How to bend tube to evaluate stress?

[gillyboy]

Am a newbie to Cosmos and want to perform a 90 degree bend on a straight extrusion to evaluate the stress or distortion at the inner radius (if any).

Besides fixing one end of the extrusion, I am not sure which other restraint to apply to perform a 90 degree bend?

Please advise.

 

[corus]

Use quarter symmetry and restrain both symmetry faces. Apply a presrcibed displacement to the free end to bend it.

corus

 

[gillyboy]

What displacement method should I apply to the free end for bending? I tried to apply a restraint on the free end with reference to an axis and a 1.5 radian (to make the free end rotate 90 degrees about the axis), but it didn't quite work. It keeps prompting to turn on large displacement flag. If I pressed 'yes', the static study will fail. If 'no', it gave some result but only with a slight bend and nowhere near 90 degrees.

 

[bassmanjax]

Do what Corus says and use 1/4 symmetry to cut down the model size for starters...

I suspect that you will definitely need to perform this as a large-displacement (non-linear) analysis, so you will have to have the large displacement flag on.

I'm not all that conversant with COSMOS (we only have the linear-static bundle here and use ABAQUS for everything else ), but could you create a rigid (non-deformable) roller to bend the extrusion round with a prescribed displacement (as Corus suggests) on the free end. Or use another rigid surface to 'push' the free end round the roller?

Martin

 

[Mech151]

This is a nonlinear analysis (not to be confused with large displacement), you will need advanced professional for a legitimate analysis, as you are interested in behavior well beyond the yield point of the material.

 

[corus]

Try a smaller displacement value to see if you get an answer with large displacement switched on. The radius of the bend might be a good estimate of the translation needed. I don't see why the job failed though as you're applying a prescribed displacement rather than a force which might cause instability problems.

corus

 

[ERT]

I agree with Mech151, it is essential to consider this as a non-linear analysis.

The material's yield and plastic deformation stress/strain behaviour are key to analyzing this problem.

Also, the final inner distortion of the tube will depend on how you model the bending dies and what assumptions are made for clamping and friction. Also, springback may be significant as well.

ERT

http://www.akeng.com

 

[Mech151]

Thanks ERT,
I would add the following:
1. For such a large displacement, your model is probably experiencing element locking (deformed elements exceed geometry limits).
2. Even better than an analysis for this problem would be to talk to a fabricator (tube bender). They will be able accurately predict the distortion that you will see. Depending on the radius of the bend, they may take counter measures like filling the tube with sand. We use a couple of good tube benders, let me know if you would like contact information.

 

[Tmoose]

Unless the bend radius is huge or the tubing wall constrained with a mandrel or shoe, inner sand fill, or the correct answer may well be buckling or kinking.

http://structures-

http://www.cv.imperial.ac.uk/staff/wadee/pics/tubes/bendkink.jpg

http://structures-

http://www.cv.imperial.ac.uk/staff/wadee/tubes.html