I may be getting off track a little bit but this thread reminds about a stress audit some of my colleagues and I were involved with some years ago.

We reviewed some stress calcs done in an overseas office. We found in several calcs, strategically placed rotational restraints just before equipment nozzles. We couldn't find these restraints on the drawings or on the stress comments, only in the computer analysis. We quickly dubbed them as MOMENT FILTERS, they worked wonders for nozzle load compliance.

Now, back to REAL WORLD moment restraints. They can be designed, but you're limited to a location where you can actually install them and restrain the rotation for real. It's controversial on how much rotation they will actually restrain but if designed properly, they'll help. I can't think of a design right off that won't restrict any forces. I think you'll have to restrain at least one linear direction.

Moment restraints should be the last resort. Limiting the linear movement causing the rotation by using conventional restraints is preferred as foglamp pointed out.