I am currently working on a simulation to test the upright/knuckle of a car I'm working on with a team at school.  In order to create more accurate results we decided that we should add bolt connectors.  However, there are joints on the upright where a bolt connector will be attached to the same part on both ends and I remember how this complicated things for me in the past, so I decided to run a smaller, simpler simulation to be sure it will work.

Above is a simplified version of the upright to wishbone (control arm) joint.  It is fixed at the wishbone, like in the full scale assembly, and a force of 100 lbf is applied to the tab where it would be attached to the upright.

This tends either fail, need large displacement (which will fail), or solve with a diagram showing the entire thing yielding.  I don't think the yielding is accurate, considering it is 1/8" walls and steel. We also ran it with a finer and adjusted mesh on a more powerful computer, but it ends up failing. 

Can anyone point out what we are doing wrong. We would greatly appreciate any advice.

FYI, I originally tried with a more open tab, and no washers, but that completely fails, so I enclosed the joint and added in the washers, all with no penetration contacts and a bolt series option selecting the inner faces of the joint and washers. Still no good though.

Thank you.