Losing group variables when removing/importing FEM

[trelau]

I have HyperSizer v6.2 integrated into a larger multidisciplinary environment and am driving it using VBA. As of right now, the general procedure is as follows:

1) Remove old FEM
2) Import new FEM
3) Push the new components into the appropriate groups (groups/variable bounds/available materials/failure modes are already defined in the project)
4) Size the structure
5) Update the FEM (generate _i.bdf, etc)
6) Manually re-submit the job to Nastran and repeat (mimicking HyperFEA)

Everything seems to be working but when I remove the old FEM, I lose all specified laminates (top skin, stringer web, and frame web) and failure modes for the rod/bulb stiffened panel family (PRSEUS).  Other variables (e.g., stringer height) and other group materials/variables remain intact.

I also noticed that if I try and save this group information without an imported FEM I get an error.


Thanks.

-Trevor

[Phil]

I'm not sure what you mean by  "I lose all specified laminates", but I know what you mean about losing the failure modes.  When you remove the FEM, the components are deleted, so any component based settings are deleted as well.  Since failure mode specification is component based, they are lost.

When you import a new fem and assign these components to groups, then the default component settings in the database are retrieved.

One way I can see around this is before you remove the FEM, store component designs for each component and then after the new model is imported and components assigned to groups, retrieve the stored component design then your failure mode settings should be back in place.

I hope this helps.

[trelau]

Phil,

Sorry for the vague description. What I meant was when I remove the FEM, the laminates that I have made available for things such as the PRSEUS skin, stringer web, and flange web, are lost. This might be related to the same issue causing the loss of the failure modes. I just checked and it looks like only the laminates are lost (any available isotropic materials remain).

The main challenge is that my component numbers change (e.g., number of wing ribs changes), so for each iteration I have to push my new components into the appropriate groups. I was hoping that everything else would remain unchanged.

Thanks,
Trevor

[Phil]

Is it possible that the laminates are not lost but rather the continuous/laminate radio button is switched from laminate to continuous for each variable?  The laminates should not be lost when you remove the FEM.

If that is the case, I think there is an object model hook to change the setting from continuous to laminate.  When you set a group variable bound, one of the settings is "vbdLaminateOnly" If this is set to true, then it will change that variable from continuous to laminate... if set to false it gets changed to continuous.

[trelau]

It appears that the laminate switch is left on laminate, but the laminates are removed. I also noticed, that if I remove the FEM, then re-assign the lost laminates, and then try to save the group, I get the error shown in the attached screenshot. I would send you the database, but I think the file size is too large (~160 MB).

-Trevor

[Phil]

Laminate are different from other materials in that they can be set for specific components during the automated sequencing process.

It appears that when you remove your FEM, HyperSizer is removing all laminates even though they aren't associated with particular components.

Then, it won't let you assign any laminates unless you have at lease one component in the model.  This is something that can be fixed, but the fix won't be immediate.

The best workarounds I can think of:
a) assign your laminates back to groups after your import your model
b) before removing the FEM, store all of your groups (store group design for project) and then after importing the FEM, retrieve the group designs.

[trelau]

I think I have a couple temporary fixes. First, I have a separate database for this particular project so I can set the desired failure modes and then save those settings as the database defaults.

Instead of assigning laminates I'll create assemblies and use the composites form to design laminates. I was going to do this anyway, but it would be nice if I could specify laminates just to get the whole environment running.

Thanks,
Trevor