i am filling a mold which has 2 injection locations. When i look at the filling pattern of the molten polymer, it looks a bit strange (please see the attached pic). This is a snapshot at an intermediate time step. You can see that there are patches which are surrounded by molten polymer. I dont think this can happen in reality. Has anyone seen this kind of behavior or know why is it happening?
PS: the mold temp is uniform all throughout. there aren't any local cold spots. This is a solid mesh.
I am using solid mesh and I'm simulating insert over molding . The mesh has almost 1.5 million elements.
I guess you are using Hybrid mesh and may be your boundary layers elements need to be thicker, you can increase the default value (0.1mm) of minimum thickness of boundary layers
example:
Hi Loftil,
Yes i am using hybrid mesh. Currently, I have specified the minimum thickness of boundary layer elements to be 0.001 mm. The reason being, the insert is very thin (0.127 mm) and the part is thick (~6mm). Meshing became extremely difficult because of the difference in the aspect ratios of the insert and the part. Hence, to get substantial number of elements through thickness in the insert, i had reduced the minimum thickness to 0.001. (please see 1st pic)
I will try and increase the minimum thickness by to 0.01 and see if that is helping me in the fill pattern. Another question comes to mind, we are interested in the tempearture, deformation, stresses etc. in the insert. Will increasing the minimum thickness reduce the number of elements through the thickness, which may reduce the accuracy of results?
Also, in the current model, the fill pattern is strange (like i had mentioned before). The stresses are also similar to the fill pattern and they don't make any sense. (please see pic 2)
Nevertheless, i will increase the minimum thickness and see what happens.
Hi Lotfi,
I did ran a model by increasing the minimum thickness of the and the results haven't changed much. Please see the picture below.
I guess there is a symmetric plan we can use symmetric analysis so that you can deal with more accurate cells and better aspect ratio for boundary layers
as well the option of direct solver may help (see Options in Advanced tab of Fill settings), but the CPU time will be longer
the best option is the pure Hexa mesh with non-orthogonal mesh, but the risk is to increase the cells numbers too much ( again the symmetric option could help), 3 layers at least could be OK
in case you can share the model we can ask the R&D team to have a look or you may file a support request
ok... this is the plan i will follow
1. try the hex mesh
2. direct solver with hybrid (current) mesh
3. symmetric analysis
yes, i can share the model. Please let me know how to do it?
1. i tried the hex mesh. But the non-orthogonal mesh type fuses the insert and the cavity. Is there any button i need to turn on to keep them separate?
2. running the hybrid mesh with direct solver. is this the way to switch it on?
3. can you let me know how to switch on the 'symmetric analysis?
4. lastly, how can i send the model?
Thanks a lot for your help! I really appreciate it.
Nikhil
I tried using the direct solver and the results haven't changed.
Hi Lofti,
Is it possible to generate a support request for the issue I am facing?
Nikhil
are you using Solid mesh or Shell mesh?