Learning Forum

[Muhammedcreed1]

Hi everyone,

I am trying to simulate a pipe bending process but i get a convergence error. I turned on the auto time steeping and also refined the mesh but no luck. Here are some pictures of the errors. Also i get a partial solution. I also have the student version and i am still learning so excuse me if i do something wrong.

[Wenlong]

Hi Muhammedcreed1,

You have a solver pivot warning message, which usually means the model is not properly constrained. Can you check your "joints" and make sure all the parts are properly constrained. Maybe you can add several images here. 

This is not relevant but using hex mesh on the pipe will generally give a more accurate result.

Regards,

Wenlong

 

 

[Muhammedcreed1]

Hi Wenlong,

Thanks for your reply,

I have made a new model similar to the one posted above and have managed to remove the solver pivot warning. I have also used the hex dominant method to make a finer mesh. Here is what i got in the pictures below. I still got a partial solution.

The force convergence graph also shows some substeps are converging. ( Note: this is not the full graph)

[Wenlong]

Hi,

Thanks for the detailed reply. So the Newton-Raphson residual shows the problematic region is where the pipe contacts the big roller. This is just a wild guess here, but can you please try varying the normal stiffness factor in contact settings (maybe try 0.1 in the first run and 10 in the second run)?

Regards,

Wenlong

 

[Muhammedcreed1]

Hi,

I have tried increasing the  normal stiffness factor and it gives different results but still not the required one.

In  0.1 it glitches and in 10 its the same as before with a warning that says 

Also i brought the pipe a bit closer so that now it bends more but still it is partial solution.