Scribd Logo
Upload

Welcome to Scribd!

  • 50 views

    T 2 F D N: HE Actorial Esig

    Uploaded by

    Saheed A Busura
    The document discusses the 2k factorial design, which involves k factors with two levels each. It is a widely used experimental design for industrial applications. The simplest case is the 22 factorial design with two factors at two levels. Analysis of a 2k factorial design involves estimating the main effects of each factor and interaction effects. Statistical tests like ANOVA are used to evaluate the significance of effects in the model.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    T 2 F D N: HE Actorial Esig

    Uploaded by

    Saheed A Busura
    50 views33 pages
    The document discusses the 2k factorial design, which involves k factors with two levels each. It is a widely used experimental design for industrial applications. The simplest case is the 22 factorial design with two factors at two levels. Analysis of a 2k factorial design involves estimating the main effects of each factor and interaction effects. Statistical tests like ANOVA are used to evaluate the significance of effects in the model.

    Original Title

    part06_2K_Factorial_Desing.pdf

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document discusses the 2k factorial design, which involves k factors with two levels each. It is a widely used experimental design for industrial applications. The simplest case is the 22 factorial design with two factors at two levels. Analysis of a 2k factorial design involves estimating the main effects of each factor and interaction effects. Statistical tests like ANOVA are used to evaluate the significance of effects in the model.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    50 views33 pages

    T 2 F D N: HE Actorial Esig

    Uploaded by

    Saheed A Busura
    The document discusses the 2k factorial design, which involves k factors with two levels each. It is a widely used experimental design for industrial applications. The simplest case is the 22 factorial design with two factors at two levels. Analysis of a 2k factorial design involves estimating the main effects of each factor and interaction effects. Statistical tests like ANOVA are used to evaluate the significance of effects in the model.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 33

    THE 2K FACTORIAL DESIGN

     Text reference, Chapter 6


     Special case of the general factorial design; k
    factors, all at two levels
     The two levels are usually called low and high
    (they could be either quantitative or qualitative)
     Very widely used in industrial experimentation
     Form a basic “building block” for other very useful
    experimental designs (DNA)
     Special (short-cut) methods for analysis
     We will make use of Computer package
    THE SIMPLEST CASE: THE 22
    “-” and “+” denote
    the low and high
    levels of a factor,
    respectively
    Low and high are
    arbitrary terms
    Geometrically, the
    four runs form the
    corners of a square
    Factors can be
    quantitative or
    qualitative, although
    their treatment in the
    final model will be
    different
    CHEMICAL PROCESS EXAMPLE

    A = reactant concentration, B = catalyst amount,


    y = recovery
    ANALYSIS PROCEDURE FOR A
    FACTORIAL DESIGN
     Estimate
    factor effects
     Formulate model
     With replication, use full model
     With an unreplicated design, use normal
    probability plots
     Statistical
    testing (ANOVA)
     Refine the model
     Analyze residuals (graphical)
     Interpret results
    ESTIMATION OF FACTOR EFFECTS
    A = y A+ − y A −
    See textbook, For manual
    ab + a b + (1)
    = − calculations
    2n 2n
    = 21n [ab + a − b − (1)] The effect estimates are: A
    = 8.33, B = -5.00, AB = 1.67
    B = yB+ − yB−
    Practical interpretation?
    ab + b a + (1)
    = −
    2n 2n
    = 21n [ab + b − a − (1)]
    ab + (1) a + b
    AB = −
    2n 2n
    = 21n [ab + (1) − a − b]
    ESTIMATION OF FACTOR EFFECTS FORM TENTATIVE MODEL

    Factorial Fit: Recovery versus A1, B1

    Estimated Effects and Coefficients for Recovery (coded units)

    Term Effect Coef SE Coef T P


    Constant 27.500 0.5713 48.14 0.000
    A1 8.333 4.167 0.5713 7.29 0.000
    B1 -5.000 -2.500 0.5713 -4.38 0.002
    A1*B1 1.667 0.833 0.5713 1.46 0.183

    S = 1.97906 PRESS = 70.5


    R-Sq = 90.30% R-Sq(pred) = 78.17% R-Sq(adj) = 86.66%

    Analysis of Variance for Recovery (coded units)

    Source DF Seq SS Adj SS Adj MS F P


    Main Effects 2 283.333 283.333 141.667 36.17 0.000
    A1 1 208.333 208.333 208.333 53.19 0.000
    B1 1 75.000 75.000 75.000 19.15 0.002
    2-Way Interactions 1 8.333 8.333 8.333 2.13 0.183
    A1*B1 1 8.333 8.333 8.333 2.13 0.183
    Residual Error 8 31.333 31.333 3.917
    Pure Error 8 31.333 31.333 3.917
    Total 11 323.000
    STATISTICAL TESTING - ANOVA

    The F-test for the “model” source is testing the significance


    of the overall model; that is, is either A, B, or AB or some combination
    of these effects important?
    Regression Model for the Process
    Estimated Effects and Coefficients for Recovery
    (coded units)

    Term Effect Coef SE Coef T P


    Constant 27.500 0.5713 48.14 0.000
    A1 8.333 4.167 0.5713 7.29 0.000
    B1 -5.000 -2.500 0.5713 -4.38 0.002
    A1*B1 1.667 0.833 0.5713 1.46 0.183
    RESIDUALS AND DIAGNOSTIC
    CHECKING
    THE RESPONSE SURFACE
    THE 23 FACTORIAL DESIGN
    EFFECTS IN THE 23 FACTORIAL DESIGN

    A = y A+ − y A −
    B = yB+ − yB−
    C = yC + − yC −
    etc, etc, ...

    Analysis
    done via
    computer
    AN EXAMPLE OF A 23 FACTORIAL
    DESIGN

    A = gap, B = Flow, C = Power, y = Etch Rate


    Table of – and + Signs for the 23 Factorial Design (pg. 214)
    PROPERTIES OF THE TABLE
     Except for column I, every column has an equal number of +
    and – signs
     The sum of the product of signs in any two columns is zero
     Multiplying any column by I leaves that column unchanged
    (identity element)
     The product of any two columns yields a column in the table:

    A  B = AB
    AB  BC = AB 2C = AC
     Orthogonal design
     Orthogonality is an important property shared by all factorial
    designs
    ESTIMATION OF FACTOR EFFECTS
    Estimated Effects and Coefficients for Etch rate (coded units)

    Term Effect Coef SE Coef T P


    Constant 776.06 11.87 65.41 0.000
    A -101.62 -50.81 11.87 -4.28 0.003
    B 7.38 3.69 11.87 0.31 0.764
    C 306.13 153.06 11.87 12.90 0.000
    A*B -24.88 -12.44 11.87 -1.05 0.325
    A*C -153.63 -76.81 11.87 -6.47 0.000
    B*C -2.13 -1.06 11.87 -0.09 0.931
    A*B*C 5.63 2.81 11.87 0.24 0.819
    ANOVA SUMMARY – FULL MODEL
    Analysis of Variance for Etch rate (coded units)

    Source DF Seq SS Adj SS Adj MS F P


    Main Effects 3 416378 416378 138793 61.62 0.000
    A 1 41311 41311 41311 18.34 0.003
    B 1 218 218 218 0.10 0.764
    C 1 374850 374850 374850 166.41 0.000
    2-Way Interactions 3 96896 96896 32299 14.34 0.001
    A*B 1 2475 2475 2475 1.10 0.325
    A*C 1 94403 94403 94403 41.91 0.000
    B*C 1 18 18 18 0.01 0.931
    3-Way Interactions 1 127 127 127 0.06 0.819
    A*B*C 1 127 127 127 0.06 0.819
    Residual Error 8 18020 18020 2253
    Pure Error 8 18021 18021 2253
    Total 15 531421
    MODEL COEFFICIENTS – FULL MODEL
    Factorial Fit: Etch rate versus A, B, C

    Estimated Effects and Coefficients for Etch rate (coded units)

    Term Effect Coef SE Coef T P


    Constant 776.06 11.87 65.41 0.000
    A -101.62 -50.81 11.87 -4.28 0.003
    B 7.38 3.69 11.87 0.31 0.764
    C 306.13 153.06 11.87 12.90 0.000
    A*B -24.88 -12.44 11.87 -1.05 0.325
    A*C -153.63 -76.81 11.87 -6.47 0.000
    B*C -2.13 -1.06 11.87 -0.09 0.931
    A*B*C 5.63 2.81 11.87 0.24 0.819

    S = 47.4612 PRESS = 72082


    R-Sq = 96.61% R-Sq(pred) = 86.44% R-Sq(adj) = 93.64%
    REFINE MODEL – REMOVE NONSIGNIFICANT
    FACTORS
    Factorial Fit: Etch rate versus A, C
    Estimated Effects and Coefficients for Etch rate (coded units)
    Term Effect Coef SE Coef T P
    Constant 776.06 10.42 74.46 0.000
    A -101.62 -50.81 10.42 -4.88 0.000
    C 306.12 153.06 10.42 14.69 0.000
    A*C -153.62 -76.81 10.42 -7.37 0.000

    S = 41.6911 PRESS = 37080.4


    R-Sq = 96.08% R-Sq(pred) = 93.02% R-Sq(adj) = 95.09%

    Analysis of Variance for Etch rate (coded units)

    Source DF Seq SS Adj SS Adj MS F P


    Main Effects 2 416161 416161 208080 119.71 0.000
    A 1 41311 41311 41311 23.77 0.000
    C 1 374850 374850 374850 215.66 0.000
    2-Way Interactions 1 94403 94403 94403 54.31 0.000
    A*C 1 94403 94403 94403 54.31 0.000
    Residual Error 12 20858 20858 1738
    Pure Error 12 20858 20858 1738
    Total 15 531421
    MODEL COEFFICIENTS – REDUCED
    MODEL

    Estimated Effects and Coefficients for Etch rate (coded


    units)
    Term Effect Coef SE Coef T P
    Constant 776.06 10.42 74.46 0.000
    A -101.62 -50.81 10.42 -4.88 0.000
    C 306.12 153.06 10.42 14.69 0.000
    A*C -153.62 -76.81 10.42 -7.37 0.000
    MODEL SUMMARY STATISTICS FOR REDUCED MODEL

     R2 and adjusted R2
    SS Model 5.106 10 5
    R = 2
    = = 0.9608
    SST 5.314 105

    SS E / df E 20857.75 /12
    R 2
    = 1− = 1− = 0.9509
    5.314 10 /15
    Adj 5
    SST / dfT
     R2 for prediction (based on PRESS)

    PRESS 37080.44
    R 2
    = 1− = 1− = 0.9302
    5.314 10
    Pred 5
    SST
    MODEL SUMMARY STATISTICS
     Standard error of model coefficients (full
    model)
    MS E2 2252.56
    se( ˆ ) = V ( ˆ ) =
    k
    = k
    = = 11.87
    n2 n2 2(8)
     Confidence interval on model coefficients

    ˆ − t / 2,df se( ˆ )    ˆ + t / 2,df se( ˆ )


    E E
    THE REGRESSION MODEL
    MODEL INTERPRETATION

    Cube plots are


    often useful visual
    displays of
    experimental
    results
    THE GENERAL 2K FACTORIAL DESIGN
     Section 6-4, pg. 224, Table 6-9, pg. 225
     There will be k main effects, and

    k 
      two-factor interactions
     2
    k 
      three-factor interactions
     3

    1 k − factor interaction
    UNREPLICATED 2K FACTORIAL
    DESIGNS
     These are 2k factorial designs with one
    observation at each corner of the “cube”
     An unreplicated 2k factorial design is also
    sometimes called a “single replicate” of
    the 2k
     These designs are very widely used
     Risks…if there is only one observation at
    each corner, is there a chance of unusual
    response observations spoiling the results?
     Modeling “noise”?

    26
    SPACING OF FACTOR LEVELS IN THE
    UNREPLICATED 2K FACTORIAL DESIGNS

    27
    If the factors are spaced too closely, it increases the chances
    that the noise will overwhelm the signal in the data
    More aggressive spacing is usually best
    UNREPLICATED 2K FACTORIAL
    DESIGNS
     Lack
    of replication causes potential
    problems in statistical testing
     Replication admits an estimate of “pure error” (a
    better phrase is an internal estimate of error)
     With no replication, fitting the full model results
    in zero degrees of freedom for error
     Potential solutions to this problem
     Pooling high-order interactions to estimate error
     Normal probability plotting of effects
    (Daniels, 1959)

    28
    EXAMPLE OF AN UNREPLICATED 2K
    DESIGN

     A 24 factorial was used to investigate the effects of


    four factors on the filtration rate of a resin
     The factors are A = temperature, B = pressure, C =
    mole ratio, D= stirring rate
     Experiment was performed in a pilot plant

    29
    UN-REPLICATED DESIGN (EXAMPLE 6.2)
    FILTRATION RATE
    HOW TO ANALYZE?
     Can not put a full model (no DF for error).
     Need to decide which elements of the model to
    include.
     Use the Normal plot, or Half normal Plot or the
    Pareto plot to decide.
     See the example below:
    NORMAL/HALF NORMAL PLOT
    ANOVA FOR EXAMPLE 6.2
    Source DF Seq SS Adj SS Adj MS F P
    Main Effects 3 3116.19 3116.19 1038.73 46.29 0.000
    Temperature 1 1870.56 1870.56 1870.56 83.37 0.000
    Concentration 1 390.06 390.06 390.06 17.38 0.003
    Stirring Rate 1 855.56 855.56 855.56 38.13 0.000
    2-Way Interactions 3 2424.69 2424.69 808.23 36.02 0.000
    Temperature*Concentration 1 1314.06 1314.06 1314.06 58.57 0.000
    Temperature*Stirring Rate 1 1105.56 1105.56 1105.56 49.27 0.000
    Concentration*Stirring Rate 1 5.06 5.06 5.06 0.23 0.647
    3-Way Interactions 1 10.56 10.56 10.56 0.47 0.512
    Temperature*Concentration*Stirring Rate 1 10.56 10.56 10.56 0.47 0.512
    Residual Error 8 179.50 179.50 22.44
    Pure Error 8 179.50 179.50 22.44
    Total 15 5730.94

    You might also like