations in the experiment.
           ■ Step 3:Assign factors to appropriate columns using the following rules:
             Rule 1: Assign interactions according to the linear graph and inter-
               action table.
             Rule 2: Use special techniques, such as dummy level and column
               merging, when the original orthogonal array is not able to accom-
               modate the factor levels in the experiment.
             Rule 3: Keep some column(s) empty if not all columns can be assigned.

           In selecting orthogonal arrays, Table 13.5 can be used as a reference.

             Example 13.2 In an experiment, there are seven factors. We will consider
             main effects only. First, we compute DOF   1   7(2   1)   8. Therefore, the
             selected orthogonal array should have at least eight runs. By examining
             Table 13.5, we find that the  L 8 array can accommodate 7 two-level factors.
             Therefore, we can use L 8 and assign those seven factors to seven columns of L 8 .

             Example 13.3 In an experiment, there is one two-level factor  A, and 6
             three-level factors, B,C,D,E,F,G. First, DOF   1   (2   1)   6(3   1)   14.
   509   510   511   512   513   514   515   516   517   518   519