I have been running both dual domain and 3D side by side. At this point I am seeing large diffference in warpage predictions (in more parts than what is attached) and am wondering which solver version I should be presenting to the customer. This particular analysis entailed cooling, flow, pack, fiber and warp. Any thoughts would be appreciated.
we had similar problems with DD and 3D mesh results differences.
You should also compare other results - inj. pressure especially.
Here in discussion you can find, that 3D mesh is prefered.
Also our plastic material supplier recomend to use only 3D mesh.
Agreed - we'd only ever use 3d meshed models for anything for presentation purposes - our toolmakers perfer it as well - DD in ok for "first-pass" analysis, we typically only use it for checking models now.
My experience has been that 3D warpage is more dependable than Dual Domain. At times, Dual Domain has resulted in significantly lower warpage than a midplane model analysis, even though the two should be close. Perhaps it would be informative to construct a midplane model and see where that falls...
We now try to use 3D for warpage predictions whenever we can.
I don't agree with the others.
I have a big experience with Moldflow (around 100 studies by year).
But I don't think that 3D is more reliable then DD.
I noticed sometimes big differences between 3D and DD deformations.
In fact the material data used for the warp simulations are not the same.
With some plastics, the variations are important, not with others.
You should try different grades of your material.
The taking into account of the ribs could also explain your trouble.
I already met a case close to yours (part with ribs, material with fibres, big differences between DD and 3D).
Finally the reality was between the 2 predictions.
Maybe you could make a simulation with a Midplane model.
I am responding to a 2009 post that apparently was never resolved!
I am also seeing huge differences in warp prediction between DD and 3D mesh types with all other parameters identical. Increasing the density of the mesh has no effect on results. This is a long and slender part with some "chunky" features. The difference in maximum displacement in the DD and 3D cases is 0.0109" and 1.14", respectively. Moldflow needs to explain this kind of disparity. I'm using Insight 2014.
Even we had observed thes ame kind of variatons and its diffcult to recommned something as moldlfow user.
There are so many variables when you do analysis of DD, 3D and MID.
1. Meshing of model (key for difference)
2. Process variables
- We did a small experiments where we had kept point 1 and 2 same and take simple plate (250X150mm, thickness 3 mm) so that there will not be any variations in mesh too.
- We prepared MID first and then prepare DD of that mid by translating the mid by 1.5 mm up and 1.5 mm down to make it DD to avoid any sort of mesh issue.
- We ahve created the 3D for the DD from moldlfow.
- Results were really surprising with lot of variataions.
- We can say that it matches till flow (~90-95%) but varies in pressure and after V/P.
- From previous experice we got correlation in DD too but now we prefer to go to 3D as our cusomer observed the 3D is better in measuring deflection patter and defelction direction.
- Mid can give u better defelction but direction may be exactly oposite.
- In short there is no soultions we need to rely on our actual correlation experice.
Team Lead, Process CAE
Tata Technologies, India
3D option is more dependable.
There should not much difference in DD or 3D.
Try to run DD with
1. More finer mesh density
2. Aspect < 6. i.e 4 or 5
3. No of laminae > 12. i.e 14 or 16
4. Mesh match > 85 i.e around 90
Try to run 3D with
1. More finer mesh density
2. Aspect < 50
3. No of laminae > 6, i.e 8
4. With gravity and inertia effect.
5. warp without mesh aggregation and upgrade 2nd order for all with small deflection option.
FIAT-CHRYSLER ATC, Chennai
Please use . Accept as Solution and Give Kudos as appropriate to further enhance these forums. Thank you! .....
Material and process conditions were constant in the comparison. Only the mesh type was changed. I also greatly refined the mesh in both the DD and 3D cases with no significant change in results. The difference in deflection is huge. In cases like this, which is typically correct, DD or 3D mesh?
Would Moldflow Technical Service please comment on this.
You need to realize that the warpage solver for dd and 3d are different solvers based on different assumptions. To start with DD uses the CRIMS shrinkage data (if it's available) and the 3d uses mechanical data. With both of those options there are questions about the quality of the data. Is it tested or is it generic data? Then the solver differences will have an effect as well.
As for which is more correct? It's a similar question even if you only ran one simulation: how close are the warpage results? Sometimes close, sometimes not. Now take two "sometimes close, sometimes not" answers and make a judgement call. I would generically say it's usually somewhere in between. What is your experience with other similar parts? This is where experience with injection molding, and experience with Moldflow come in to play.
Great question though.