I see one problem immediately with this approach:
As you step the torsions, one might expect internal bond angles to change.
In particular, as VDW interactions rise as 1 / R^12, as two atoms get
closer during the torsion rotation, internal bond angles are likely
to open to relieve the VDW energy. Look at the example of
the central torsion of butane:
If T = 180, C1 and C2 are maximally distant and VDW energy in minimized
If T = 0, C1 and C2 are closest and one or more of the internal C-C-C bond
angles opens to relieve the strain.
Hope this helps,
Dr. Andrew Smellie,
1804 Embarcadero Road,
Palo Alto, CA 94303
email: asmellie-: at :-combichem.com