Simulation failed

[GoncaloRodrigues]

Hello people,

I'm trying to run a Cool + Fill + Pack + Warp Analysis and some errors ocurred.
The steps I made and the errors were the follows:

- Surfaces imported from a iges file and meshed (dual domain).

- Mesh "repaired" with Moldflow tools.

- Feed and cooling systems modeled

- Mesh Statistics OK.

- Start a Cool + Fill + Pack + Warp Analysis.

- ANALISYS FAILED
1ST FAIL - Not enough disk space available (56 Gb needed ???). We consider this very strange because we have older similar simulations that only needed 3Gb.
2ND FAIL - After change the temporary folder to an EXTERNAL disk with more space available the simulation failed with the error: "** ERROR 220046 ** Failure in reading Cool interface file : file name."
3TH FAIL - After change the temporary folder to an INTERNAL disk with 300GB of available space, the simulation failed without error messages (as you can see in the picture attached)

Anybody can help me, please.

Best Regards,
Gonçalo Rodrigues
MOLIPOREX,SA
Portugal Edited by: GoncaloRodrigues on Apr 15, 2010 9:38 AM

[nordhb]

Hi,
the geometry influence cooling parameter is set to Ideal.

In the Ideal method, the geometry influence calculation for each element takes into account all other elements in the model.
This method will produce the most accurate results but will require considerable disk space, and computing time.
In models with a high number of elements, the Ideal method may become impractical due to hardware limitations, in particular available disk space.

As a large model, for this model the suggestion is to change the cool solver parameter "Method of calculating geometrical influence" to Automatic
In the Automatic method, the software calculates the minimum number of surrounding elements required to produce accurate results for each element.
The number of surrounding elements used will vary from element to element.

If it still not converge, the next step would be to use GIP, geometry influence parameter.

Regards,
Berndt

[yannick.moret]

I would also consider reducing the size of the mesh.400K elements for a fusion model is huge. That is why you require so much disk space.
Ensure that your node numbering has been compacted to the minimum (you can use global merge, with really low tolerance. That should not merge nodes if you apply the right tolerance, but renumber the nodes).

Is it possible to remove some fillets, use a bigger edge length, those kind of things ?

Is the part symmetric ? Can it be possible to do the cool on half of the part ?
That can already help you to assess your cooling lines, and see if there is anything obviously wrong. That doesn't replace a simulation on the full model, but can help to already identify hot spots.

[pk_vel]

Hai,

Adding the cooling system or feed system into the other study file will increase the element ID (incorrectly). Renumbering can be done using the Global merge (0.0001). Try using the Parametric method in the cool solver parameter. Also increase the convergence tolerance and give Loose 0.2 in the cool solver (advanced settings).

Regards
velu