Difference of modeling jointed bolts using CFAST vs CBEAM(CBAR)+RBE2 in NX / MSC Nastran?

I would analyze the screws by hand. I would model the structure so I could easily extract the load that the fasteners experience. There are different ways to do this, depending on how many fasteners you're talking about ... <10, >100.

another day in paradise, or is paradise one day closer ?

Why would you want to attempt to model bolt preload?

Hi SWComposites, bolt preload would be important to get pre-stiffening effects in the structural attaching zone, after bolt preloading.

Essentially every airplane designed to date has not included bolt preload in the stress analysis.

If you want to investigate preload effects you will need to use a very detailed 3D FEM, and even that is likely to be only an approximation of little practical value.

What exactly are to trying to analyze?

If you are modelling individual fasteners and are worried about preload effects, a hugely significant factor is going to be preload scatter ... which will lead to an "infinite" number of models. Of course you could say ...
1) assume all fasteners nominal preload,
2) assume all fasteners minimum preload, or
3) assume all fasteners maximum preload ...
but these are so coarse that the assumption will swamp any significant effect (which is IMHO insignificant).

From a hundred years of designing preloaded bolts, I think we understand how to design safely. And how "sharpening the pencil" will usually equate to maintenance costs (inspection or replacement). Preload is generally only significant in tension joints under fatigue loads.

Now, of course, you could be in a field where money spent on analysis is not viewed as a cost, and extracting the last "psi" (stress optimised) or the last "gram" (weight optimised) from the structure is more important. In that case "fill your boots" ... pick any appraoch and analyze away. Remember you'll need a full ground test program to validate this analysis.

another day in paradise, or is paradise one day closer ?