Majd Shhadi, PhD

First Semester 2018/2019 1

 Factorial Experiments
2

 When several factors are of interest in an experiment, a

factorial experiment should be used.

 By factorial experiment, we mean that in each complete trial

or replicate of the experiment, all possible combinations of
the levels of the factors are investigated.

 The total number of experiments for studying k factors at 2-

levels is 2k.

 One of the assumptions we make for factors at 2-levels is that

the response is approximately linear over the range of the
factor settings chosen. Majd Shhadi, PhD
 Factorial Experiments
3

 The first design in the 2k series is one with only two factors, say,
A and B, each factor to be studied at 2-levels (low (-) and high
levels (+)). This is called a 22 full factorial design.

 In the 22 design are the main effects A and B and the two-
factor interaction AB. Let the symbols (1), a, b and ab also
represent the totals of all n observations taken at these design
points. It is easy to estimate the effects of these factors.

Majd Shhadi, PhD

 Factorial Experiments: 22 design
4

 The treatment for this design is shown in Table 14.12

Table14-12: Signs for Effects in the 22 Design

Factorial Effect
Treatment
Run 1 A B AB
Combination
1 1 1 -1 -1 1
2 a 1 1 -1 -1
3 b 1 -1 1 -1
4 ab 1 1 1 1

Majd Shhadi, PhD

 Factorial Experiments: 22 design
5

 The 22 design can also be presented in a geometric way, using

a square with the four treatment combinations lying at the four
corners.

Majd Shhadi, PhD

 Analytical tools of DOE
6

1. Main effects plot

 A main effect plot is a plot of the mean response values at each

level of a design parameter or process variable.

 One can use this plot to compare the relative strength of the effects
of various factors. The sign and magnitude of a main effect would
tell us the following:

 The sign of a main effect tells us of the direction of the effect,

i.e. if the average response value increases or decreases.

 The magnitude tells us of the strength of the effect.

Majd Shhadi, PhD
 Analytical tools of DOE: Main effects plot
7

 If the effect of a design or process parameter is positive, it implies

that the average response is higher at high level than at low level
of the parameter setting.
 In contrast, if the effect is negative, it means that the average
response at the low level setting of the parameter is more than at
the high level.

 The effect of a process or design parameter (or factor) can be

mathematically calculated using the following simple equation:
𝐸𝑓 = 𝑦ത(+1) − 𝑦ത(−1)
 Where 𝑦ത(+1) average response at high level setting of a factor, and
𝑦ത(−1) average response at low level setting of a factor.
Majd Shhadi, PhD
 Analytical tools of DOE: Main effects plot
8

The following Figure illustrates main effect of temperature on the

tensile strength of a steel specimen. As you can see from the figure,
tensile strength increases when the setting of temperature varies
from low to high (i.e. -1 to 1).

Majd Shhadi, PhD

 Analytical tools of DOE: Interactions plots
9

2. Interactions Plots

 An interactions plot is a powerful graphical tool which plots the

mean response of two factors at all possible combinations of
their settings.

 If the lines are parallel, then it connotes that there is no an

interaction between the factors.

 Non-parallel lines is an indication of the presence of

interaction between the factors.

Majd Shhadi, PhD

 Analytical tools of DOE: Interactions plots
10

2. Interactions Plots

Majd Shhadi, PhD

 Analytical tools of DOE: Cube Plots
11

3. Cube Plots

 Cube plots display the average response values at all

combinations of process or design parameter settings.

 One can easily determine the best and the worst combinations
of factor levels for achieving the desired optimum response.

 A cube plot is useful to determine the path of steepest ascent

or descent for optimization problems.

Majd Shhadi, PhD

 Analytical tools of DOE: Cube Plots
12

3. Cube Plots

Figure 4.2 illustrates an example of a cube plot for a cutting tool life optimization
study with three tool parameters; cutting speed, tool geometry and cutting angle.
The graph indicates that tool life increases when cutting speed is set at low level and
cutting angle and tool geometry are set at high levels.
The worst condition occurs when all factors are set at low levels. Majd Shhadi, PhD
 Analytical tools of DOE:
Pareto Plot Of Factor Effects
13

4. Pareto Plot Of Factor Effects

 The Pareto plot allows one to detect the factor and interaction
effects which are most important to the process or design
optimization study one has to deal with.

 It displays the absolute values of the effects, and draws a

reference line on the chart.

 Any effect that extends past this reference line is potentially

important.

Majd Shhadi, PhD

 Analytical tools of DOE:
Pareto Plot Of Factor Effects
14

4. Pareto Plot Of Factor Effects

The graph shows

that factors B and
C and interaction
AC are most
important

Majd Shhadi, PhD

 Analytical tools of DOE:
Normal Probability Plot of factor effects
15

5. Normal Probability Plot (NPP) of factor effects

 For NPPs, the main and interaction effects of factors or process

(or design) parameters should be plotted against cumulative
probability (per cent).

 Inactive main and interaction effects tend to fall roughly along

a straight line whereas active effects tend to appear as
extreme points falling off each end of the straight line. These
active effects are judged to be statistically significant.

Majd Shhadi, PhD

 Analytical tools of DOE:
Normal Probability Plot of factor effects
16

5. Normal Probability Plot (NPP) of factor effects

Majd Shhadi, PhD

 Analytical tools of DOE:
Normal Probability Plot of residuals
17

6. Normal Probability Plot of residuals

 In order to check the data for normality, it is best to construct a
NPP of the residuals. Normal probability plots are useful for
evaluating the normality of a data set, even when there is a
fairly small number of observations.

 Here a residual means difference in the observed value

(obtained from the experiment) and the predicted value or
fitted value.

 If the residuals fall approximately along a straight line, the

residuals are then normally distributed.
Majd Shhadi, PhD
 Analytical tools of DOE:
Normal Probability Plot of residuals
18

6. Normal Probability Plot of residuals

 In contrast, if the residuals do not fall fairly close to a straight
line, the residuals are then not normally distributed and hence
the data do not come from a normal population.

Majd Shhadi, PhD

 Analytical tools of DOE:
Response Surface Plots And Regression Models
19

7. Response Surface Plots And Regression Models

 Response surface plots such as contour and surface plots are
useful for establishing desirable response values and operating
conditions.
 In a contour plot, the response surface is viewed as a two-
dimensional plane where all points that have the same
response are connected to produce contour lines of constant
responses.
 A surface plot generally displays a three-dimensional view that
may provide a clearer picture of the response.
Majd Shhadi, PhD
 Analytical tools of DOE:
Response Surface Plots And Regression Models
20

7. Response Surface Plots And Regression Models

 Response surface plots such as contour and surface plots are
useful for establishing desirable response values and operating
conditions.
 In a contour plot, the response surface is viewed as a two-
dimensional plane where all points that have the same
response are connected to produce contour lines of constant
responses.
 A surface plot generally displays a three-dimensional view that
may provide a clearer picture of the response.
Majd Shhadi, PhD
 Analytical tools of DOE:
Response Surface Plots And Regression Models
21

7. Response Surface Plots And Regression Models

Contour plot of cutting tool life

Majd Shhadi, PhD
 Analytical tools of DOE:
Response Surface Plots And Regression Models
22

7. Response Surface Plots And Regression Models

Surface plot of cutting tool life

Majd Shhadi, PhD
 Example of a 22 Full Factorial Design
23

 Consider a simple nickel plating process with two plating

process parameters; plating time and plating solution
temperature.

 Each process parameter is studied at 2-levels. The response of

interest to the experimenters was plating thickness. Table 6.1
illustrates the two process parameters and their chosen levels
for the experiment.

Process Parameters Labels Low Level High Level

Plating Time A 4 second 12 second
Plating Solution Temperature B 16 C 32 C
Majd Shhadi, PhD
 Example of a 22 Full Factorial Design
24

 Table 6.2 shows the design layout of the experiment with

response values. Each experimental condition was replicated
five times.

Trail
A B Plating Thickness
number
1 4 16 116.1 116.9 112.6 118.7 114.9
2 4 32 106.7 107.5 105.9 107.1 106.5
3 12 16 116.5 115.5 119.2 114.7 118.3
4 12 32 123.2 125.1 124.5 124.0 124.7

Majd Shhadi, PhD

 Example of a 22 Full Factorial Design
25

The following are the four objectives set by the experimenter:

1. Which main effects or interactions might affect the mean
plating thickness?

2. Which main effects or interactions might influence variability

in plating thickness?

3. What is the best setting of factors to minimize variability in

thickness?

4. How to achieve a target plating thickness of 120 units?

Majd Shhadi, PhD

 Example of a 22 Full Factorial Design
26

6.2.1 Objective 1" Determination of main/interaction effects

which influence mean plating thickness.

 In order to determine the effect of process parameters A and B

and its interaction AB, we need to construct a coded design
matrix with mean plating thickness values as shown in Table
6.3.

 The column AB is obtained by simply multiplying the coded

values in columns 'A' and 'B'. Interaction AB yields a combined
effect of two factors, A and B.

Majd Shhadi, PhD

 Example of a 22 Full Factorial Design
27

6.2.1 Objective 1" Determination of main/interaction effects

which influence mean plating thickness.

Table 6.3 Coded design matrix with mean plating thickness values

Mean
Std Actual Treatment
Combination A B AB Plating
Run Run
Thickness
1 2 1 -1 -1 1 115.84
2 3 a 1 -1 -1 116.84
3 1 b -1 1 -1 106.74
4 4 ab 1 1 1 124.3
Majd Shhadi, PhD
 Example of a 22 Full Factorial Design
28

6.2.1 Objective 1" Determination of main/interaction effects

which influence mean plating thickness.

Cube Plot (data means) for mean plating thickness

106.74 124.30
1

Plating Solution Temperature

115.84 116.84
-1

-1 1
Plating Time

Majd Shhadi, PhD