SlideShare a Scribd company logo
1 of 77
Download to read offline
M6141 QUALITY ENGINEERING                                                             CONTENTS
                                                                               1. Introduction
                  DESIGN OF EXPERIMENTS
                                                                               2. Fundamentals of Probability and Statistics (cancelled)
                          A/P Wu Zhang                                         3. Single Factor Experiments
                            MAE, NTU
                                                                               4. Randomized Blocks (cancelled)
                       Office: N3.2-02-14
                         Tel: 6790 4445                                        5. Factorial Experiments
                     Email: mzwu@ntu.edu.sg                                    6. 2k Experiments
                                                                               7. Fractional Experiments
TEXT                                                                           8. A Case Study
   Montgomery, D. C., (2009), Design and Analysis of Experiments,
   John Wiley & Sons.

                                                                   1-1                                                                     1-2




1 INTRODUCTION                                                           Example 1.1:    Two factors: Machine, Feed rate
                                                                                         Response:     Process capability
A designed experiment is a test or series of tests, in which, the
experimenter chooses certain factors for study. He purposely varies                              Table 1.1
those factors in a controlled fashion, and then observes the effect of                           Process capability Cp
such factors on the response function.                                                            Feed rate (mm/sec.)
                                                                               Machine         0.2        0.4        0.6
                                                                                USA         2.1, 2.2 1.7, 1.6 1.2, 1.3
                                                                                Japan       0.9, 1.0 0.7, 0.6 0.4, 0.3

                                                                         Conclusion: using USA machine at lowest feed rate will result in
                                                                                     max. process capability.




                                                                   1-3                                                                     1-4
Factor (input variable)                                                        (3) Run: A run is one set of levels for all factors in an experiment.
                                                                                   In above experiment, there are three factors: (A) temperature,
(1) Types                                                                          (B) machine and (C) switch. The total number of runs is
     Quantitative: Temperature, time: They can be set at any value.
                                                                                    R = 4 × 3 × 2 = 24
     Qualitative: Machines, operators: They have different categories
                  Switch:                  It has different status                  For example, [1]           A = 10, B = Japan-made, C = on
                                                                                                         ..........
(2) Levels
                                                                                                    [24]    A = 40, B = UK-made, C = off
     Temperature (oC):     10, 20, 30, 40 [4 levels]
     Machine:              made by Japan, by USA, by UK [3 levels]                  For single factor experiments, run is equivalent to level of the
                                                                                    factor. The number R of runs is equal to the number a of levels.
     Switch:               on, or off [2 levels]
                                                                               (4) Replicates (n): number of observations at each run. A larger
                                                                                   value of n increases the accuracy of the experimental results, but
                                                                                   makes the experiment more expensive.
                                                                         1-5                                                                      1-6




Run        A        B         C      Run      A         B         C            Response (or output, observation)
No       Temp.   Machine    Switch   No     Temp.    Machine    Switch
 1        10      Japan      On       13     30       Japan      On
                                                                               Response is dependent on the values of the factors.
 2        10      Japan      Off     14       30       Japan     Off
 3        10      USA        On      15       30       USA       On            Types of response:
 4        10      USA        Off     16       30       USA       Off
 5        10       UK        On      17       30       UK        On               Larger the better:     product yield, process capability, student’s
 6        10       UK        Off     18       30       UK        Off
                                                                                                         mark
 7        20      Japan      On      19       40       Japan     On
                                                                                  Smaller the better: product cost, number of errors
 8        20      Japan      Off     20       40       Japan     Off
 9        20      USA        On      21       40       USA       On               Nominal the best:      diameter of a shaft
 10       20      USA        Off     22       40       USA       Off
 11       20       UK        On      23       40       UK        On            By DOE, we can find out which factors have effects on the response,
 12       20       UK        Off     24       40       UK        Off
                                                                               and the direction of the influence (i.e., whether the increase of a
                                                                               factor will increase or decrease the response ?)
                                                                         1-7                                                                      1-8
Example 1.2:      Single factor: A text book     (a = 2, n = 3, R = 2)       Example 1.3:     Two factors: A text book, B studying hours
                  Response:      Students’ marks                                              Response: Students’ marks

                      Table 1.2                                                               (a = 2, b = 3, n = 2, R = 6)
                        Students’ marks
                                                                                              Table 1.3
     Text 1      76, 74, 78           Ave = 76
                                                                                                  Students’ marks
     Text 2      48, 50, 52           Ave = 50                                               1 hr       2 hr      3 hr
                                                                                  Text 1    45, 47    59, 58     76, 75
                                                                                  Text 2    32, 33    51, 54     70, 71
   Text book (a qualitative factor)
       Text 1:    book by Montgomery
       Text 2:    book by Taguchi



                                                                       1-9                                                                          1-10




The objectives of an experiment may include                                  What Makes Designed Experiment (DOE) Special?

(1) Determining which factors xs are most influential on the response,       (1) It conducts the experiments in a systematic and efficient way.
    y.                                                                           Reduce design and development time, as well as the cost.

                                                                             (2) It presents experiment results in the simplest and clearest way.
(2) Determining where to set the influential x's, so that y is near the
    desired requirement.                                                     (3) It extracts the maximum amount of information from a given set
                                                                                 of experiments.
(3) Determining where to set the influential x's, so that variability in y
    is small.                                                                (4) It draws the right conclusions, despite the variability presented in
                                                                                 the data.

                                                                             Examples:     Increase process capability
                                                                                           Improve students’ study
                                                                                           Reduce output variability.

                                                                      1-11                                                                          1-12
Example 1.4 Reducing Defect Level in a Process                           In this application, DOE
A manufacturing engineer is going to apply DOE to a process for
                                                                         (1) Determines which factors affect the occurrence of defects
soldering electronic components to printed circuit boards, in order to
reduce defect levels even further.
                                                                         (2) Determines the direction of adjustment for the effective factors to
Factors to be investigated:                                                  further reduce the number of defects per unit.
          Solder temperature
          Preheat temperature                                            (2) Decides whether changing the factors together produces different
          Conveyor speed                                                     results than just adjusting individual factor separately?
          Flux type
          Flux specific gravity
          Solder wave depth
          Conveyor angle
          Thickness of the printed circuit board
          Types of components used on the board

                                                                  1-13                                                                      1-14
3 SINGLE FACTOR EXPERIMENTS                                                                                Table 3.1 Tensile strength of paper
                                                                          ___________________________________________________________________________________________________________

                                                                                Hardwood                                                  Observations
3.1 Introduction                                                                                              _____________________________________________________________________________________________

                                                                          Concen. (%)                         1           2       3         4          5         6           Total            Aver.
                                                                          _________________________________________________________________________________________________________________________________
Single factor experiment can be used in the situation where the effect
(influence) of a single factor on the response is dominant. The effects            5                          7           8      15 11                 9        10            60             10.00
of all other factors are negligible.                                             10                           12 17 13 18 19 15                                               94             15.67

The procedures and formulae developed for the single factor                      15                           14 18              19 17 16 18                                 102             17.00
experiments can be easily modified, and then used for the several            20         19 25 22 23 18 20 127          21.17
factors experiments.                                                      __________________________________________________
                                                                                                                                                                             383             15.96

                                                                          Study the effect (influence) of hardwood concentration on the tensile
                                                                          strength of paper.

                                                                    3-1                                                                                                                                       3-2




(1)    One factor: hardwood concentration.                                       Table 3.3 General data table for a single factor experiment

(2)    Four levels: 5%, 10%, 15%, 20%: a = 4.                                    Level                               Observation                                     Total            Average
                                                                                     1                  y11         y12          .           .        y1n              y1.                  y1.
(3)    Response is the tensile strength of paper.
                                                                                     2                  y21         y22          .           .        y2n              y2.                  y2.
(4)    Each level has six observations or replicates: n = 6.
                                                                                       .                  .           .          .           .          .
(5)    Total number of observations: N = an = 24.                                      .                  .           .          .           .          .
                                                                                     a                  ya1         ya2          .           .        yan              ya.                  ya.
                                                                                                                                                                       y..                   y..


                                                                          yij              the jth observation taken under the ith level of the factor.

                                                                    3-3                                                                                                                                       3-4
yi.   the total of the observations at the ith level                   3.2 The Analysis of Variance (ANOVA)
              n
                                                                       ANOVA is the most important statistical tool used in DOE. Its task
       yi. = ∑ yij                                       (3.1)
             j =1
                                                                       is to decide whether and which factors or interactions have
                                                                       significant effect (influence) on the response.
yi.   the average of the observations at the ith level
       yi. = yi. / n                                     (3.2)         Model of ANOVA
                                                                       Yij = μ + τ i + ε ij = μi + ε ij                                   (3.5)
y..   the grand total of all observations
              a     n     a
       y.. = ∑∑ yij = ∑ yi.                              (3.3)
                                                                       μ          the overall mean. μ ≈ y..            (mean ≈ average)
             i =1 j =1   i =1
                                                                       μi         the ith level mean. μ i ≈ yi.
y..   the grand average of all observations.
                                                                       τi         the ith level effect. τ i = μi − μ ≈ yi. − y..
       y.. = y.. / N                                     (3.4)
                                                                       ε ij       the random error. ε ij = yij − μ i ≈ yij − yi.
N = an is the total number of observations
                                                                 3-5                                                                              3-6




Machining Example (1)                                                         y1. = 2.1 + 2.0 + 1.9 = 6.0,       y1. = y1. / 3 = 2.0
Study the effect of machine on the process capability                         y 2. = 0.9 + 1.0 + 1.1 = 3.0,      y 2. = y 2. / 3 = 1.0
               Table 3.4                                                      y.. = (2.1 + 2.0 + 1.9) + (0.9 + 1.0 + 1.1) = 9.0
       Machine     Process capability Cp
                                                                              y.. = 9.0 / 6 = 1.5
         USA             2.1, 2.0, 1.9
        Japan            0.9, 1.0, 1.1                                        μ1 ≈ y1. = 2.0,       μ 2 ≈ y 2. = 1.0
                                                                              μ ≈ y.. = 1.5
(1)   One factor: machine (qualitative).
                                                                              τ1 = μ1 − μ ≈ y1. − y.. = 2.0 − 1.5 = 0.5
(2)   Two levels: USA-made, Japan-made, a = 2.
                                                                              τ 2 = μ 2 − μ ≈ y2. − y.. = 1.0 − 1.5 = −0.5
(3)   Response is the process capability.
(4)   Each level has 3 observations or replicates: n = 3.
(5)   Total number of observations: N = an = 2 × 3 = 6.
                                                                 3-7                                                                              3-8
Evaluation of Errors
                                                                               y
                                                                                               ε11 = 0.1
     For instance, if i = 1, j = 1, yij = y11 = 2.1
     ε ij = yij − μi                                                                                               μ1 = 2
                                                                                                    τ1 = 0.5
     ε11 = y11 − μ1 ≈ y11 − y1. = 2.1 − 2.0 = 0.1                                                                  μ = 1.5
                                                                                   y11 = 2.1        τ2 = -0.5
     Errors for the machining example (1)                                                                           μ2 = 1
                        Table 3.5
         Machine              Errors
            USA            0.1, 0.0, -0.1                                                                  y11 = μ + τ1 + ε11
           Japan          -0.1, 0.0,      0.1                                                                 = 1.5 + 0.5 + 0.1 = 2.1

                                                                                                                             Figure 3.2

                                                               3-9                                                                             3-10




                                              μ1              μ2
                                                                                                                             μ1
                                                    τ1   τ2
                                                                                                                                   τ1
μ1                         μ2                                        μ1        μ2                     μ3
                                  μ                            μ          τ1         τ2        τ3                                                μ
μ3                          μ4                                                                       μ                 τ2          τ3     τ4
                                                                                                                  μ2          μ3                μ4
                                                    τ3   τ4
                                                                                     τ4
                                              μ3              μ4                               μ4

                                 Figure 3.3
                                                                                                     Figure 3.4


                                                              3-11                                                                             3-12
Hypothesis Test: to test if a factor has effect on the response.             Null:         Ho :   the factor has no effect on the response

If a factor has no effect on the response, the response is always the        Alternative: H1:     the factor has effect on the response
same, regardless the change of the factor,
                                                                                Ho : τ1 = τ 2 =L = τ a = 0           (|τi| is small, no effect)
           μ1 = μ 2 = L = μ a = μ
                                                                                H1 : τ i = o for at lease one i
                                                                                         /                           (|τi| is larger, has effect)   (3.6)
           ( μ1 − μ ) = ( μ 2 − μ ) = L = ( μ a − μ ) = 0
          τ 1 = τ 2 = Lτ a = 0                                               The larger the value of any |τi|, the more effective the factor will be.
                                                                             Here, the magnitude of τi, rather than its sign, makes sense.
If a factor has effect, the value of the response will be different, along
with the change of the factor. The more the response differs, the            Type I error: the null hypothesis Ho is rejected when it is actually
greater the effect of the factor is.                                         true (The factor is concluded effective, when it is not)
          τ i = o for at least one i
              /                                                              Type II error: the null hypothesis is accepted when it is actually
                                                                             false (The factor is concluded ineffective, when it is effective).

                                                                      3-13                                                                              3-14




Machining Example (2)                                                                 y
                       Table 3.6
       Machine          Process capability Cp                                                          τ1 = τ2 = 0
         USA              1.6, 1.5, 1.4                                                                               μ = μ1 = μ2 =1.5
         Japan            1.4, 1.5, 1.6
                                                                                                         Minor variation is caused by
                                                                                                         errors, rather than the machines
   μ1 ≈ 1.5, μ 2 ≈1.5 μ ≈ 1.5
   τ 1 = μ1 − μ = 15 − 15 = 0
                   .    .
   τ 2 = μ2 − μ = 15 − 15 = 0
                    .    .

Conclusion: in this example, the factor (machine) has no effect on
            the process capability. Because, when using different                                            Figure 3.5
            machines, the response is always the same, on average.
                                                                      3-15                                                                              3-16
Analysis of Variance (ANOVA)                                                                  (1) Calculate sum of squares (SS)

Major Steps                                                                                    Three types of differences:
                                                                                                 [1] Difference between an individual observation and the grand
(1) Calculate sum of squares (SS).                                                                   average:
                                                                                                       yij - y..
(2) Determine degrees of freedom (D.O.F.).
                                                                                                 [2] Difference between a level average and the grand average:
(3) Calculate mean square (MS).
                                                                                                        yi. - y.. = τ i
(4) Calculate the test ratio F0.                                                                 [3] Difference between an individual observation and the
                                                                                                     corresponding level average:
(5) Make conclusion.                                                                                   yij - yi. = ε ij

                                                                                                 Note: yij - y.. = ( yi. - y.. ) + ( yij - yi. )

                                                                                       3-17                                                                              3-18




 a    n                     a       n             a       n
                                                                                              Sum of squares is the sum of squares of the above three differences.
∑∑ ( y
i =1 j =1
            ij   - y.. )   ∑∑ ( yi. - y.. )
                           i =1 j =1
                                                  ∑∑ ( y
                                                  i =1 j =1
                                                                  ij   - yi. )
                                                                                              SST    The total sum of squares, which is a measure of total
 a    n                         a       n             a       n                                      variability in the data.
∑∑ ( y           - y.. )    ∑∑ ( yi. - y.. )          ∑∑ ( y                - yi . )
                       2                    2                                      2
            ij                                                         ij
i =1 j =1                   i =1 j =1                 i =1 j =1                                                 a    n
                                                                                                     SS T = ∑ ∑ ( yij - y.. ) 2                                  (3.7)
                                                                                                               i =1 j =1
  SST                  =            SSF       +               SSE
                                                                                              SSF    the sum of squares due to the factor, that is the sum of squares
                                                                                                     of differences between factor level averages and the grand
                                                                                                     average (difference between levels).
                                                                                                                a    n                      a            a
                                                                                                      SS F = ∑∑ ( yi. − y.. ) 2 = n∑ ( yi. − y.. ) 2 ≈ n∑ τ i2   (3.8)
                                                                                                               i =1 j =1                   i =1         i =1




                                                                                       3-19                                                                              3-20
SSE             the sum of squares due to error, that is the sum of squares           Alternative Formulae of Calculating the Sum of Squares
                of differences between observations and their level
                averages (difference within levels).                                                a    n
                                                                                                                   y . .2
                       a     n                   a     n
                                                                                         SS T =    ∑ ∑ yij 2 -
                                                                                                   i =1 j =1       an
                                                                                                                                                       (3.11)
                SS E = ∑∑ ( yij − yi. ) 2 ≈∑∑ ε ij
                                                 2
                                                                      (3.9)
                       i =1 j =1                 i =1 j =1                                          a
                                                                                                        yi. 2   y . .2
                                                                                          SS F =   ∑ n an
                                                                                                   i =1
                                                                                                              -                                        (3.12)


         SS T = SS F + SS E                                           (3.10)              SS E = SST - SS F                                            (3.13)


                                                                                      Formulae (3.7) to (3.9) are used to describe the underlying ideas.
                                                                                      Formulae (3.11) to (3.13) are used for actually computation.


                                                                               3-21                                                                             3-22




Machining Example (1)                                                                 If SSF is large, it is due to differences among the means at the
                                                                                      different factor levels. See equation (3.8), a large SSF means that τ2i
                             Table 3.7
                                                                                      (or |τi|) is great. It is an indication that the factor has significant effect
          Machine             Process capability Cp
                                                                                      on the response.
            USA                    2.1, 2.0, 1.9
                                                                                      Usually, SSF is standardized by taking SSF / SSE . By comparing SSF
           Japan                   0.9, 1.0, 1.1
                                                                                      to SSE, we can see how much variability is due to changing factor
                                                                                      levels and how much is due to error.
  a = 2, n = 3
      y1⋅ = 6        y2⋅ = 3          y⋅⋅ = 9                                         Machining Examples
      y1⋅ = 2        y2⋅ = 1          y⋅⋅ = 15
                                             .                                                     Table 3.8
                                                                                                       Effective ?          SSF/SSE
   SST = (2.12 + 2.0 2 + 1.9 2 + 0.9 2 + 1.0 2 + 1.12 ) − 9 2 / 6 = 1.54                 Example 1        Yes                37.5
   SS F = (6 2 + 32 ) / 3 − 9 2 / 6 = 1.50                                               Example 2         No                  0
   SS E = 1.54 − 1.50 = 0.04
                                                                               3-23                                                                             3-24
(2) Determine degrees of freedom (D.O.F.).                                    Machining Example (1)
                                                                                                   Table 3.9
   total number of observations N:    an
                                                                                     Machine        Process capability Cp
   degrees of freedom of SST:         N - 1 = an - 1             (3.14)                 USA          2.1, 2.0, 1.9

   degrees of freedom of SSF:         a-1                        (3.15)                 Japan        0.9, 1.0, 1.1

   degrees of freedom of SSE:         R (n - 1) = a (n – 1)      (3.16)          a = 2, n = 3
      where, (n - 1) is the degrees of freedom of errors in a run.               total number of observations N:          2×3=6
                                                                                 degrees of freedom of SST:      N–1 =6–1=5
Equation of DOF
                                                                                 degrees of freedom of SSF:      a–1 =2–1=1
   an - 1 = (a - 1) + a (n - 1)                                  (3.17)
                                                                                 degrees of freedom of SSE:      a (n-1) = 2 × (3 - 1) = 4
   DOF of SST = DOF of SSF + DOF of SSE
                                                                              Check equation (3.17): 5 = 1 + 4
                                                                       3-25                                                                           3-26




(3) Calculate mean square (MS).                                               (4) Calculate the test ratio F0.

   The ratio between a sum of squares and its DOF.                                       MS F
                                                                                 Fo =                                                        (3.20)
                                                                                         MS E

                 SS F
   MS F   =                                                   (3.18)
                a − 1                                                         If H0 is true, F0 follows a theoretical F distribution, which is
                                                                              completely determined by the two parameters ν1 and ν2.
                  SS E
   MS E =                                                     (3.19)             ν1 = a - 1     the D.O.F. of MSF   (numerator)
              a ( n - 1)
                                                                                 ν2 = a(n - 1) the D.O.F. of MSE    (denominator)

   There is no need to calculate MS for the SST


                                                                       3-27                                                                           3-28
f(x)                                                                    (5) Make conclusion.
                                ν1 = 1
                                ν2 = 4        (1) Decide the curve from
                                                                                                     SS F
                                                  ν1 and ν2.                                 MS F            a (n − 1) SS F     SS
                                              (2) Decide Fα,ν1,ν2 from α                Fo =      = a −1 =            ⋅     = Q⋅ F
                                                                                             MS E    SS E      a − 1 SS E       SS E
                                                                                                   a (n − 1)

                                                                                                a (n − 1)
                                                             α                     Where, Q =             is a constant.
                                                                                                  a −1

                                                                                                   SS F
                                               Fα ,ν1 ,ν 2                         F0 is equivalent to   (the standardized SSF) except for a constant Q.
                                                                                                   SS E
                             Figure 3.6                                                                                                       SS
    The density function curve of an example of the F distribution                 However, while F0 follows the theoretical F distribution, F does
                                                                                                                                              SS E
                                                                                   not.
                                                                            3-29                                                                       3-30




                       SS F                                                             Table 3.10 Control Limit Fα ,ν 1 ,ν 2 found from the F Table
If F0 is large ---->        is large ---> SSF is large ----> τ2i is large --->
                       SS E
    |τi| is large ---> the factor has significant effect on the response.                                           α = 0.05

Control limit: Fα ,ν 1 ,ν 2                                                              ν1     1         2        3       4      5        .       ∞
                                                                                   ν2
    ν1 = a - 1 and ν2 = a(n - 1) are the two parameters of an F                    1          161.4      199.5
    distribution
                                                                                    2         18.51      19.00
    α is the probability of Type I error, which means concluding that
    the factor has significant effect on the response when it in fact has           3         10.13      9.55
                                                                                                                               Examples:
    no effect. Usually, set α = 1% or 5%                                            4          7.71      6.94
                                                                                                                               F0.05,2,1 = 199.5
If Fo < Fα ,ν 1 ,ν 2 ,   conclude that the factor has no effect.                    .
                                                                                                                               F0.05,1,4 = 7.71
                                                                                   ∞
If Fo > Fα ,ν 1 ,ν 2 ,   conclude that the factor has significant effect.
                                                                            3-31                                                                       3-32
Table 3.11                                                                                     Table 3.12




                                                            3-33                                                                                     3-34




                           Table 3.13                              Machining Example (1) (a = 2, n = 3, N = 6), from Table 3.7
      Analysis of Variance for a Single-Factor Experiment
___________________________________________________                                                                      Table 3.14
                                                                   ___________________________________________________


Source of     Sum of   Degrees of   Mean       Fo                  Source of                          Sum of             Degrees of   Mean     Fo
Variation     Squares Freedom       Square                         Variation                          Squares            Freedom      Square
                                                                   ___________________________________________________
___________________________________________________
                                                                   Machine                               1.50              1 (= ν1)   1.50     150
Factor         SSF         a-1          MSF        MS F            Error                                 0.04              4 (= ν2)    0.01
                                                   MS E
                                                                   Total                                 1.54               5
                                                                   ___________________________________________________


Error          SSE        a(n-1)        MSE

Total         SST        an-1
___________________________________________________

                                                            3-35                                                                                     3-36
Specify α = 0.05, Control limit Fα ,ν 1 ,ν 2 = F0.05,1, 4 = 7.71                                                                           Machining Example (2) (a = 2, n = 3, N = 6),
                                                                                                                                                                                  Table 3.15
Since F0 > Fα ,ν 1 ,ν 2 , machines have significant effect on the process
                                                                                                                                           Source of                Sum of        Degrees of    Mean           Fo
                          capability.
                                                                                                                                           Variation                Squares       Freedom       Square
Sum of squares:                              1.50 + 0.04 = 1.54
                                                                                                                                           Machines                   0             1 (= ν1)      0            0
Degrees of freedom 1 + 4 = 5
                                                                                                                                           Error                     0.04           4 (= ν2)      0.01
                                                                                                                                           Total          0.04      5
                                                                                                                                           ___________________________________________________

                                                                                                                                           Specify α = 0.05, Control limit Fα ,ν 1 ,ν 2 = F0.05,1, 4 = 7.71

                                                                                                                                           Since F0 < Fα ,ν1 ,ν2 , machines don’t have significant effect on the
                                                                                                                                                                   process capability.
                                                                                                                                    3-37                                                                            3-38




                    Table 3.16 Example: Tensile strength of paper                                                                                        4    6
                                                                                                                                                                          y.. 2
___________________________________________________________________________________________________________                                    SS T = ∑ ∑ yij 2 -
     Hardwood                                                  Observations                                                                             i =1 j =1         an
                                   _____________________________________________________________________________________________

Concen. (%)                        1          2         3         4         5          6         Total              Aver.                                                              (383) 2
_________________________________________________________________________________________________________________________________                  = (7) 2 + (8) 2 + L + (20) 2 -              = 512.96
                                                                                                                                                                                         24
         5                         7          8        15 11                9        10             60             10.00
                                                                                                                                                         4
                                                                                                                                                           yi.2 y..2
       10                          12 17 13 18 19 15                                                94             15.67                       SS F = ∑        -
                                                                                                                                                      i =1 n     an
       15                          14 18               19 17 16 18                               102               17.00
                                                                                                                                                        (60) 2 + (94) 2 + (102) 2 + (127) 2 (383) 2
                                                                                                                                                    =                                      -        = 382.79
   20         19 25 22 23 18 20 127          21.17                                                                                                                       6                    24
__________________________________________________
                                                                                                                                               SS E = SS T - SS F
                                                                                                  383             15.96                             = 512.96 - 382.79 = 130.17
a = 4, n = 6, N = an = 24.


                                                                                                                                    3-39                                                                            3-40
Table 3.17 ANOVA for tensile strength of paper                                                                              SUMMARY: ANOVA for Single Factor Experiment
___________________________________________________________________________________________________________________________________

Source of                          Sum of                  Degrees of                       Mean                         Fo
Variation                          Squares                 Freedom                          Square                                              1. Decide the factor to be investigated and the following two
__________________________________________________________________________________________________________________________________                 parameters:
Hardwood                            382.79                       3 (= ν1)                   127.60                19.61                                1) number of levels a;
concentration                                                                                                                                          2) replicate n.
Error                                130.17                     20 (= ν2)                          6.51
                                                                                                                                                2. Carry out the experiments, obtain N (= an) observations yij by a
Total                                512.96                     23                                                                                 random manner.
______________________________________________________________________________________________________________________________________




Specify α = 0.01, Control limit Fα ,ν 1 ,ν 2 = F0.01,3, 20 = 4.94                                                                               3. Check if the residuals (estimates of errors) satisfy the
                                                                                                                                                   requirements.
Since F0 > Fα ,ν 1 ,ν 2 , hardwood concentration has significant effect on
                          the tensile strength of paper.                                                                                        4. Calculate sum of squares SST, SSF and SSE (3.11, 3.12, 3.13).


                                                                                                                                         3-41                                                                      3-42




5. Calculate the degrees of freedom for each of SST, SSF, SSE (3.14,                                                                            3.3 Test on Individual Level Means
   3.15, 3.16).
                                                                                                                                                ANOVA results in a single parameter F0, which can tell whether a
6. Calculate the mean squares MSF and MSE (3.18, 3.19).                                                                                         factor has effect on the response, or whether the mean response
                                                                                                                                                values will be different at different levels of the factor from an
7. Calculate the ratio F0 between MSF and MSE (3.20).                                                                                           overall viewpoint.

8. Specify a type I error α, and find the control value Fα ,ν 1 ,ν 2 from                                                                       However, ANOVA cannot decide the direction of influence of the
       the F distribution Tables (ν1 = a - 1, ν2 = a (n – 1) ).                                                                                 factors, nor identify which factor level mean is different from another
                                                                                                                                                level mean.
9. Conclude if the factor has significant effect on the response.




                                                                                                                                         3-43                                                                      3-44
The plots indicate that changing the hardwood concentration has a
                                                                         strong effect on the paper strength; specifically, higher hardwood
                                                                         concentrations produce higher observed paper strength.

                                                                         Box plots show the variability of the observations within a factor
                                                                         level and the variability between factor levels.

                                                                         However, a less subjective approach is required to test the individual
                                                                         level means.




          Figure 3.7 Box plots of paper strength example
                                                                  3-45                                                                       3-46




Multiple Comparison Procedures                                           One of the major problems with making comparisons among level
                                                                         means is that unrestricted use of these comparisons can lead to an
It is the techniques for making comparisons between two or more
                                                                         excessively high probability of a Type I error.
level means subsequent to an analysis of variance.
When we run an analysis of variance and obtain a significant F0          For example, if we have 10 levels in which the complete null
value, we have shown simply that the overall null hypothesis is false.   hypothesis is true (with α = .05)
We do not know which of a number of possible alternative
hypotheses is true.                                                      Ho: µ1 = µ2 = µ3 = … = µ10

   H1 :   µ1 ≠ µ2 ≠ µ3 ≠ µ4 ≠ µ5                                         A series t tests between all pairs of level means will lead to making at
Or H1:    µ1 ≠ µ2 = µ3 = µ4 = µ5                                         least one Type I error 57.8% of the time. In other words, the
                                                                         experimenter who thinks he is working at the (α = .05) level of
Multiple comparison techniques allow us to investigate hypotheses        significance is actually working at (α = 0.578).
that involve means of individual levels. For example, we might be
interested in whether µ1 is different from µ2, or whether µ2 is
different from µ3.
                                                                  3-47                                                                       3-48
The probability of making at least one Type I error increases as we        Least Significant Difference (LSD) Test is a useful tool to test the
increase the number of independent t tests we make between pairs of        differences between the individual level means.
level means. While it is nice to find significant differences, it is not
nice to find ones that are not really there. We need to find some way      The basic requirement for a LSD test is that the F0 for the overall
to make the comparisons we need but keep the probability of                analysis of variance (ANOVA) must be significant. If the F0 was not
incorrect rejections of Ho under control.                                  significant, no comparisons between level means are allowed. You
                                                                           simply declare that there are no differences among the level means
                                                                           and stop right there.

                                                                           On the other hand, if the overall F0 is significant, you can proceed to
                                                                           make any or all pairwise comparisons between individual level
                                                                           means by the use of a t test.




                                                                    3-49                                                                          3-50




The t distributions                                                        To test between the ith level mean (µi) and jth level mean (µj),
                                                                              Ho: µi = µj
                                                                              H1: µi ≠ µj                                                     (3.21)

                                                                           we calculate
                                                                                    yi⋅ − y j⋅
                                                                              t=                                                              (3.22)
                                                                                    2 MS E / n

                                                                           The value of MSE has already been obtained during the overall
                                                                           ANOVA. The DOF of t is equal to the DOF of MSE, i.e., a(n-1) or
t(30) is a t distribution with a DOF of 30.                                (N-a).
t(∞) means a standard normal distribution with (µ = 0 and σ = 1).          Then, using a two-tailed t test for a specified level of type I error α
                                                                           (e.g., α = 0.05), we can decide a pair of two-sided control limits
                                                                           ±tα(dof) from the t table.
                                                                    3-51                                                                          3-52
If the t value determined by Eq (3.22) falls within the limits ±tα(dof),
                                                                                                                                           we cannot reject the null hypothesis in (3.21). We will therefore
                                                                                                                                           conclude that there is no difference between µi and µj.

                                                                                                                                           On the other hand, if the t value falls beyond the limits ±tα(dof), we
                                                                                                                                           will reject the null hypothesis and conclude that there is a difference
                                                                                                                                           between µi and µj.




                                                                                                                                    3-53                                                                       3-54




                         The Example of Tensile Strength of Paper                                                                          Problem: if hardwood concentrations 5% and 20% produce the
                                                            Table 3.18                                                                     same paper strength ?
___________________________________________________________________________________________________________

     Hardwood                                                  Observations                                                                                        Ho : μ 1 = μ 4
                                   _____________________________________________________________________________________________
                                                                                                                                           Hypothesis:
Concen. (%)                        1         2          3         4         5          6         Total              Aver.                                          H1 : μ1 = μ 4
                                                                                                                                                                           /
_________________________________________________________________________________________________________________________________


         5                         7          8        15 11                9        10             60             10.00                   Five Steps:
       10                          12 17 13 18 19 15                                                94             15.67
                                                                                                                                           (1) Use Eq (3.22)
       15                          14 18               19 17 16 18                               102               17.00
   20         19 25 22 23 18 20 127          21.17                                                                                                  yi ⋅ − y j ⋅        10.00 − 21.17
                                                                                                                                              t=                    =                  = −7.58
__________________________________________________                                                                                                  2 MS E / n            2 × 6.51 / 6
a = 4, n = 6
                                                                                                                                               (Note: MSE = 6.51 was calculated during ANOVA, which is
y1⋅ = 10.00,                 y2⋅ = 15.67,                   y3⋅ = 17.00,                y4⋅ = 21.17                                            always carried out before the test on individual level means)

                                                                                                                                    3-55                                                                       3-56
(2) Determine the DOF, DOF = a (n – 1) = 4 × (6 – 1) = 20                   General Steps to Test the Individual Level Means

(3) Decide a type I error level. We use α = 0.01 as in the ANOVA for        (1) Calculate t using Eq (3.22) and find the MSE value obtained
    this example.                                                              during ANOVA.

(4) Find the limits from the t table: ±tα(dof) = ±t0.01(20) ≈ 3.00          (2) Determine the DOF, DOF = a (n – 1).

(5) Make conclusion: since the calculated t value (= -7.58) falls           (3) Specify a type I error α.
    beyond the limits ±3.00, we will reject the null hypothesis.
                                                                            (4) Find the control limits from the t table: ±tα(dof) based on α and
   Conclusion: There is a significant difference in paper strength              DOF.
               between using hardwood concentrations 5% and
               20%. Moreover, since t < 0, µ1 < µ4.                         (5) Make conclusion by comparing the calculated t value against the
                                                                                control limits.


                                                                     3-57                                                                       3-58




The Bonferroni Procedure                                                    To put this in a way that is slightly more useful to us, if you want the
                                                                            overall family-wise error rate to be no more than (α = 0.05), and you
In a Bonferroni procedure the family-wise error rate (Type I error) is      want to run c tests, then each run of them should be at (α' = 0.05/c).
divided by the number of comparisons.                                       To run these tests, you do exactly what you did in Fisher's LSD test,
                                                                            though you omit any requirement about the significance of overall F.
The basic idea behind this procedure is that if you run several tests
(say c tests), each at a significance level (type I error) represented by
α', the probability of at least one Type I error can never exceed cα'.

Thus, for example, if you ran 5 tests, each at α' = 0.05, the family-
wise error rate would be at most 5(0.05) = 0.25. But that is a too high
Type I error rate to make anyone happy.

But suppose that you ran each of those 5 tests at the α' = 0.01. Then
the maximum family-wise error rate would be 5(0.01) = 0.05, which
is certainly acceptable.
                                                                     3-59                                                                       3-60
a     ni           2
                                                                                                    y..
                                                                          SST = ∑∑ yij −
3.5 Unbalanced Experiments                                                          2
                                                                                                            DOF = N -1
                                                                                  i =1 j =1         N
In an unbalanced experiment, the number of observations taken under
each run is different.                                                             a
                                                                                         yi2. y..
                                                                                                2
                                                                          SS F = ∑           −              DOF = a -1
ni     is the number of observations taken under the ith run (for                 i =1   ni N
       single factor experiment, it is simply the ith level).
                                                                                                                       a
             a                                                            SS E = SST − SS F                 DOF =    ∑ (ni − 1)
       N = ∑ ni = n1 + n2 + L + na                                                                                    i =1
            i =1

       is the total number of observations.

Note, in unbalanced experiments, N ≠ an.



                                                               3-61                                                                          3-62




Machining Example                                                     SST = (2.12 + 2.0 2 + 1.9 2 ) + (0.9 2 + 1.0 2 ) - 7.9 2 / 5 = 1.348

                                                                      DOET = N – 1 = 5 – 1 = 4
       Machine         Process capability Cp
         USA             2.1, 2.0, 1.9                                SSF = (6.02/3 + 1.92/2) – 7.92/5 = 1.323
         Japan           0.9, 1.0,
                                                                      DOEF = a – 1 = 2 – 1 = 1

n1 = 3, n2 = 2, N = 3 + 2 = 5
                                                                      SSE = 1.348 – 1.323 = 0.025
y1. = 6.0, y2. = 1.9, y.. = 7.9
                                                                      DOEE = (n1 – 1) + (n2 – 1) = (3 – 1) + (2 – 1) = 3




                                                               3-63                                                                          3-64
ANOVA for Machining Example (a = 2, n1 = 3, n2 = 2, N = 5)                    3.7 Guidelines for Designing Experiments
                                                                              (1) Recognition and statement of the problem.
Source of          Sum of      Degrees of        Mean           Fo             Fully develop all ideas about the problem and the specific
Variation          Squares     Freedom           Square                        objectives of the experiment;
                                                                               Solicit input from all concerned parties -- engineering, quality,
Machine             1.323         1 (= ν1)        1.323        158.8           marketing, customer, management and operators.
Error               0.025         3 (= ν2)        0.0083
                                                                              (2) Choice of factors and levels.
Total          1.348     4                                                     Choose the factors, their ranges and levels. Process knowledge is
___________________________________________________                            required.
Specify α = 0.05, Control limit: Fα ,ν 1 ,ν 2 = F0.05,1,3 = 10.13              Investigate all factors that may be of importance and avoid being
                                                                               overly influenced by past experience.
Since F0 > Fα ,ν 1 ,ν 2 , machines have significant effect on the process      Keep the number of factor levels low (Most often two levels are
                           capability.                                         used.) in the early stage
                                                                       3-65                                                                     3-66




(3) Selection of the response                                                 (6) Data analysis
 Most often, the average or standard deviation (or both) of the                Analyze the data so that results and conclusions are objective rather
 measured characteristic will be the response variable.                        than judgmental.
                                                                               Use software packages and simple graphical methods
(4) Choice of experimental design.
                                                                               Carry out residual analysis.
 Decide the factorial fraction experiment, select the replicate
 (sample size) and run order for the experimental trials.
                                                                              (7) Conclusions and recommendations.
 Decide whether or not blocking or other randomization methods are
 involved.                                                                    (8) Experimentation is an important part of the learning process.
                                                                                New hypotheses are formulated based on the investigation of the
(5) Performing the experiment.                                                  tentative hypotheses. Don’t design a single, large comprehensive
                                                                                experiment at the start of a study. As a rule of thumb, the first
 Ensure that everything is being done accordingly to plan.
                                                                                experiment should spend no more than 25% of the total budget of
                                                                                the study.

                                                                       3-67                                                                     3-68
4 RANDOMIZED BLOCKS                                                         Response:     coded roughness of a machined surface.

4.1 Randomized Block Design                                                 Main factor (factor to be studied in the experiment)
Nuisance factor probably has an effect on the response, but we are             Cutter of a machine tool
not interested in that effect.
                                                                            Known nuisance factor (factor not to be studied in the experiment,
For unknown and uncontrollable nuisance factors, use Complete               but we are well aware of the existence of its effect on the response)
Randomization technique to get rid of them.
                                                                               Different machine tools of the same type
For known and controllable nuisance factors, use Randomized
Blocking technique instead.                                                 Unknown nuisance factor (unknown and uncontrollable factors,
                                                                            which may have some effect on the response)
The variability of the nuisance factor will contribute to the variability
observed in the experimental data. As a result, the experimental error         Temperature, humidity
will reflect both the random error of the unknown nuisance factors
and variability of the known nuisance factor.
                                                                      4-1                                                                          4-2




                              Table 4.1                                            MS F   MS F         DOEE
                                                                            F0 =        =       = MS F                                    (4.4)
                                 Machine tool (Block)                              MS E    SS E         SS E
            Cutter             1      2       3         4                                 DOEE
              1               -2     -1       1         5
              2               -1     -2       3         4                   So, whether F0 will be increased by the addition of the known
              3               -3     -1       0         2                   nuisance factors depends on which of SSE or DOEE will increase
              4               2       1       5         7                   more.
                                                                            If SSE increases more by including the known nuisance factors, F0
SST = SS F + SS known + SSunknown                            (4.1)          becomes less sensitive. Therefore, Randomized Blocking technique
                                                                            is preferred.
If the effects of the known nuisance factors are included in the error      If DOEE increases more, F0 becomes more sensitive. Therefore,
                                                                            Complete Randomization technique is preferred.
SS E = SS known + SSunknown                                  (4.2)
                                                                            Note, the increase of DOEE also reduces the control limit Fα ,ν 1 ,ν 2 (ν2
DOEE = DOEknown + DOEunknown                                 (4.3)          = DOEE), and therefore, makes F0 relatively more sensitive.
                                                                      4-3                                                                          4-4
Randomized Block Design                                                              Block 1          Block 2                       Block b
   It means blocking all the runs according to the different levels of a
   known nuisance factor. Within a block, the order in which the                       y11                 y12                        y1b
   runs of the main factors are tested is randomly determined.
                                                                                       y21                 y22                        y2b
   The blocks form a more homogeneous experimental unit in terms
   of the known nuisance factor.                                                                                    .............
                                                                                       y31                 y32                        y3b
   Effectively, this design strategy improves the accuracy of the                       .                   .                          .
   comparisons of the responses by eliminating the variability of the                   .                   .                          .
   known nuisance factor.                                                               .                   .                          .
                                                                                       ya1                 ya2                        yab
Examples of blocks
   Equipment, machinery, batches of raw material, people, time slot                                        Figure 4.1


                                                                     4-5                                                                            4-6




The statistical model                                                      Hypotheses of interest are
                                ⎧ i = 1,2,L, a                                  H0: μ1 = μ2 = = … μa
   yij = μ + τ i + β j + ε ij   ⎨                           (4.5)
                                ⎩ j = 1,2,L, b                                  H1: at least one μi ≠ μj                                    (4.6)
where, μ is an overall mean, τi is the effect of the ith run, βj is the    Or
effect of the jth block, and εij is the usual NID(0, σ2) random error           H0: τ1 = τ2 = … = τa = 0
term.
                                                                                H1: at least one τi ≠ 0                                     (4.7)

                                                                           βj will not be studied, because it is the effect of the known nuisance
                                                                           factor.




                                                                     4-7                                                                            4-8
yi.                the total of all observations taken under run i (total of row)      yi. the average of the observations taken under the ith run
y.j                the total of all observations in block j (total of column)          y. j the average of the observations in block j
y..                the grand total of all observations
                                                                                       y.. the grand average of all observations.
N = ab             the total number of observations.
               b
                                                                                          yi. = yi . / b   y. j = y . j / a   y.. = y.. / N                (4.9)
      yi. = ∑ yij              i = 1,2,L, a
            j =1
               a
      y. j = ∑ yij             j = 1,2,L, b                           (4.8)
            i =1
            a          b
      y.. = ∑∑ yij
           i =1 j =1




                                                                                 4-9                                                                               4-10




Sum of square                                                                          Degrees of Freedom

      SST = SS F + SS Block + SS E                                    (4.10)              SST              N - 1 = ab - 1

                   a       b
                                  y..2                                                    SSF              a-1
      SST = ∑∑ yij 2 -                                                (4.11)
                i =1 j =1         N
                                                                                          SSBlocks         b-1
                       a           2
               1          y..
      SS F =     ∑ yi.2 - N
               b i=1
                                                                      (4.12)              SSE              (ab − 1) − (a − 1) − (b − 1) = ab − a − b + 1      (4.15)

                1 b 2 y..2
      SS Block = ∑ y. j -                                             (4.13)
                a j =1    N
      SS E = SST - SS F − SS Block                                    (4.14)



                                                                                4-11                                                                               4-12
Test Statistic                                                                                   Table 4.2 ANOVA Table
                                                                               _______________________________________________________
           MS F                                                                Source of    Sum of      Degrees of  Mean           F0
    F0 =
           MS E                                                                Variation    Squares     Freedom     Square
           SS F                                                                _______________________________________________________
    MS F =                                                     (4.16)
           a −1                                                                                                                   SS F               MS F
                                                                               Factor            SS F            a -1                         Fo =
                SS E                                                                                                              a −1               MS E
    MS E =
           ab − a − b + 1                                                                                                        SS Blocks
                                                                               Blocks            SSBlocks        b–1
We would reject Ho in (4.7), if Fo > Fα,(a-1)(ab-a-b+1).                                                                          b −1
                                                                                                                                 SS E
                                                                               Error              SS E      ab – a – b + 1
                                                                                                                             ab − a − b + 1
                                                                               Total         SST        N -1
                                                                               _______________________________________________________
                                                                        4-13                                                                           4-14




We may also examine the ratio of MSBlocks (= SSBlocks/(b-1)) to MSE. If        Example 4.1
this ratio is large, it implies that the blocking factor has a large effect
                                                                                                               Table 4.3
and that the noise reduction obtained by blocking was probably                                                 Machine tool (Block)
                                                                                        Cutter               1       2      3            4       yi.
helpful in improving the effectiveness of the experiment. Otherwise,
                                                                                          1                 -2      -1      1            5        3
blocking may actually be unnecessary, i.e, the known nuisance factor                      2                 -1      -2      3            4        4
should be handled as random error.                                                        3                 -3      -1      0            2       -2
                                                                                          4                 2       1       5            7       16
                                                                                          y.j               -4          -3   9          18     y.. = 20




                                                                        4-15                                                                           4-16
The sums of squares are obtained as follows:                                                          Table 4.4 ANOVA Table
                                                                                    _______________________________________________________
            a     b
                             y..2         20 2                                      Source of    Sum of      Degrees of  Mean          F0
   SST = ∑∑ yij 2 -               = 154 −      = 129
            i =1 j =1        N            16                                        Variation    Squares     Freedom     Square
                                                                                    _______________________________________________________
                   = (3 + 4 + (−2) + 15 ) −
         1 a 2 y..2 1 2                     20 2
   SS F = ∑ yi. -          2      2    2
                                                 = 38.5                             Cutter          38.50            3           12.83         14.44
         b i=1    N 4                       16
                                                                                    Machine         82.50            3           27.50

                  ∑ y.2j - N = 4 ((−4) 2 + (−3) 2 + 92 + 182 ) − 16 = 82.5
                      b           2                              2                  (Block)
                1          y.. 1                                 20
   SS Block =
                a j =1                                                              Error            8.00            9           0.89

   SS E = SST - SS F − SS Block = 129 − 38.5 − 82.5 = 8                             Total         129.00     15
                                                                                    _______________________________________________________

                                                                                    Specify α = 0.05, F0.05,3,9 = 3.86. Since 14.44 > 3.86, we conclude
                                                                                    that the type of cutter affects the mean hardness reading.
                                                                             4-17                                                                   4-18




If the randomized block design is not used, F0.05,3,12 = 3.49, the
hypothesis of equal mean values from the four cutters cannot be
rejected.

                             Table 4.5 ANOVA Table
Source of               Sum of        Degrees of       Mean            F0
Variation               Squares       Freedom          Square

Cutter                    38.50             3           12.83        1.70
Error                     90.50             12          7.54
Total                     129.00       15


Here, the error is actually the sum of error and block in Table 4.4.

                                                                             4-19
5 FACTORIAL EXPERIMENTS                                                      Case one
                                                                                    Two factors: A: Machine, B: Feed rate
5.1 Introduction
                                                                                                   Response:          Process capability
Factorial experiments study the effects of several factors on the
response.                                                                                                   Table 5.1
                                                                                                            Process capability Cp
In factorial experiments, all possible runs (combinations of the levels                                      Feed rate (mm/sec.)
of the factors) are investigated.                                                      Machine            0.2        0.4        0.6
For example, for two factors A: temperature (with a levels) and B:                      USA            2.1, 2.2 1.7, 1.6 1.2, 1.3
humidity (with b levels), the factorial experiment has ab possible                      Japan          0.9, 1.0 0.7, 0.6 0.4, 0.3
runs.
                                                                                  a = 2, b = 3, n = 2, R =ab = 6
If for each run, n observations are taken, then, the total number N of            N = Rn = abn = 2 × 3 × 2 = 12
the observations is equal to abn.

                                                                       5-1                                                                                5-2




Case two                                                                     5.2 ANOVA for Two-Factor Factorial Experiments
  Three factors: A, B and C.
                                                                             Mathematical model
  a = 2, b = 3, c = 4, n = 2, R = abc = 24
                                                                             Yijk = μ + τ i + β j + (τβ )ij + ε ijk                               (5.1)
  N = Rn = abcn = 2 × 3 × 4 × 2 = 48
                                                                             μ                 the overall mean
                                                                             τi                the effect incurred at the ith level of factor A
Number of runs is equal to the product of the levels of all factors.
                                                                             βj                the effect incurred at the jth level of factor B
                                                                             (τβ)ij            the effect incurred at the interaction between the ith
                                                                                               level of A and the jth level of B
                                                                             εijk              the random error (normally and independently
                                                                                               distributed, zero mean and constant variance)

                                                                       5-3                                                                                5-4
Interaction indicates the influence of one factor on the effect of        Table 5.2 Data Arrangement for a Two-Factor Factorial Design
             another factor, and vice versa.                                                                Factor B
                                                                                             1           2           …        b         Total   Average
Factors           A: temperature                                                     1   y111, y112, y121, y122,     …    y1b1, y1b2,    y1..     y1..
                  B: pressure                                             Factor A        …, y11n     …, y12n              …, y1bn

                                                                                     2   y211, y212, y221, y222,   …     Y2b1, y2b2,     y2..     y 2..
Response          Process yield                                                           …, y21n     …, y22n             …, y2bn
                                                                                     :       …           …         …        …            …        …
If:    (1) Increasing temperature alone cannot increase the yield; and
                                                                                     a   ya11, ya12, ya21, ya22,   …      yab1, yab2,    ya..     y a..
       (2) Increasing pressure alone cannot increase the yield; and                       …, ya1n     …, ya2n              …, yabn
       (3) Increasing temperature and pressure at the same time does       Total            y.1.        y.2.       …         y.b.        y…
            increase the yield                                            Average            y .1.       y .2.     …          y .b.               y ...


Then: There is interaction between temperature (factor A) and             yijk is the kth observation taken at the ith level of factor A and the jth
      pressure (factor B)                                                       level of factor B.
                                                                          The total number of observations: N = abn
                                                                    5-5                                                                                   5-6




Total and average at the ith level of factor A:                           Hypotheses
               b    n
                                y
      yi .. = Σ Σ yijk  y i .. = i .. i = 1,2,...,a                       Factor A:
              j =1 k =1         bn                          (5.2)
                                                                                H o : τ 1 = τ 2 = ... = τ a = 0     Factor A has no effect on y
Total and average at the jth level of factor B:
                 a   n           y. j .                                         H 1 : at least one τ i ≠ 0          Factor A has effect on y
      y. j . = Σ Σ yijk  y. j. =        j = 1,2,..., b
               i =1 k =1         an                         (5.3)         Factor B:

Total and average when A at the ith level and B at the jth level:               H 0 : β1 = β2 =L = βb = 0           Factor B has no effect on y
               n             y                                                  H 1 : at least one β j ≠ 0
      yij . = Σ yijk y ij . = ij .                                                                                  Factor B has effect on y
              k =1            n                             (5.4)
                                                                          Interaction AB:
Grand total and average
              a    b    n
                                   y...                                         H o : (τβ ) ij = 0 for all i, j     Interaction AB has no effect y
       y...= ΣΣΣ yijk y...=                                 (5.5)
             i =1 j =1 k =1        abn                                          H1 : at least one (τβ ) ij ≠ 0     Interaction AB has effect on y
                                                                    5-7                                                                                   5-8
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering
quality engineering

More Related Content

Similar to quality engineering

Applied Reliability Symposium 2009 M Turner
Applied Reliability Symposium 2009 M TurnerApplied Reliability Symposium 2009 M Turner
Applied Reliability Symposium 2009 M TurnerMark Turner CRP
 
Matlab 3
Matlab 3Matlab 3
Matlab 3asguna
 
Solution manual for design and analysis of experiments 9th edition douglas ...
Solution manual for design and analysis of experiments 9th edition   douglas ...Solution manual for design and analysis of experiments 9th edition   douglas ...
Solution manual for design and analysis of experiments 9th edition douglas ...Salehkhanovic
 
Fault simulation – application and methods
Fault simulation – application and methodsFault simulation – application and methods
Fault simulation – application and methodsSubash John
 
Principles of design of experiments (doe)20 5-2014
Principles of  design of experiments (doe)20 5-2014Principles of  design of experiments (doe)20 5-2014
Principles of design of experiments (doe)20 5-2014Awad Albalwi
 
Using capability assessment during product design
Using capability assessment during product designUsing capability assessment during product design
Using capability assessment during product designMark Turner CRP
 
Model-Based Testing: Theory and Practice. Keynote @ MoTiP (ISSRE) 2012.
Model-Based Testing: Theory and Practice. Keynote @ MoTiP (ISSRE) 2012.Model-Based Testing: Theory and Practice. Keynote @ MoTiP (ISSRE) 2012.
Model-Based Testing: Theory and Practice. Keynote @ MoTiP (ISSRE) 2012.Wolfgang Grieskamp
 
Qm0021 statistical process control
Qm0021 statistical process controlQm0021 statistical process control
Qm0021 statistical process controlsmumbahelp
 
DS-004-Robust Design
DS-004-Robust DesignDS-004-Robust Design
DS-004-Robust Designhandbook
 
Gestión de la calidad sem 2
Gestión de la calidad sem 2Gestión de la calidad sem 2
Gestión de la calidad sem 2youffre
 
デザインキット・D級アンプのスタートアップガイド
デザインキット・D級アンプのスタートアップガイドデザインキット・D級アンプのスタートアップガイド
デザインキット・D級アンプのスタートアップガイドTsuyoshi Horigome
 
Bridging the Gap: Machine Learning for Ubiquitous Computing -- Evaluation
Bridging the Gap: Machine Learning for Ubiquitous Computing -- EvaluationBridging the Gap: Machine Learning for Ubiquitous Computing -- Evaluation
Bridging the Gap: Machine Learning for Ubiquitous Computing -- EvaluationThomas Ploetz
 
JF608: Quality Control - Unit 5
JF608: Quality Control - Unit 5JF608: Quality Control - Unit 5
JF608: Quality Control - Unit 5Asraf Malik
 

Similar to quality engineering (20)

Applied Reliability Symposium 2009 M Turner
Applied Reliability Symposium 2009 M TurnerApplied Reliability Symposium 2009 M Turner
Applied Reliability Symposium 2009 M Turner
 
Catapult DOE Case Study
Catapult DOE Case StudyCatapult DOE Case Study
Catapult DOE Case Study
 
Matlab 3
Matlab 3Matlab 3
Matlab 3
 
Robust Design
Robust DesignRobust Design
Robust Design
 
Solution manual for design and analysis of experiments 9th edition douglas ...
Solution manual for design and analysis of experiments 9th edition   douglas ...Solution manual for design and analysis of experiments 9th edition   douglas ...
Solution manual for design and analysis of experiments 9th edition douglas ...
 
Junit 4.0
Junit 4.0Junit 4.0
Junit 4.0
 
Fault simulation – application and methods
Fault simulation – application and methodsFault simulation – application and methods
Fault simulation – application and methods
 
Principles of design of experiments (doe)20 5-2014
Principles of  design of experiments (doe)20 5-2014Principles of  design of experiments (doe)20 5-2014
Principles of design of experiments (doe)20 5-2014
 
Using capability assessment during product design
Using capability assessment during product designUsing capability assessment during product design
Using capability assessment during product design
 
Why do a designed experiment
Why do a designed experimentWhy do a designed experiment
Why do a designed experiment
 
Gage r&r
Gage r&rGage r&r
Gage r&r
 
Model-Based Testing: Theory and Practice. Keynote @ MoTiP (ISSRE) 2012.
Model-Based Testing: Theory and Practice. Keynote @ MoTiP (ISSRE) 2012.Model-Based Testing: Theory and Practice. Keynote @ MoTiP (ISSRE) 2012.
Model-Based Testing: Theory and Practice. Keynote @ MoTiP (ISSRE) 2012.
 
Qm0021 statistical process control
Qm0021 statistical process controlQm0021 statistical process control
Qm0021 statistical process control
 
DS-004-Robust Design
DS-004-Robust DesignDS-004-Robust Design
DS-004-Robust Design
 
Gestión de la calidad sem 2
Gestión de la calidad sem 2Gestión de la calidad sem 2
Gestión de la calidad sem 2
 
デザインキット・D級アンプのスタートアップガイド
デザインキット・D級アンプのスタートアップガイドデザインキット・D級アンプのスタートアップガイド
デザインキット・D級アンプのスタートアップガイド
 
The lab report
The lab reportThe lab report
The lab report
 
9. design of experiment
9. design of experiment9. design of experiment
9. design of experiment
 
Bridging the Gap: Machine Learning for Ubiquitous Computing -- Evaluation
Bridging the Gap: Machine Learning for Ubiquitous Computing -- EvaluationBridging the Gap: Machine Learning for Ubiquitous Computing -- Evaluation
Bridging the Gap: Machine Learning for Ubiquitous Computing -- Evaluation
 
JF608: Quality Control - Unit 5
JF608: Quality Control - Unit 5JF608: Quality Control - Unit 5
JF608: Quality Control - Unit 5
 

Recently uploaded

Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Alan Dix
 
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxMaking_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxnull - The Open Security Community
 
Artificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraArtificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraDeakin University
 
Unleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubUnleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubKalema Edgar
 
APIForce Zurich 5 April Automation LPDG
APIForce Zurich 5 April  Automation LPDGAPIForce Zurich 5 April  Automation LPDG
APIForce Zurich 5 April Automation LPDGMarianaLemus7
 
08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking MenDelhi Call girls
 
Benefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other FrameworksBenefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other FrameworksSoftradix Technologies
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreternaman860154
 
How to convert PDF to text with Nanonets
How to convert PDF to text with NanonetsHow to convert PDF to text with Nanonets
How to convert PDF to text with Nanonetsnaman860154
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsRizwan Syed
 
Build your next Gen AI Breakthrough - April 2024
Build your next Gen AI Breakthrough - April 2024Build your next Gen AI Breakthrough - April 2024
Build your next Gen AI Breakthrough - April 2024Neo4j
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationSafe Software
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsMemoori
 
CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):comworks
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Allon Mureinik
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticscarlostorres15106
 
Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for PartnersEnhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for PartnersThousandEyes
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 

Recently uploaded (20)

E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptxE-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
 
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
 
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptxMaking_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
Making_way_through_DLL_hollowing_inspite_of_CFG_by_Debjeet Banerjee.pptx
 
Artificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning eraArtificial intelligence in the post-deep learning era
Artificial intelligence in the post-deep learning era
 
Unleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubUnleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding Club
 
APIForce Zurich 5 April Automation LPDG
APIForce Zurich 5 April  Automation LPDGAPIForce Zurich 5 April  Automation LPDG
APIForce Zurich 5 April Automation LPDG
 
08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men
 
Benefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other FrameworksBenefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other Frameworks
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreter
 
How to convert PDF to text with Nanonets
How to convert PDF to text with NanonetsHow to convert PDF to text with Nanonets
How to convert PDF to text with Nanonets
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL Certs
 
Build your next Gen AI Breakthrough - April 2024
Build your next Gen AI Breakthrough - April 2024Build your next Gen AI Breakthrough - April 2024
Build your next Gen AI Breakthrough - April 2024
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
 
DMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special EditionDMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special Edition
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial Buildings
 
CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
 
Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for PartnersEnhancing Worker Digital Experience: A Hands-on Workshop for Partners
Enhancing Worker Digital Experience: A Hands-on Workshop for Partners
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 

quality engineering

  • 1. M6141 QUALITY ENGINEERING CONTENTS 1. Introduction DESIGN OF EXPERIMENTS 2. Fundamentals of Probability and Statistics (cancelled) A/P Wu Zhang 3. Single Factor Experiments MAE, NTU 4. Randomized Blocks (cancelled) Office: N3.2-02-14 Tel: 6790 4445 5. Factorial Experiments Email: mzwu@ntu.edu.sg 6. 2k Experiments 7. Fractional Experiments TEXT 8. A Case Study Montgomery, D. C., (2009), Design and Analysis of Experiments, John Wiley & Sons. 1-1 1-2 1 INTRODUCTION Example 1.1: Two factors: Machine, Feed rate Response: Process capability A designed experiment is a test or series of tests, in which, the experimenter chooses certain factors for study. He purposely varies Table 1.1 those factors in a controlled fashion, and then observes the effect of Process capability Cp such factors on the response function. Feed rate (mm/sec.) Machine 0.2 0.4 0.6 USA 2.1, 2.2 1.7, 1.6 1.2, 1.3 Japan 0.9, 1.0 0.7, 0.6 0.4, 0.3 Conclusion: using USA machine at lowest feed rate will result in max. process capability. 1-3 1-4
  • 2. Factor (input variable) (3) Run: A run is one set of levels for all factors in an experiment. In above experiment, there are three factors: (A) temperature, (1) Types (B) machine and (C) switch. The total number of runs is Quantitative: Temperature, time: They can be set at any value. R = 4 × 3 × 2 = 24 Qualitative: Machines, operators: They have different categories Switch: It has different status For example, [1] A = 10, B = Japan-made, C = on .......... (2) Levels [24] A = 40, B = UK-made, C = off Temperature (oC): 10, 20, 30, 40 [4 levels] Machine: made by Japan, by USA, by UK [3 levels] For single factor experiments, run is equivalent to level of the factor. The number R of runs is equal to the number a of levels. Switch: on, or off [2 levels] (4) Replicates (n): number of observations at each run. A larger value of n increases the accuracy of the experimental results, but makes the experiment more expensive. 1-5 1-6 Run A B C Run A B C Response (or output, observation) No Temp. Machine Switch No Temp. Machine Switch 1 10 Japan On 13 30 Japan On Response is dependent on the values of the factors. 2 10 Japan Off 14 30 Japan Off 3 10 USA On 15 30 USA On Types of response: 4 10 USA Off 16 30 USA Off 5 10 UK On 17 30 UK On Larger the better: product yield, process capability, student’s 6 10 UK Off 18 30 UK Off mark 7 20 Japan On 19 40 Japan On Smaller the better: product cost, number of errors 8 20 Japan Off 20 40 Japan Off 9 20 USA On 21 40 USA On Nominal the best: diameter of a shaft 10 20 USA Off 22 40 USA Off 11 20 UK On 23 40 UK On By DOE, we can find out which factors have effects on the response, 12 20 UK Off 24 40 UK Off and the direction of the influence (i.e., whether the increase of a factor will increase or decrease the response ?) 1-7 1-8
  • 3. Example 1.2: Single factor: A text book (a = 2, n = 3, R = 2) Example 1.3: Two factors: A text book, B studying hours Response: Students’ marks Response: Students’ marks Table 1.2 (a = 2, b = 3, n = 2, R = 6) Students’ marks Table 1.3 Text 1 76, 74, 78 Ave = 76 Students’ marks Text 2 48, 50, 52 Ave = 50 1 hr 2 hr 3 hr Text 1 45, 47 59, 58 76, 75 Text 2 32, 33 51, 54 70, 71 Text book (a qualitative factor) Text 1: book by Montgomery Text 2: book by Taguchi 1-9 1-10 The objectives of an experiment may include What Makes Designed Experiment (DOE) Special? (1) Determining which factors xs are most influential on the response, (1) It conducts the experiments in a systematic and efficient way. y. Reduce design and development time, as well as the cost. (2) It presents experiment results in the simplest and clearest way. (2) Determining where to set the influential x's, so that y is near the desired requirement. (3) It extracts the maximum amount of information from a given set of experiments. (3) Determining where to set the influential x's, so that variability in y is small. (4) It draws the right conclusions, despite the variability presented in the data. Examples: Increase process capability Improve students’ study Reduce output variability. 1-11 1-12
  • 4. Example 1.4 Reducing Defect Level in a Process In this application, DOE A manufacturing engineer is going to apply DOE to a process for (1) Determines which factors affect the occurrence of defects soldering electronic components to printed circuit boards, in order to reduce defect levels even further. (2) Determines the direction of adjustment for the effective factors to Factors to be investigated: further reduce the number of defects per unit. Solder temperature Preheat temperature (2) Decides whether changing the factors together produces different Conveyor speed results than just adjusting individual factor separately? Flux type Flux specific gravity Solder wave depth Conveyor angle Thickness of the printed circuit board Types of components used on the board 1-13 1-14
  • 5. 3 SINGLE FACTOR EXPERIMENTS Table 3.1 Tensile strength of paper ___________________________________________________________________________________________________________ Hardwood Observations 3.1 Introduction _____________________________________________________________________________________________ Concen. (%) 1 2 3 4 5 6 Total Aver. _________________________________________________________________________________________________________________________________ Single factor experiment can be used in the situation where the effect (influence) of a single factor on the response is dominant. The effects 5 7 8 15 11 9 10 60 10.00 of all other factors are negligible. 10 12 17 13 18 19 15 94 15.67 The procedures and formulae developed for the single factor 15 14 18 19 17 16 18 102 17.00 experiments can be easily modified, and then used for the several 20 19 25 22 23 18 20 127 21.17 factors experiments. __________________________________________________ 383 15.96 Study the effect (influence) of hardwood concentration on the tensile strength of paper. 3-1 3-2 (1) One factor: hardwood concentration. Table 3.3 General data table for a single factor experiment (2) Four levels: 5%, 10%, 15%, 20%: a = 4. Level Observation Total Average 1 y11 y12 . . y1n y1. y1. (3) Response is the tensile strength of paper. 2 y21 y22 . . y2n y2. y2. (4) Each level has six observations or replicates: n = 6. . . . . . . (5) Total number of observations: N = an = 24. . . . . . . a ya1 ya2 . . yan ya. ya. y.. y.. yij the jth observation taken under the ith level of the factor. 3-3 3-4
  • 6. yi. the total of the observations at the ith level 3.2 The Analysis of Variance (ANOVA) n ANOVA is the most important statistical tool used in DOE. Its task yi. = ∑ yij (3.1) j =1 is to decide whether and which factors or interactions have significant effect (influence) on the response. yi. the average of the observations at the ith level yi. = yi. / n (3.2) Model of ANOVA Yij = μ + τ i + ε ij = μi + ε ij (3.5) y.. the grand total of all observations a n a y.. = ∑∑ yij = ∑ yi. (3.3) μ the overall mean. μ ≈ y.. (mean ≈ average) i =1 j =1 i =1 μi the ith level mean. μ i ≈ yi. y.. the grand average of all observations. τi the ith level effect. τ i = μi − μ ≈ yi. − y.. y.. = y.. / N (3.4) ε ij the random error. ε ij = yij − μ i ≈ yij − yi. N = an is the total number of observations 3-5 3-6 Machining Example (1) y1. = 2.1 + 2.0 + 1.9 = 6.0, y1. = y1. / 3 = 2.0 Study the effect of machine on the process capability y 2. = 0.9 + 1.0 + 1.1 = 3.0, y 2. = y 2. / 3 = 1.0 Table 3.4 y.. = (2.1 + 2.0 + 1.9) + (0.9 + 1.0 + 1.1) = 9.0 Machine Process capability Cp y.. = 9.0 / 6 = 1.5 USA 2.1, 2.0, 1.9 Japan 0.9, 1.0, 1.1 μ1 ≈ y1. = 2.0, μ 2 ≈ y 2. = 1.0 μ ≈ y.. = 1.5 (1) One factor: machine (qualitative). τ1 = μ1 − μ ≈ y1. − y.. = 2.0 − 1.5 = 0.5 (2) Two levels: USA-made, Japan-made, a = 2. τ 2 = μ 2 − μ ≈ y2. − y.. = 1.0 − 1.5 = −0.5 (3) Response is the process capability. (4) Each level has 3 observations or replicates: n = 3. (5) Total number of observations: N = an = 2 × 3 = 6. 3-7 3-8
  • 7. Evaluation of Errors y ε11 = 0.1 For instance, if i = 1, j = 1, yij = y11 = 2.1 ε ij = yij − μi μ1 = 2 τ1 = 0.5 ε11 = y11 − μ1 ≈ y11 − y1. = 2.1 − 2.0 = 0.1 μ = 1.5 y11 = 2.1 τ2 = -0.5 Errors for the machining example (1) μ2 = 1 Table 3.5 Machine Errors USA 0.1, 0.0, -0.1 y11 = μ + τ1 + ε11 Japan -0.1, 0.0, 0.1 = 1.5 + 0.5 + 0.1 = 2.1 Figure 3.2 3-9 3-10 μ1 μ2 μ1 τ1 τ2 τ1 μ1 μ2 μ1 μ2 μ3 μ μ τ1 τ2 τ3 μ μ3 μ4 μ τ2 τ3 τ4 μ2 μ3 μ4 τ3 τ4 τ4 μ3 μ4 μ4 Figure 3.3 Figure 3.4 3-11 3-12
  • 8. Hypothesis Test: to test if a factor has effect on the response. Null: Ho : the factor has no effect on the response If a factor has no effect on the response, the response is always the Alternative: H1: the factor has effect on the response same, regardless the change of the factor, Ho : τ1 = τ 2 =L = τ a = 0 (|τi| is small, no effect) μ1 = μ 2 = L = μ a = μ H1 : τ i = o for at lease one i / (|τi| is larger, has effect) (3.6) ( μ1 − μ ) = ( μ 2 − μ ) = L = ( μ a − μ ) = 0 τ 1 = τ 2 = Lτ a = 0 The larger the value of any |τi|, the more effective the factor will be. Here, the magnitude of τi, rather than its sign, makes sense. If a factor has effect, the value of the response will be different, along with the change of the factor. The more the response differs, the Type I error: the null hypothesis Ho is rejected when it is actually greater the effect of the factor is. true (The factor is concluded effective, when it is not) τ i = o for at least one i / Type II error: the null hypothesis is accepted when it is actually false (The factor is concluded ineffective, when it is effective). 3-13 3-14 Machining Example (2) y Table 3.6 Machine Process capability Cp τ1 = τ2 = 0 USA 1.6, 1.5, 1.4 μ = μ1 = μ2 =1.5 Japan 1.4, 1.5, 1.6 Minor variation is caused by errors, rather than the machines μ1 ≈ 1.5, μ 2 ≈1.5 μ ≈ 1.5 τ 1 = μ1 − μ = 15 − 15 = 0 . . τ 2 = μ2 − μ = 15 − 15 = 0 . . Conclusion: in this example, the factor (machine) has no effect on the process capability. Because, when using different Figure 3.5 machines, the response is always the same, on average. 3-15 3-16
  • 9. Analysis of Variance (ANOVA) (1) Calculate sum of squares (SS) Major Steps Three types of differences: [1] Difference between an individual observation and the grand (1) Calculate sum of squares (SS). average: yij - y.. (2) Determine degrees of freedom (D.O.F.). [2] Difference between a level average and the grand average: (3) Calculate mean square (MS). yi. - y.. = τ i (4) Calculate the test ratio F0. [3] Difference between an individual observation and the corresponding level average: (5) Make conclusion. yij - yi. = ε ij Note: yij - y.. = ( yi. - y.. ) + ( yij - yi. ) 3-17 3-18 a n a n a n Sum of squares is the sum of squares of the above three differences. ∑∑ ( y i =1 j =1 ij - y.. ) ∑∑ ( yi. - y.. ) i =1 j =1 ∑∑ ( y i =1 j =1 ij - yi. ) SST The total sum of squares, which is a measure of total a n a n a n variability in the data. ∑∑ ( y - y.. ) ∑∑ ( yi. - y.. ) ∑∑ ( y - yi . ) 2 2 2 ij ij i =1 j =1 i =1 j =1 i =1 j =1 a n SS T = ∑ ∑ ( yij - y.. ) 2 (3.7) i =1 j =1 SST = SSF + SSE SSF the sum of squares due to the factor, that is the sum of squares of differences between factor level averages and the grand average (difference between levels). a n a a SS F = ∑∑ ( yi. − y.. ) 2 = n∑ ( yi. − y.. ) 2 ≈ n∑ τ i2 (3.8) i =1 j =1 i =1 i =1 3-19 3-20
  • 10. SSE the sum of squares due to error, that is the sum of squares Alternative Formulae of Calculating the Sum of Squares of differences between observations and their level averages (difference within levels). a n y . .2 a n a n SS T = ∑ ∑ yij 2 - i =1 j =1 an (3.11) SS E = ∑∑ ( yij − yi. ) 2 ≈∑∑ ε ij 2 (3.9) i =1 j =1 i =1 j =1 a yi. 2 y . .2 SS F = ∑ n an i =1 - (3.12) SS T = SS F + SS E (3.10) SS E = SST - SS F (3.13) Formulae (3.7) to (3.9) are used to describe the underlying ideas. Formulae (3.11) to (3.13) are used for actually computation. 3-21 3-22 Machining Example (1) If SSF is large, it is due to differences among the means at the different factor levels. See equation (3.8), a large SSF means that τ2i Table 3.7 (or |τi|) is great. It is an indication that the factor has significant effect Machine Process capability Cp on the response. USA 2.1, 2.0, 1.9 Usually, SSF is standardized by taking SSF / SSE . By comparing SSF Japan 0.9, 1.0, 1.1 to SSE, we can see how much variability is due to changing factor levels and how much is due to error. a = 2, n = 3 y1⋅ = 6 y2⋅ = 3 y⋅⋅ = 9 Machining Examples y1⋅ = 2 y2⋅ = 1 y⋅⋅ = 15 . Table 3.8 Effective ? SSF/SSE SST = (2.12 + 2.0 2 + 1.9 2 + 0.9 2 + 1.0 2 + 1.12 ) − 9 2 / 6 = 1.54 Example 1 Yes 37.5 SS F = (6 2 + 32 ) / 3 − 9 2 / 6 = 1.50 Example 2 No 0 SS E = 1.54 − 1.50 = 0.04 3-23 3-24
  • 11. (2) Determine degrees of freedom (D.O.F.). Machining Example (1) Table 3.9 total number of observations N: an Machine Process capability Cp degrees of freedom of SST: N - 1 = an - 1 (3.14) USA 2.1, 2.0, 1.9 degrees of freedom of SSF: a-1 (3.15) Japan 0.9, 1.0, 1.1 degrees of freedom of SSE: R (n - 1) = a (n – 1) (3.16) a = 2, n = 3 where, (n - 1) is the degrees of freedom of errors in a run. total number of observations N: 2×3=6 degrees of freedom of SST: N–1 =6–1=5 Equation of DOF degrees of freedom of SSF: a–1 =2–1=1 an - 1 = (a - 1) + a (n - 1) (3.17) degrees of freedom of SSE: a (n-1) = 2 × (3 - 1) = 4 DOF of SST = DOF of SSF + DOF of SSE Check equation (3.17): 5 = 1 + 4 3-25 3-26 (3) Calculate mean square (MS). (4) Calculate the test ratio F0. The ratio between a sum of squares and its DOF. MS F Fo = (3.20) MS E SS F MS F = (3.18) a − 1 If H0 is true, F0 follows a theoretical F distribution, which is completely determined by the two parameters ν1 and ν2. SS E MS E = (3.19) ν1 = a - 1 the D.O.F. of MSF (numerator) a ( n - 1) ν2 = a(n - 1) the D.O.F. of MSE (denominator) There is no need to calculate MS for the SST 3-27 3-28
  • 12. f(x) (5) Make conclusion. ν1 = 1 ν2 = 4 (1) Decide the curve from SS F ν1 and ν2. MS F a (n − 1) SS F SS (2) Decide Fα,ν1,ν2 from α Fo = = a −1 = ⋅ = Q⋅ F MS E SS E a − 1 SS E SS E a (n − 1) a (n − 1) α Where, Q = is a constant. a −1 SS F Fα ,ν1 ,ν 2 F0 is equivalent to (the standardized SSF) except for a constant Q. SS E Figure 3.6 SS The density function curve of an example of the F distribution However, while F0 follows the theoretical F distribution, F does SS E not. 3-29 3-30 SS F Table 3.10 Control Limit Fα ,ν 1 ,ν 2 found from the F Table If F0 is large ----> is large ---> SSF is large ----> τ2i is large ---> SS E |τi| is large ---> the factor has significant effect on the response. α = 0.05 Control limit: Fα ,ν 1 ,ν 2 ν1 1 2 3 4 5 . ∞ ν2 ν1 = a - 1 and ν2 = a(n - 1) are the two parameters of an F 1 161.4 199.5 distribution 2 18.51 19.00 α is the probability of Type I error, which means concluding that the factor has significant effect on the response when it in fact has 3 10.13 9.55 Examples: no effect. Usually, set α = 1% or 5% 4 7.71 6.94 F0.05,2,1 = 199.5 If Fo < Fα ,ν 1 ,ν 2 , conclude that the factor has no effect. . F0.05,1,4 = 7.71 ∞ If Fo > Fα ,ν 1 ,ν 2 , conclude that the factor has significant effect. 3-31 3-32
  • 13. Table 3.11 Table 3.12 3-33 3-34 Table 3.13 Machining Example (1) (a = 2, n = 3, N = 6), from Table 3.7 Analysis of Variance for a Single-Factor Experiment ___________________________________________________ Table 3.14 ___________________________________________________ Source of Sum of Degrees of Mean Fo Source of Sum of Degrees of Mean Fo Variation Squares Freedom Square Variation Squares Freedom Square ___________________________________________________ ___________________________________________________ Machine 1.50 1 (= ν1) 1.50 150 Factor SSF a-1 MSF MS F Error 0.04 4 (= ν2) 0.01 MS E Total 1.54 5 ___________________________________________________ Error SSE a(n-1) MSE Total SST an-1 ___________________________________________________ 3-35 3-36
  • 14. Specify α = 0.05, Control limit Fα ,ν 1 ,ν 2 = F0.05,1, 4 = 7.71 Machining Example (2) (a = 2, n = 3, N = 6), Table 3.15 Since F0 > Fα ,ν 1 ,ν 2 , machines have significant effect on the process Source of Sum of Degrees of Mean Fo capability. Variation Squares Freedom Square Sum of squares: 1.50 + 0.04 = 1.54 Machines 0 1 (= ν1) 0 0 Degrees of freedom 1 + 4 = 5 Error 0.04 4 (= ν2) 0.01 Total 0.04 5 ___________________________________________________ Specify α = 0.05, Control limit Fα ,ν 1 ,ν 2 = F0.05,1, 4 = 7.71 Since F0 < Fα ,ν1 ,ν2 , machines don’t have significant effect on the process capability. 3-37 3-38 Table 3.16 Example: Tensile strength of paper 4 6 y.. 2 ___________________________________________________________________________________________________________ SS T = ∑ ∑ yij 2 - Hardwood Observations i =1 j =1 an _____________________________________________________________________________________________ Concen. (%) 1 2 3 4 5 6 Total Aver. (383) 2 _________________________________________________________________________________________________________________________________ = (7) 2 + (8) 2 + L + (20) 2 - = 512.96 24 5 7 8 15 11 9 10 60 10.00 4 yi.2 y..2 10 12 17 13 18 19 15 94 15.67 SS F = ∑ - i =1 n an 15 14 18 19 17 16 18 102 17.00 (60) 2 + (94) 2 + (102) 2 + (127) 2 (383) 2 = - = 382.79 20 19 25 22 23 18 20 127 21.17 6 24 __________________________________________________ SS E = SS T - SS F 383 15.96 = 512.96 - 382.79 = 130.17 a = 4, n = 6, N = an = 24. 3-39 3-40
  • 15. Table 3.17 ANOVA for tensile strength of paper SUMMARY: ANOVA for Single Factor Experiment ___________________________________________________________________________________________________________________________________ Source of Sum of Degrees of Mean Fo Variation Squares Freedom Square 1. Decide the factor to be investigated and the following two __________________________________________________________________________________________________________________________________ parameters: Hardwood 382.79 3 (= ν1) 127.60 19.61 1) number of levels a; concentration 2) replicate n. Error 130.17 20 (= ν2) 6.51 2. Carry out the experiments, obtain N (= an) observations yij by a Total 512.96 23 random manner. ______________________________________________________________________________________________________________________________________ Specify α = 0.01, Control limit Fα ,ν 1 ,ν 2 = F0.01,3, 20 = 4.94 3. Check if the residuals (estimates of errors) satisfy the requirements. Since F0 > Fα ,ν 1 ,ν 2 , hardwood concentration has significant effect on the tensile strength of paper. 4. Calculate sum of squares SST, SSF and SSE (3.11, 3.12, 3.13). 3-41 3-42 5. Calculate the degrees of freedom for each of SST, SSF, SSE (3.14, 3.3 Test on Individual Level Means 3.15, 3.16). ANOVA results in a single parameter F0, which can tell whether a 6. Calculate the mean squares MSF and MSE (3.18, 3.19). factor has effect on the response, or whether the mean response values will be different at different levels of the factor from an 7. Calculate the ratio F0 between MSF and MSE (3.20). overall viewpoint. 8. Specify a type I error α, and find the control value Fα ,ν 1 ,ν 2 from However, ANOVA cannot decide the direction of influence of the the F distribution Tables (ν1 = a - 1, ν2 = a (n – 1) ). factors, nor identify which factor level mean is different from another level mean. 9. Conclude if the factor has significant effect on the response. 3-43 3-44
  • 16. The plots indicate that changing the hardwood concentration has a strong effect on the paper strength; specifically, higher hardwood concentrations produce higher observed paper strength. Box plots show the variability of the observations within a factor level and the variability between factor levels. However, a less subjective approach is required to test the individual level means. Figure 3.7 Box plots of paper strength example 3-45 3-46 Multiple Comparison Procedures One of the major problems with making comparisons among level means is that unrestricted use of these comparisons can lead to an It is the techniques for making comparisons between two or more excessively high probability of a Type I error. level means subsequent to an analysis of variance. When we run an analysis of variance and obtain a significant F0 For example, if we have 10 levels in which the complete null value, we have shown simply that the overall null hypothesis is false. hypothesis is true (with α = .05) We do not know which of a number of possible alternative hypotheses is true. Ho: µ1 = µ2 = µ3 = … = µ10 H1 : µ1 ≠ µ2 ≠ µ3 ≠ µ4 ≠ µ5 A series t tests between all pairs of level means will lead to making at Or H1: µ1 ≠ µ2 = µ3 = µ4 = µ5 least one Type I error 57.8% of the time. In other words, the experimenter who thinks he is working at the (α = .05) level of Multiple comparison techniques allow us to investigate hypotheses significance is actually working at (α = 0.578). that involve means of individual levels. For example, we might be interested in whether µ1 is different from µ2, or whether µ2 is different from µ3. 3-47 3-48
  • 17. The probability of making at least one Type I error increases as we Least Significant Difference (LSD) Test is a useful tool to test the increase the number of independent t tests we make between pairs of differences between the individual level means. level means. While it is nice to find significant differences, it is not nice to find ones that are not really there. We need to find some way The basic requirement for a LSD test is that the F0 for the overall to make the comparisons we need but keep the probability of analysis of variance (ANOVA) must be significant. If the F0 was not incorrect rejections of Ho under control. significant, no comparisons between level means are allowed. You simply declare that there are no differences among the level means and stop right there. On the other hand, if the overall F0 is significant, you can proceed to make any or all pairwise comparisons between individual level means by the use of a t test. 3-49 3-50 The t distributions To test between the ith level mean (µi) and jth level mean (µj), Ho: µi = µj H1: µi ≠ µj (3.21) we calculate yi⋅ − y j⋅ t= (3.22) 2 MS E / n The value of MSE has already been obtained during the overall ANOVA. The DOF of t is equal to the DOF of MSE, i.e., a(n-1) or t(30) is a t distribution with a DOF of 30. (N-a). t(∞) means a standard normal distribution with (µ = 0 and σ = 1). Then, using a two-tailed t test for a specified level of type I error α (e.g., α = 0.05), we can decide a pair of two-sided control limits ±tα(dof) from the t table. 3-51 3-52
  • 18. If the t value determined by Eq (3.22) falls within the limits ±tα(dof), we cannot reject the null hypothesis in (3.21). We will therefore conclude that there is no difference between µi and µj. On the other hand, if the t value falls beyond the limits ±tα(dof), we will reject the null hypothesis and conclude that there is a difference between µi and µj. 3-53 3-54 The Example of Tensile Strength of Paper Problem: if hardwood concentrations 5% and 20% produce the Table 3.18 same paper strength ? ___________________________________________________________________________________________________________ Hardwood Observations Ho : μ 1 = μ 4 _____________________________________________________________________________________________ Hypothesis: Concen. (%) 1 2 3 4 5 6 Total Aver. H1 : μ1 = μ 4 / _________________________________________________________________________________________________________________________________ 5 7 8 15 11 9 10 60 10.00 Five Steps: 10 12 17 13 18 19 15 94 15.67 (1) Use Eq (3.22) 15 14 18 19 17 16 18 102 17.00 20 19 25 22 23 18 20 127 21.17 yi ⋅ − y j ⋅ 10.00 − 21.17 t= = = −7.58 __________________________________________________ 2 MS E / n 2 × 6.51 / 6 a = 4, n = 6 (Note: MSE = 6.51 was calculated during ANOVA, which is y1⋅ = 10.00, y2⋅ = 15.67, y3⋅ = 17.00, y4⋅ = 21.17 always carried out before the test on individual level means) 3-55 3-56
  • 19. (2) Determine the DOF, DOF = a (n – 1) = 4 × (6 – 1) = 20 General Steps to Test the Individual Level Means (3) Decide a type I error level. We use α = 0.01 as in the ANOVA for (1) Calculate t using Eq (3.22) and find the MSE value obtained this example. during ANOVA. (4) Find the limits from the t table: ±tα(dof) = ±t0.01(20) ≈ 3.00 (2) Determine the DOF, DOF = a (n – 1). (5) Make conclusion: since the calculated t value (= -7.58) falls (3) Specify a type I error α. beyond the limits ±3.00, we will reject the null hypothesis. (4) Find the control limits from the t table: ±tα(dof) based on α and Conclusion: There is a significant difference in paper strength DOF. between using hardwood concentrations 5% and 20%. Moreover, since t < 0, µ1 < µ4. (5) Make conclusion by comparing the calculated t value against the control limits. 3-57 3-58 The Bonferroni Procedure To put this in a way that is slightly more useful to us, if you want the overall family-wise error rate to be no more than (α = 0.05), and you In a Bonferroni procedure the family-wise error rate (Type I error) is want to run c tests, then each run of them should be at (α' = 0.05/c). divided by the number of comparisons. To run these tests, you do exactly what you did in Fisher's LSD test, though you omit any requirement about the significance of overall F. The basic idea behind this procedure is that if you run several tests (say c tests), each at a significance level (type I error) represented by α', the probability of at least one Type I error can never exceed cα'. Thus, for example, if you ran 5 tests, each at α' = 0.05, the family- wise error rate would be at most 5(0.05) = 0.25. But that is a too high Type I error rate to make anyone happy. But suppose that you ran each of those 5 tests at the α' = 0.01. Then the maximum family-wise error rate would be 5(0.01) = 0.05, which is certainly acceptable. 3-59 3-60
  • 20. a ni 2 y.. SST = ∑∑ yij − 3.5 Unbalanced Experiments 2 DOF = N -1 i =1 j =1 N In an unbalanced experiment, the number of observations taken under each run is different. a yi2. y.. 2 SS F = ∑ − DOF = a -1 ni is the number of observations taken under the ith run (for i =1 ni N single factor experiment, it is simply the ith level). a a SS E = SST − SS F DOF = ∑ (ni − 1) N = ∑ ni = n1 + n2 + L + na i =1 i =1 is the total number of observations. Note, in unbalanced experiments, N ≠ an. 3-61 3-62 Machining Example SST = (2.12 + 2.0 2 + 1.9 2 ) + (0.9 2 + 1.0 2 ) - 7.9 2 / 5 = 1.348 DOET = N – 1 = 5 – 1 = 4 Machine Process capability Cp USA 2.1, 2.0, 1.9 SSF = (6.02/3 + 1.92/2) – 7.92/5 = 1.323 Japan 0.9, 1.0, DOEF = a – 1 = 2 – 1 = 1 n1 = 3, n2 = 2, N = 3 + 2 = 5 SSE = 1.348 – 1.323 = 0.025 y1. = 6.0, y2. = 1.9, y.. = 7.9 DOEE = (n1 – 1) + (n2 – 1) = (3 – 1) + (2 – 1) = 3 3-63 3-64
  • 21. ANOVA for Machining Example (a = 2, n1 = 3, n2 = 2, N = 5) 3.7 Guidelines for Designing Experiments (1) Recognition and statement of the problem. Source of Sum of Degrees of Mean Fo Fully develop all ideas about the problem and the specific Variation Squares Freedom Square objectives of the experiment; Solicit input from all concerned parties -- engineering, quality, Machine 1.323 1 (= ν1) 1.323 158.8 marketing, customer, management and operators. Error 0.025 3 (= ν2) 0.0083 (2) Choice of factors and levels. Total 1.348 4 Choose the factors, their ranges and levels. Process knowledge is ___________________________________________________ required. Specify α = 0.05, Control limit: Fα ,ν 1 ,ν 2 = F0.05,1,3 = 10.13 Investigate all factors that may be of importance and avoid being overly influenced by past experience. Since F0 > Fα ,ν 1 ,ν 2 , machines have significant effect on the process Keep the number of factor levels low (Most often two levels are capability. used.) in the early stage 3-65 3-66 (3) Selection of the response (6) Data analysis Most often, the average or standard deviation (or both) of the Analyze the data so that results and conclusions are objective rather measured characteristic will be the response variable. than judgmental. Use software packages and simple graphical methods (4) Choice of experimental design. Carry out residual analysis. Decide the factorial fraction experiment, select the replicate (sample size) and run order for the experimental trials. (7) Conclusions and recommendations. Decide whether or not blocking or other randomization methods are involved. (8) Experimentation is an important part of the learning process. New hypotheses are formulated based on the investigation of the (5) Performing the experiment. tentative hypotheses. Don’t design a single, large comprehensive experiment at the start of a study. As a rule of thumb, the first Ensure that everything is being done accordingly to plan. experiment should spend no more than 25% of the total budget of the study. 3-67 3-68
  • 22. 4 RANDOMIZED BLOCKS Response: coded roughness of a machined surface. 4.1 Randomized Block Design Main factor (factor to be studied in the experiment) Nuisance factor probably has an effect on the response, but we are Cutter of a machine tool not interested in that effect. Known nuisance factor (factor not to be studied in the experiment, For unknown and uncontrollable nuisance factors, use Complete but we are well aware of the existence of its effect on the response) Randomization technique to get rid of them. Different machine tools of the same type For known and controllable nuisance factors, use Randomized Blocking technique instead. Unknown nuisance factor (unknown and uncontrollable factors, which may have some effect on the response) The variability of the nuisance factor will contribute to the variability observed in the experimental data. As a result, the experimental error Temperature, humidity will reflect both the random error of the unknown nuisance factors and variability of the known nuisance factor. 4-1 4-2 Table 4.1 MS F MS F DOEE F0 = = = MS F (4.4) Machine tool (Block) MS E SS E SS E Cutter 1 2 3 4 DOEE 1 -2 -1 1 5 2 -1 -2 3 4 So, whether F0 will be increased by the addition of the known 3 -3 -1 0 2 nuisance factors depends on which of SSE or DOEE will increase 4 2 1 5 7 more. If SSE increases more by including the known nuisance factors, F0 SST = SS F + SS known + SSunknown (4.1) becomes less sensitive. Therefore, Randomized Blocking technique is preferred. If the effects of the known nuisance factors are included in the error If DOEE increases more, F0 becomes more sensitive. Therefore, Complete Randomization technique is preferred. SS E = SS known + SSunknown (4.2) Note, the increase of DOEE also reduces the control limit Fα ,ν 1 ,ν 2 (ν2 DOEE = DOEknown + DOEunknown (4.3) = DOEE), and therefore, makes F0 relatively more sensitive. 4-3 4-4
  • 23. Randomized Block Design Block 1 Block 2 Block b It means blocking all the runs according to the different levels of a known nuisance factor. Within a block, the order in which the y11 y12 y1b runs of the main factors are tested is randomly determined. y21 y22 y2b The blocks form a more homogeneous experimental unit in terms of the known nuisance factor. ............. y31 y32 y3b Effectively, this design strategy improves the accuracy of the . . . comparisons of the responses by eliminating the variability of the . . . known nuisance factor. . . . ya1 ya2 yab Examples of blocks Equipment, machinery, batches of raw material, people, time slot Figure 4.1 4-5 4-6 The statistical model Hypotheses of interest are ⎧ i = 1,2,L, a H0: μ1 = μ2 = = … μa yij = μ + τ i + β j + ε ij ⎨ (4.5) ⎩ j = 1,2,L, b H1: at least one μi ≠ μj (4.6) where, μ is an overall mean, τi is the effect of the ith run, βj is the Or effect of the jth block, and εij is the usual NID(0, σ2) random error H0: τ1 = τ2 = … = τa = 0 term. H1: at least one τi ≠ 0 (4.7) βj will not be studied, because it is the effect of the known nuisance factor. 4-7 4-8
  • 24. yi. the total of all observations taken under run i (total of row) yi. the average of the observations taken under the ith run y.j the total of all observations in block j (total of column) y. j the average of the observations in block j y.. the grand total of all observations y.. the grand average of all observations. N = ab the total number of observations. b yi. = yi . / b y. j = y . j / a y.. = y.. / N (4.9) yi. = ∑ yij i = 1,2,L, a j =1 a y. j = ∑ yij j = 1,2,L, b (4.8) i =1 a b y.. = ∑∑ yij i =1 j =1 4-9 4-10 Sum of square Degrees of Freedom SST = SS F + SS Block + SS E (4.10) SST N - 1 = ab - 1 a b y..2 SSF a-1 SST = ∑∑ yij 2 - (4.11) i =1 j =1 N SSBlocks b-1 a 2 1 y.. SS F = ∑ yi.2 - N b i=1 (4.12) SSE (ab − 1) − (a − 1) − (b − 1) = ab − a − b + 1 (4.15) 1 b 2 y..2 SS Block = ∑ y. j - (4.13) a j =1 N SS E = SST - SS F − SS Block (4.14) 4-11 4-12
  • 25. Test Statistic Table 4.2 ANOVA Table _______________________________________________________ MS F Source of Sum of Degrees of Mean F0 F0 = MS E Variation Squares Freedom Square SS F _______________________________________________________ MS F = (4.16) a −1 SS F MS F Factor SS F a -1 Fo = SS E a −1 MS E MS E = ab − a − b + 1 SS Blocks Blocks SSBlocks b–1 We would reject Ho in (4.7), if Fo > Fα,(a-1)(ab-a-b+1). b −1 SS E Error SS E ab – a – b + 1 ab − a − b + 1 Total SST N -1 _______________________________________________________ 4-13 4-14 We may also examine the ratio of MSBlocks (= SSBlocks/(b-1)) to MSE. If Example 4.1 this ratio is large, it implies that the blocking factor has a large effect Table 4.3 and that the noise reduction obtained by blocking was probably Machine tool (Block) Cutter 1 2 3 4 yi. helpful in improving the effectiveness of the experiment. Otherwise, 1 -2 -1 1 5 3 blocking may actually be unnecessary, i.e, the known nuisance factor 2 -1 -2 3 4 4 should be handled as random error. 3 -3 -1 0 2 -2 4 2 1 5 7 16 y.j -4 -3 9 18 y.. = 20 4-15 4-16
  • 26. The sums of squares are obtained as follows: Table 4.4 ANOVA Table _______________________________________________________ a b y..2 20 2 Source of Sum of Degrees of Mean F0 SST = ∑∑ yij 2 - = 154 − = 129 i =1 j =1 N 16 Variation Squares Freedom Square _______________________________________________________ = (3 + 4 + (−2) + 15 ) − 1 a 2 y..2 1 2 20 2 SS F = ∑ yi. - 2 2 2 = 38.5 Cutter 38.50 3 12.83 14.44 b i=1 N 4 16 Machine 82.50 3 27.50 ∑ y.2j - N = 4 ((−4) 2 + (−3) 2 + 92 + 182 ) − 16 = 82.5 b 2 2 (Block) 1 y.. 1 20 SS Block = a j =1 Error 8.00 9 0.89 SS E = SST - SS F − SS Block = 129 − 38.5 − 82.5 = 8 Total 129.00 15 _______________________________________________________ Specify α = 0.05, F0.05,3,9 = 3.86. Since 14.44 > 3.86, we conclude that the type of cutter affects the mean hardness reading. 4-17 4-18 If the randomized block design is not used, F0.05,3,12 = 3.49, the hypothesis of equal mean values from the four cutters cannot be rejected. Table 4.5 ANOVA Table Source of Sum of Degrees of Mean F0 Variation Squares Freedom Square Cutter 38.50 3 12.83 1.70 Error 90.50 12 7.54 Total 129.00 15 Here, the error is actually the sum of error and block in Table 4.4. 4-19
  • 27. 5 FACTORIAL EXPERIMENTS Case one Two factors: A: Machine, B: Feed rate 5.1 Introduction Response: Process capability Factorial experiments study the effects of several factors on the response. Table 5.1 Process capability Cp In factorial experiments, all possible runs (combinations of the levels Feed rate (mm/sec.) of the factors) are investigated. Machine 0.2 0.4 0.6 For example, for two factors A: temperature (with a levels) and B: USA 2.1, 2.2 1.7, 1.6 1.2, 1.3 humidity (with b levels), the factorial experiment has ab possible Japan 0.9, 1.0 0.7, 0.6 0.4, 0.3 runs. a = 2, b = 3, n = 2, R =ab = 6 If for each run, n observations are taken, then, the total number N of N = Rn = abn = 2 × 3 × 2 = 12 the observations is equal to abn. 5-1 5-2 Case two 5.2 ANOVA for Two-Factor Factorial Experiments Three factors: A, B and C. Mathematical model a = 2, b = 3, c = 4, n = 2, R = abc = 24 Yijk = μ + τ i + β j + (τβ )ij + ε ijk (5.1) N = Rn = abcn = 2 × 3 × 4 × 2 = 48 μ the overall mean τi the effect incurred at the ith level of factor A Number of runs is equal to the product of the levels of all factors. βj the effect incurred at the jth level of factor B (τβ)ij the effect incurred at the interaction between the ith level of A and the jth level of B εijk the random error (normally and independently distributed, zero mean and constant variance) 5-3 5-4
  • 28. Interaction indicates the influence of one factor on the effect of Table 5.2 Data Arrangement for a Two-Factor Factorial Design another factor, and vice versa. Factor B 1 2 … b Total Average Factors A: temperature 1 y111, y112, y121, y122, … y1b1, y1b2, y1.. y1.. B: pressure Factor A …, y11n …, y12n …, y1bn 2 y211, y212, y221, y222, … Y2b1, y2b2, y2.. y 2.. Response Process yield …, y21n …, y22n …, y2bn : … … … … … … If: (1) Increasing temperature alone cannot increase the yield; and a ya11, ya12, ya21, ya22, … yab1, yab2, ya.. y a.. (2) Increasing pressure alone cannot increase the yield; and …, ya1n …, ya2n …, yabn (3) Increasing temperature and pressure at the same time does Total y.1. y.2. … y.b. y… increase the yield Average y .1. y .2. … y .b. y ... Then: There is interaction between temperature (factor A) and yijk is the kth observation taken at the ith level of factor A and the jth pressure (factor B) level of factor B. The total number of observations: N = abn 5-5 5-6 Total and average at the ith level of factor A: Hypotheses b n y yi .. = Σ Σ yijk y i .. = i .. i = 1,2,...,a Factor A: j =1 k =1 bn (5.2) H o : τ 1 = τ 2 = ... = τ a = 0 Factor A has no effect on y Total and average at the jth level of factor B: a n y. j . H 1 : at least one τ i ≠ 0 Factor A has effect on y y. j . = Σ Σ yijk y. j. = j = 1,2,..., b i =1 k =1 an (5.3) Factor B: Total and average when A at the ith level and B at the jth level: H 0 : β1 = β2 =L = βb = 0 Factor B has no effect on y n y H 1 : at least one β j ≠ 0 yij . = Σ yijk y ij . = ij . Factor B has effect on y k =1 n (5.4) Interaction AB: Grand total and average a b n y... H o : (τβ ) ij = 0 for all i, j Interaction AB has no effect y y...= ΣΣΣ yijk y...= (5.5) i =1 j =1 k =1 abn H1 : at least one (τβ ) ij ≠ 0 Interaction AB has effect on y 5-7 5-8