6.1.8. Cubic Face Centred

Your Ad Here

Figure 6.13. Cubic Face Centered Panel

This method is equivalent to a full factorial with two levels plus the mid-points of the design space hypercube.

The experiments are placed in the design variables hyper-cube as follows:

On each vertex (2^nvariables points).
On the centre of each face (2*nvariables points).
On the hyper-cube's centre (1 point).

Figure 6.14. Cubic Face Centred of 3 variables (15 experiments)

Hence the total number of generated experiments is given by:

        N=2^nvariables+2*nvariables+1

This method allows the computation of 2^nd order interactions and can be useful when the problem is weakly non linear and a full factorial with three levels is too expensive.

Note:

Even if N is greater, the maximum number of generated designs is limited to 64000.