I simulated a part (2K) and the comparison for the real part shows, that the deflection of the real part is ~5-10 * deflection (Sim).
How can i improve the Sim? Anybody an idea?
I used for the Sim following setting:
Part: .x_t (Using MDL2013) importet and meshed both parts (MDL)
Warpage analyse type: small deflection (isolated cause of warpage & all tetras upgraded to second order)
The warpage analysis settings creates great results in 1K.
A large part of results with warp has to do with material quality. I have seen parts warp the completely wrong direction in moldflow, and others match exactly with what we were seeing at the press. In every case, it has to do with material having CRIMS data, or the model type. For materials without CRIMS data, or materials that have very old data, I always use 3D meshes. It takes a longer time to solve, but the accuracy is significantly improved.
For 2K components, it is especially important to have mesh contact set to "precise match" - this allows warpage of shot 1 to affect the warpage of shot 2. If the mesh isn't matched perfectly, the software has to interpolate the stress (and thus displacement) values for any nodes that do not line up between the two parts, and result accuracy will suffer.
Let me know your thoughts on this as I am still experimenting with this process and doing some validation on test parts.
Thank you and best regards,
i used "perfect match" for the contact faces for component2.
As a see the real part, i know why the part is deformatet. The 2nd component remelt a critical area and the 1st component "relax" in this area and deformate. (it´s a symmetrical part design) The deformation through the remelt of the 1st component causes in a ovality of the part in a range of 8-10mm (diameter = 90mm). The Sim-deflection is around 0,8 to 1mm.
I mesh the two parts at the same time and i saw that the mesh match betwenn 1st and 2nd componend is nealy perfect.
i run the simulation in 3D.
This sounds interesting!
I'm pretty sure you have a pretty good Dual DOmain mesh for both parts, before switching to a 3D, in terms of Quality criteria, so I will scratch that out; and I will also assume material properties are Gold-Gold-Gold for the sake of the post extension. =)
So, the first step I would follow is to evaluate the amount of information you have included in your analysis, in terms of waterlines, material feeding system and so on. how complete is the model...
After that, I would focus on correlating the filling progression you have in reality with your predictions. Any difference is a red flag here.
And last but not least, I would make sure you are simulating the same packing profile used in reality as well as the cooling time on both shots.
also, why not take a look at the way you are ejecting the part from the mold.
There are many cross check points
- Cad geometry : part, feed system sizes.
- Mesh : type and quality
- Process paramters : melt temperatures, mold temperatures, injection time, packing pressure, hydraulic pressure, mold temeprature , coolant tmeprature
- coolant channel configurations : coolant flow rate, coolant channels modeling and hose connections, coolant temperatures & Mold temperature varations
- cycle time: components and varaitions.
- Mold machine paramters : screw size & effect of injection.
- Raw material : Properties of material, grade used & quality of data
FIAT-CHRYSLER ATC, Chennai
Please use . Accept as Solution and Give Kudos as appropriate to further enhance these forums. Thank you! .....