Force field design is relatively detailed work, and people usually tend to underthink more than overthink. Your question is a pertinent one. First of all, for the record, only one potential energy scan is needed per rotatable bond. This leaves you to fit 3 redundant dihedral parameters to 1 energy profile. What I usually do in this kind of situation is constrain the parameters to be synergetic and have the same absolute amplitude, as this tends to decrease the chance of unwanted forces and improve transferability. Since your phase difference is 180°, this means that the even multiplicities should get the same phase and the odd ones opposite phase. Your case will likely at least need a 1-fold and a 2-fold, so your fitting problem would look like this:

CG2DC3 CG2DC1 CG2O5 CG2R61 x 1 0.00
CG2DC3 CG2DC1 CG2O5 CG2R61 y 2 180.00
CG331 CG2DC1 CG2O5 CG2R61 x 1 180.00
CG331 CG2DC1 CG2O5 CG2R61 y 2 180.00
CG331 CG2DC1 CG2O5 OG2D3 x 1 0.00
CG331 CG2DC1 CG2O5 OG2D3 y 2 180.00

Where x and y are the sought-after variables. In the above formalism, negative solutions are acceptable, but by convention, we make them positive again by flipping the corresponding phases in the final parameter file. In case you need further explanation with all this,

this would be a good starting point.