High Stress at Sheet Metal Curves

[Penguineer]

Hello,

I have a welded structure that I modeled using surface bodies. All bodies intersect at what would be welded joints and all lines are split to ensure node-to-node bonding. Something I typically see with models like these is that I have relatively high stress right where one of the members terminates into another, particularly at the curved portion. I sense that this is not a realistic hot-spot and that it has something to do with the math model, but I don't know how to explain it. Could someone provide me an explanation as to why I see higher stresses where the geometry transitions from a straight section into a curved section? Your help is most appreciated.

The first attachment is the joint as modeled with the intersection highlighted. The second attachment shows the nodal VM stress.

Thanks!

 

[corus]

Refer to Roark, Stress & Strain, and look at the stress concentration factor for a fillet radius. As the radius tends to zero, then the factor increases to infinity. You have a zero radius at your juncture.

 

[Penguineer]

Thanks for the replay Corus. The point of concern to me is the stress near the curved region of the juncture; not so much the juncture itself. Following along the path of intersection, you can see relatively low stress up until you get to the bend. I don't see why it would be any higher at that region that it would on the vertical/horizontal regions.

 

[corus]

In general, the load path follows the stiffest route and for a box section this will be at the corners. In addition you may have bending about two orthogonal axes at the corners and so your stress intensity is highest there.

 

[rb1957]

i'm trying to understand the stress pic ... from the model pic i thought you modelled the bent piece with 2D elements, but the stress pic looks as though the larger piece's mesh is following the bent piece ?

are you spot welding (joining only relevent nodes) or line welding ?

is this a contact problem (have you constrained the larger piece from bending into the bent piece) ?

 

[Twoballcane]

I’m not allowed to open files, but maybe you have a “singularity” at the curve?

Tobalcane
"If you avoid failure, you also avoid success."
“Luck is where preparation meets opportunity”
"People get promoted when they provide value and when they build great relationships"

 

[Penguineer]

Perhaps this will clear up some of the confusion with the model. What I have is a C-shaped formed channel terminating at the web of another C-shaped formed channel. The orange line follows the intersection where nodes are aligned.

This is modeled as a full seam weld completely around the C profile.

I think Corus may have the answer to my question. With the load path following the stiffest region of a member, that would result in a higher load concentration near the corners of the "C" profile, which produces higher stress. My thought was maybe it had something to do with the elements transitioning across curved geometry, but perhaps it is a reasonable model.

Thanks!