I've had some nice discussion at the Euromold fair recently.
Since there are many "rants and raves" about meshing, mesh types, mesh sizes I would like to ask the Moldflow team and
others Users about their experience with meshsize and element count.
1. Could it be that a fine mesh with a homogenious mesh size calculates faster than a mesh with inhomogenius mesh size?
2. Does a mesh with a good AR calculates faster than a mesh with bad AR at the same element count?
(This questions goes without regard to solution quality!)
Very obvious I could do a benchmark on my own, but maybe we save some time if we share some experiences here.
3. If there is something lieke a rule about the questions above, does the suggestion of mesh size in MF takes that into accout?
Other question I have aremore related to different approaches like "skin and core mesh"
(Moldex, Simpoe) or Hexmesh/ Voxel approach (Sigma, Simpoe).
There seemed to be advantages as well as disadvantages about every approach, but I would be glad to have a choice
and make meshing of whole tools easier than it is now (as an example).
Calculation time is not the main factor, if you see 40.000.000 Hex elements builing a solution within 2 hours (not really bad).
Best regards
Harald
Solved! Go to Solution.
Solved by xusho. Go to Solution.
Hi Harald,
In general, a mesh with good AR will probably converge faster. Elements with bad shape will also have impact to analysis accuracy.
However, different solvers may respond differently to mesh changes. It depends on implementations.
The default values for mesh size are basically calculated based on model dimensions.
Many users do not use default values.
We recommend users to look at models/meshes closely and determine mesh size and other control parameters carefully.
Default parameters do not suit all models.