Page 564 - Design for Six Sigma a Roadmap for Product Development
P. 564

522   Chapter Fourteen


             metal ions is very harmful because of its toxicity and nonbiodegradable
             nature. It is proposed to use hydrated iron oxides [Fe(II)] with an appropriate
             pH level to remove these harmful metal ions. The objective of this project is
             to remove as many metal ions as possible from the wastewater. The output
             quality characteristic is the total remaining (metal) concentration [TRC; in
             milligrams per liter (mg/L)], which is a smaller-the-better quality charac-
             teristic. The following four control factors are identified:

                     Control factors    Level 1  Level 2  Level 3
               F  Fe(II)/(total metal in solution)  2  7  15
               T  Temperature, °C        25     50      75
               H  Aging time, h           1      2      3
               P  pH                      8     10      12
             The variability in the composition of wastewater is considered to be the noise
             factor. In this Taguchi parameter design project, artificially introduced potas-
             sium permanganate is used to simulate the noise factor as follows:

                  Noise factor   Level 1   Level 2   Level 3
               N  KMnO 4       3.75 
 10  3  3.75 
 10  2  7.5 
 10  2
               concentration, mol/L

             It is assumed that there is no interaction between control factors; an  L 9
             orthogonal array is selected as the inner array. The experimental data are
             tabulated as follows:
                         Control factors
                          and levels  Output (TRC, mg/L) at different noise factor levels
                                        N1          N2         N3
               Experiment                                              S/N
                  no.    T  P  F  H  Rep. 1  Rep. 2  Rep. 1  Rep. 2  Rep. 1  Rep. 2  ratio
                   1     1  1  1  1  2.24  0.59  5.29  1.75  155.04  166.27  39.35
                   2     1  2  2  2  1.75  5.07  1.05  0.41  0.38  0.48   7.05
                   3     1  3  3  3  5.32  0.65  0.40  1.07  0.51  0.36   7.05
                   4     2  1  2  3  0.37  0.32  0.34  0.68  4.31  0.65   5.19
                   5     2  2  3  1  7.20  0.49  0.48  0.44  0.80  0.88   9.54
                   6     2  3  1  2  39.17  27.05  46.54  25.77  138.08  165.61  39.34
                   7     3  1  3  2  0.57  1.26  0.61  0.70  0.91  1.42  0.28
                   8     3  2  1  3  3.88  7.85  22.74  36.33  92.80  120.33  36.20
                   9     3  3  2  1  15.42  25.52  35.27  48.61  67.56  72.73   33.79
             The signal-to-noise ratio is computed by the following formula:

                                              1  n
                                 S/N   10 log        2
                                               n  
 y i
                                                i   1
             By using MINITAB, we can obtain the following ANOVA table for S/N ratio
             and Fig. 14.15 for percentage contribution of each control factor. Clearly,
             control factor F [Fe(II)/(total metal in solution)] is by far the most important
             factor in influencing S/N, contributing 74.91 percent to total S/N; control
             factor H (aging time) and P (pH) are also significant (more than 10 percent
             contribution). Control factor T has a negligible effect.
   559   560   561   562   563   564   565   566   567   568   569