Modeling of bolted supports

[spongebob007]

I have always modeled bolted supports using fixed boundary conditions. For example, lets say I want to do an FEA model of a cantilever beam type structure that is bolted to the wall. I would start by using beam elements and fixing the end that is attached to the wall.

Recently I wrote a structural analysis report for a structure that is bolted to the floor. At the locations where the bolt holes are I applied fixed boundary conditions. The engineer who reviewed the report for the customer commented that the fixed supports I used for the bolts should be changed to pinned supports and the model re-run.

I can't really rationalize how this is a pinned support. If there were one blot I suppose the the structure could rotate around the axis of the bolt. However, with 4 bolts that motion is restricted. I would also assume that the head of the bolt can carry some moment when the load is along the axis of the bolts.

So which is correct, pinned or fixed for bolted connections?

 

[ESPcomposites]

It should not make much of a difference, since the moment is largely reacted by a couple of the fasteners. Provided the coupling distance is relatively large, there should not be much moment in the fastener even if you do fix it. The reality is that the fastener is somewhere in between...so I would take the most conservative approach, which is probably to just assume it is pinned.

Brian

http://www.espcomposites.com

 

[corus]

The worst case for assessing the beams would be to assume pinned members. The worst case for assessing the bolts would be to assume the beams were built in and the bolts had to carry the bending moment. The true case would be somewhere inbetween. I'd run both cases and report on each.

 

[rb1957]

there a couple of things here. i think the reviewer said that your "monument" shouldn't be fixed at four points to the floor to the plane. Very correct, if you do this then 9g fwd creates a large lateral couple +ve on one side, -ve on the other (so internally reacting); and if they're single bolts (or closely pitched bolts) they probably wouldn't react much moment. Instead i think you should model the attmts as three rods (aligned X, Y, and Z), and the far end of the rods is a ground. Next you'll ask "how stiff ?", and I'd answer pretty much any stiffness, i like 1in2, but the difference should be small (try different areas 1in2, 0.1in2, ...)