Effective way for Gusset Plate and Bolt Check Design

[tatox]

Hi.

I am currently working for bridge design. For steel bridges truss, the client want to model each gusset connection in FEM. Just like I-35W Bridge Gusset Failure FEM.
They want to see the gusset plate stress, and the bolt stress. The load is static elastic for some case due vehicle load or moving load if it could to get the maximum forces at each member that connected to gusset.
I propose to use shell for gusset and the beam for the bolt, that connected via spider compression only link to get bearing pressure to the hole plate, and some stiff link for stabilize around near hole for bolt tension/compression/fix rotation. Since using compression only it will non linear situation. The gusset can be connected to truss-beam bridge whole structure using constraint rigid link at connected members.
But this quite take a while, sometimes non convergence.
If i use rigid link, like RBE2 spider it will static situation, but will not get real bearing stress at hole due shearing bolt, right?
So what the best one to efficiently get plate bearing stress at hole gusset?

 

[rb1957]

OMG ! such a fine mesh !! Personally I wouldn't model individual fasteners, but one node for the bolt group. If I wanted to really "fuss" the joint, I'd figure out a finite stiffness CBUSH spring, or use 10^6 as a typical number (as opposed to a "rigid" spring, like using a common node). Modelling the individual fasteners as spearate nodes is as far as I'd go.

I'd use a hand calc for the individual fasteners.

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

 

[tatox]

I just found the FEA of Bolted Joints - User Guide Seminar by AdaptiveCax on youtube and its model explanation on their website quite useful. They even compare the RBE2/RBE3/compression bolt/contact edge vs solid bolt. They can done the calculation using RBE3 for plate bearing load on the bolt connection using Static linear. OR i am wrong?
I thought i need several compression only spring to the cbeam in nonlinear static analysis for plate bearing bolt problem, but hey there will be alot time and singularity.
Because tony abbey said in the spider session, to get plate bearing load, use rbe first than use the force reaction rbe to get the local stress on the plate.