Dealing with Safety Factor issue

[jasonhomrighaus]

Hi, Im modeling out a machine and am using Simulations to refine and test the design.  At this point I have it pretty solid and when testing Im getting less than .1 inches deflection over a 10 foot span etc.  Almost entire machine is cool blue with almost no stress, but safety factor reads in the toilet due to little tiny spots at the inerfaces of some components where there would normally be a weld bead to spread the load out evenly.  

Can anyone suggest a methodology to clean these up, ignore them or otherwise make them not tank my Safety estimate without having to work through the entire machine creating weld beads and pentration simulations which could take more time than it took to design the machine.

Gut says this thing is plenty strong  but I cant generate a report to back that up. Note in the picture below a view of the failure spot.  its at the edge of a 3" wide tube that mates to an 1.25 steel plate boss.  once welded there is no way this will fail.  For what its worth the machine is loaded with a 7000lb off axis load and you can clearly see how stress free everything else is and how concentrated the trouble spot is.

Any thoughts?

[tyler_henderson]

Hi Jason,

This is most likely a singularity cause by a sharp angle where the tube intersects with the plate.  To verify this you could try running the simulation with Adaptive Mesh Refinement turned On (it's under the Manage panel).  Then check the Convergence Plot.  You'll be able to see if the stress converges or not.

You might also try running Nonlinear Stress with Nonlinear materials.

If the stresses do not converge, then there is no way to "hide" the singularities such that they don't affect the minimum SF.  But you could simply put a fillet in one area and re-solve to show the effects that a fillet weld will have and highlight that in your report.