Abaqus - Forming channel problem

[tnrocha]

Hi!

I'm new here!

I am doing a subject work about simulating a forming channel in ABAQUS CAE. It's my first work in ABAQUS and I can't finish it, maybe because I have a small error, but I don't know where it is. It stop the simulation with the error : TOO MANY ATTEMPTS.

So if you can help me, it would be great.

I leave a link to my dropbox with the ABAQUS file, so you can watch it. I use ABAQUS CAE 6.81.

http://dl.dropbox.com/u/3005496/trab2_SPT_3d_original.cae

Thanks!

 

[corus]

I opened the file in 6.9.1 and there were no parts for some reason. CLick on the link yourself and see if it opens. If not then you'll have to upload it again.

Tata

 

[tnrocha]

Hi!

I downloaded it and I had no problem with the parts. Can you download and try again? Maybe there was some download problem.

Thanks for helping.

 

[tnrocha]

I upload the file in another server, try to open it now.

 

[corus]

Problem solved with reading the file.

I don't know why you've not used 2D to model this instead of going the expensive 3D route. In addition I'd use analytical rigid surfaces rather than discrete rigid elements as I'm not sure your mesh size in the discrete rigid part is causing the job to stall in the 2nd step. In fact I'm not sure why you have a first step as that just seems to be the same as the 2nd step, just a smaller deflection.

In the results you have severe changes in stress at the bend. I'd consider doing the above and refining your mesh towards the peak stresses, and through the thickness.

Tata

 

[tnrocha]

I change:
-radius (5 to 10mm)
- discret rigid to analytic
- refine mesh

It works better but it don't finish yet. What can I do more?

ps:Is the mesh too much refined?It take too long to calculate.

 

[tnrocha]

Just a question.
How can I refine mesh through the peak stresses?

 

[amubashar]

From your model it looks like you are trying to do something similar to Forming a Channel Example in Abaqus Getting Started Manual (Infact your material data looks almost the same, may be I am wrong). As mentioned already, there is no point in using a 3D model for this problem. If you are doing something similar to the example problem I mentioned earlier, it would be easier to go through the manual.

 

[tnrocha]

Hi!
I did 2D simulation but now I have to do 3D do compare results.
I follow the example you mentioned (2D right?) and made some changes to make it 3D, but I can't finish the job.
The error is allways the same. Too many increments, and it's my first work in Abaqus so I don't know how to solve it, and no one can help me in my university.

 

[amubashar]

You have to think about what would be different in 3D and would you be getting any different results than 2D. If you simply want to look at results in 3D, the 2D model can be extruded during post processing.

The Abaqus 2D example discusses a number of issues encountered in contact problems like contact chattering, problems when establishing initial contact, etc. Look for similar problems in your 3D model. Since there is bending involved, incompatible mode elements may perform better. Reduce the problem size by considering symmetry.

 

[corus]

There is a difference between 2D and 3D in that in 2D you'd assume either plane strain or plane stress. Plane Strain would be more applicable but at the free edges it would behave as plane stress. A 3D model would naturally include both effects. One thing I forgot to include was *contact controls, automatic tolerances.

Tata