Discussion Forum

I am trying to do a transient analysis of vibrations through a metal cylinder with end flanges. The end flanges are clamped on by bolts, and I need to simulate this clamping effect on the vibrations on the outer surface of the end flanges.

My problem is that no matter what I do I cannot transition from an initial stress to an equivalent constant stress when the simulation starts. Inevitably there is some snapping of surfaces into position as the physics is turned on at t=0. I cannot use a fixed constraint to hold things in place because I do not want to eliminate transmission of vibrations on the clamping surface.

My best attempt was to supply an initial stress along the clamping axis (Linear Elastic Material -> Initial Stress and Strain) and then have an equal pressure applied to the clamping surface (Boundary Load). It appears that there is some buckling of the cylinder at t=0 which causes a skewed displacement in the clamping surface. This then results in vibrations which interfere with vibrations intentionally produced.

I suppose I could adiabatically apply a clamping pressure manually before producing the desired vibrations, but this would likely dramatically increase the simulation time. Is there any way to have COMSOL adiabatically apply a boundary load automatically? Alternatively, is it possible to somehow use the output of an initial simulation that has come to equilibrium as a static input to further simulations?

Any help is appreciated.