How To Run FEA on Nut-Bolt Assembly 1

What is the nature of vibration (steady-state, transient, random) ? How is the structure supposed to be supported and loaded ?

I am thinking of applying a cycle of steady-state vibrations (i.e. apply steady-state vibration n amount of times and see the affect it has on the clamping force). The c-channel beams are used for warehouse racking, the nut-bolt assemble 2 safety bars perpendicular to the c-channel, so there are 2 c-channels with 2 safety bars perpendicularly assembled to the c-channels. The loads are applied to the safety bars.

Then harmonic analysis will do the job. You just have to connect the parts (with bonded contact), apply boundary conditions and loads. They will be harmonic by default and you can also define their variation with frequency. Then specify the range of frequencies to do the sweep.

What mechanism do you suppose will change the clamping force?

I need to find a change in the clamping force due to vibrations that will occur during transportation and when loaded. Something similar to this paper: On the Anti-Loosening Characteristics of M16 Threaded
Fasteners under Vibratory Conditions

The goal is not to have me read the paper and extract the mechanisms you are looking for. The goal is for you to decide what factors there are so you can match them to what an FEA can actually simulate.

To give a hint - the major reason for loss of tension is embeddment - randomly located variations in high spots on the order of 0.0001 inch take the majority of the initial load. Repeated high loads cause these highly stressed points to deform, lowering the overall separation and the preload. To simulate that you need to have the ability to generate a randomly rough surfaces and allow for incremental deformation, possibly with very small areas of fatigue failure. The overall stress levels will remain in the elastic region so a typical FEA will never detect/simulate the effect.