First time approching DOE tecnique - help -

Hi everyone,
I am Dario and I am a PhD student.
I am new in DOE tecnique and this is the reason why I am asking you help.

I must provide a DOE tecnique in order to obtain an analytical correlation... in particular, I have 3 factors (A,B,C) and 3 level for each one.

A 0.59 0.69 0.77
B 16.2 24.2 32.2
C 3.2 3.4 3.6

I am using the book "Design and Analysis Experiments" by Douglas C. Montgomery; I read that "the 3^k design is not the most efficient way to model a quadratic relationship" and that is better use this sequential strategy:

- Use 2^k design augmented with center points and obtain an indication of curvature;
- If the curvature is important, augment the 2-level design with axial runs to obtain a CCD.

Can someone explain me clearly how to procede with this strategy?
Maybe with an example??

I don't have much time unluckly to study all the book because my advisor give me only 3 days to solve this issue.
So I really appreciate if someone could help me.
Thank you, in advance.