Extremely large distortions

[Daniel N.]

Hi all,

in the forming simulation of rotary draw bending, I have a problem with very large distortions. A bending process is simulated there, in which a U-profile is clamped between the bending die and the clamping die. Both tools then rotate around the bending axis. When bending, extremely large distortions then occur on the inner profile web. At this point, the profile is in contact with the bending die. What you see there is certainly not a physical effect but a numerical problem. But I'm already not sure if this is a problem of the contact or a problem of the elements used. The simulation is an Explicit simulation and I have already tried the General contact and also a Surface to Surface contact. I meshed the tools as rigid bodies with R3D4 elements and the profile with C3D8R elements. Does anyone have an idea what the problem or the solution could be here?

Here is some more info about the simulation:

• NLGEOM is activated
• The problem occurs both with and without mass scaling
• The speed is approximately the same as in the real process
• ALLAE is about 3-4% of ALLIE
• The mesh on the profile is 0.5 mm with 5 elements above the sheet thickness

 

[NRP99]

Maybe you can change element type or use full integration or hourglass control. Check help manual for solved forming problems. Some hint about distortion will be available.

 

[FEA way]

Based on the pictures, it seems to be a typical case of hourglassing and standard approaches advised in such cases may help. But it’s hard to notice what exactly happens there based just on the zoomed picture. It would be also necessary to take a closer look at the global deformation of the model and interactions of the parts. In Explicit similar effects occur when regions of the model get stuck or struggle to slide smoothly in contact. In such cases it’s often necessary to refine the mesh or slightly modify the model. Check the other global energies (e.g. ALLVD), they may indicate what’s going on.

What is the scale of the deformation in this case ?

 

[Daniel N.]

First of all, thank you for your tips. I had already used Hourglass control in the model. And since ALLAE is only 3-4% of ALLIE, I would have actually excluded Hourglassing as well. I have now nevertheless made another simulation without reduced integration. But even here I get these distortions. A simulation without friction on the profile webs also resulted in the same distortions. I had not set ALLVD as an output, but I can still do that. In any case, I did not notice anything with the other energies. In particular also not with ALLIE in the comparison to ALLKE. In any case, the local distortions there are significantly larger than the global strains that would be expected from the bending.

 

[FEA way]

Hourglassing can be very local and then it’s less visible in global energies.

Maybe, apart from additional mesh refinement in that region, you could also change the geometry a bit. For example add some fillets and smooth the edges of the workpiece and rigid tools. This often helps with similar issues.

 

[NRP99]

Try following.
-Second order elements with full integration and Hourglass control
-Distortion control
-Adaptive meshing
-In help manual check solved problem "Indentation of a thick plate"

 

[Daniel N.]

Hello,

thanks again to both of you for all the tips. Finally, I was able to solve the problem by rounding the profile bars. It's a pity that there is no other way, but this small radius should not play a big role for the forming behavior of the profile. Therefore, I will continue to do so for the time being.