BEM cooling not converging after changing insert property

[Adam.Mayes]

Hello all,

Ran a Dual Domain BEM cooling analysis with mold inserts. I forgot to change the insert material to moldmax so the analysis ran with the insert material as the default P-20 steel. I realized my mistake and re ran the analysis with the proper insert material(Moldmax HH 40). First analysis with incorrect insert material(P-20) ran no problem in 35 minutes. Now the second analysis with correct insert material runs for hours and then fails.

Not sure why. Log file doesn't tell me anything.

Any help appreciated,

Thanks 

Adam

[bernor_mf]

Hi Adam,

first things that comes to my mind:

Does cooling circuits go through the insert?
If, does the cooling circuit beams passing trough the insert have mold properties of insert?

Check and fix and retry.

Not sure it helps, but maybe worth trying?

If no info in analysis log, then difficult to nail.

Regards,

Berndt

[Adam.Mayes]

Hi Berndt,

Thanks for the reply.

The cooling lines do go through the insert. I did check that I changed the mold properties to the correct insert properties.

Still having issues converging. There must be something that I am missing. I have reconstructed the inserts, re imported the cooling lines and still no luck. There are trapped beams in the cooling lines but had the same beams in the analysis that ran fine before I changed the insert properties. I always seem to have trapped beams even when meshing to ideal L : D. but that is a different problem! Not sure why moldflow creates a trapped beam when meshing a baffle. If you change the length of the beam to eliminate the "trapped" beam then you get a beam with bad L : D ratio?

[bernor_mf]

@Adam.Mayes 

Hi Adam,

maybe you are aware of  this, but will highlight it anyway, as the issue you have is a bit difficult to nail.

For BEM Cooling.

When modelling inserts and channels passing through it is recommended not to have intersecting insert triangle elements with cooling channel. You could simple remove triangles at intersection:

It might work and converge with intersecting elements, but unexpected temperature results might occur.

In the insert the cooling beams need to have properties of insert material:

Have you modeled cooling and set properties in insert as above?

Regards,

Berndt

[Adam.Mayes]

Hi Berndt,

I modeled exactly as you have shown with appropriate mold properties.

One question, does the beam element have to have a node exactly on the place where it passes through the insert wall?

In my model the beam element that passes through the insert wall is half in and half out of the insert.

This maybe causes a convergence problem?

Thanks,

Adam

[bernor_mf]

Hi Adam,

from what I know the node for channel does not need to be a in plane position for an insert surface  (or a mold surface).

As you have modeled looks okay to me.

Regards,

Berndt

[bernor_mf]

Hi Adam,

just wanted to follow up, if you got the Cool analysis to converge?

Thanks.

Regards,

Berndt

[Adam.Mayes]

Hi Berndt,

Good timing, I was just about to send you an update as its been a few days.

I did not get it to work, so I got moldflow tech support involved. Sent them the 2 models. The one that ran and the one that did not run. They got the same results, It did not converge. Justin at tech support was making  some changes to see if he could get it to work. There were a few "warnings, elements too close" in the log concerning the mesh but I don't think this was causing the issue. I get similar warnings on the mesh on many models and have no issues.

Waiting to hear back from them today. 

One thing of interest was he ran the model in the new "Beta" version of moldflow 2021 and it ran with no issues!

Apparently it is a little more friendly to things that cause convergence problems.

I will let you know how things turn out.

Thanks for staying in touch!

Adam