230526 - Design of Experiments

Question 1

How to know if it is a full-factorial design?

Answer 1

2^2 = 4 runs

Answer 2

Interaction diagram: 2 lines
Mean effect: 1 line

Answer 3

if you only have the corner points, thats no proof that there’s no interaction between the effects, additional points could be

Question 4

What is a regression analysis?

Answer 4

Determine the coefficients for the regression analysis.
with corner points you can only have linear polynomials in the regression
the function can be extended to more effects (A, B and C; or A, B, C, D and E and so on)

Question 5

range of values for fatigue stresses

Answer 5

does it depend on the material? does it depend on the shape of the part?

Question 6

What is T-Test?

Question 7

What’s the goal of doing this analysis?

Answer 7

get a mathematical model of how the behavior of the system will be

Question 8

How do you know if something has no interaction effect if you only get the corner points?

Answer 8

The center points may have a different shape, since the functions may not be lines.

Question 9

Resolution of an experiment

Answer 9

Ideal resolution? How do you define the colors of each resolution level? yellow? red?
How to reduce the number of tests in order to make it more cost-effective or faster? Do the first half? the second half?
since the target is to use only a half and the number of experiments is given by: 2^a, you just need to do 2^(a-1)
(delete the one factor only interaction and replace it with the two or all factors, starting to 3?).