CH.

5 FACTORIAL DESIGN

INTRODUCTION
Factorial design is an important method to determine the effects of MULTIPLE variables on a response/output. Traditionally, experiments are designed to determine the effect of ONE variable upon ONE response/output.

https://controls.engin.umich.edu/wiki/index.php/Design_of_experiments_via_factorial_designs

Cooling Rate

%C

Tensile Strength

Heating Temp

Temp Diff

Alloying Elements

Physical Prop.

Heat Transfer Rate

Heat Transfer Area

Which variable is more important than others?

Fluid Velocity

 R.A. Fisher showed that there are advantages by combining the study of multiple variables in the same factorial experiment. Factorial design can reduce the number of experiments one has to perform by studying multiple factors simultaneously. Additionally, it can be used to find both main effects (from each independent factor) and interaction effects (when both factors must be used to explain the outcome).
https://controls.engin.umich.edu/wiki/index.php/Design_of_experiments_via_factorial_designs

 Factorial design is a useful method to design experiments in both laboratory and industrial settings. Because factorial design can lead to a large number of trials, which can become expensive and time-consuming, factorial design is best used for a small number of variables with few states (1 to 3).
https://controls.engin.umich.edu/wiki/index.php/Design_of_experiments_via_factorial_designs

A 23 FACTORIAL DESIGN: PILOT PLANT INVESTIGATION

This experiment employed a 23 factorial experimental design with two quantitative factorstemperature T and concentration C and a single qualitative factortype of catalyst K. Each data value recorded is for the response yield y averaged over two duplicate runs.

CALCULATION OF MAIN EFFECTS

CALCULATION OF MAIN EFFECTS

CALCULATION OF MAIN EFFECTS

INTERACTION EFFECTS

INTERACTION EFFECTS

INTERACTION EFFECTS

A L G O R I T M A YAT E S
To calculate
The main effects (e.g. main effect 1, main effect 2, etc.) The interaction effects (12, 13, 23, 123, etc.)

STANDARD ERROR FOR EFFECTS

Veffect

4 2 s N SE Veffect

S TA N D A R D E R R O R
Standard error = standard of deviation.
s2 SE n

s2 47,628 SE 1,323 1,15 n 36

http://hatta2stat.wordpress.com/2011/05/21/ standar-error/

EXAMPLE-2

EXAMPLE-3

FACTORIAL DESIGN WITH REPLICATION RUNS

Checked by ANOVA. Comparing variation between treatment (ST ) and variation within treatment (SR ). Conclusion..?

ST
2 sT

nt yt y
t 1

ST k 1

COMPARING TWO TREATMENTS

Checked by Fisher Test and Student Test. Conclusion..?

s s

2 max 2 min

t t97 ,5
2 97 , 5

x1 x2 t s n11 n12