Structured vs. Swept mesh tecnique

[blckwtr]

I have a long cylindrical shaft exposed to axial shocks. There has been fatigue damages to certain parts, now for investigation. Cross sections are not homogenuos through the whole length, it varies a lot.

Question(s):

Do I use swept mesh technique as far as possible, or do I use structured mesh on transitions from one cross section to another, or do I just use structured mesh technique?

For results I would prefer the first...

 

[brep]

This is going to depend totally on the topology of your model, so providing some more info (or a sat file) of the geometry will be necessary.

That being said, assuming you are using V6.6, the sweep meshing technique can be used on variable cross section geometry (e.g. lofted geometry etc)

Also, you can now use virtual topology to combine faces to create a sweepable region.

Can you provide some more detail?

 

[blckwtr]

I could, but on a more general level:

Do you get better results from a structured mesh on a part that consists of a variable revolved geometry (axisymmetric), or would you say a swept mesh technique provides the best results?

The cross section increase on one point from 16mm to 32mm and then to 16mm again. This part is almost 1m long, the "larger" section on the middle is about 30mm long.

I am interested in the stress concentration in the "groove" if you could say so...

 

[brep]

As long as the elements have acceptable shape (meeting the ABAQUS checks) then the analysis will not care how they were created. Of course the analysis code does not even know what technique was used to create the mesh (sweep, structured, manual, imported...)