Static Study Parameterization

Hi all

I am on a static simulation of an assembly composed of a fairly large number of sheet metal and volume parts in alloy 5754 and stainless steel.

The curved mesh is carried out without any problems.

On the other hand, the study is excessively long, in particular because of the function "more precise surface-to-surface contact" corresponding to the option of "more accurate solidarity contacts" which lasted more than 28 hours.

There is an option "Incompatible solidarity contact" which is by default on automatic, if I change it to simplified it makes the study crash.

Would you know which parameters to modify in order to shorten the duration, knowing that I have to do the 3 axes in 2 directions and that currently I have modifications to make because 2 points are out of yield strength.

Here are the parameters of the mesh:

Mesh Type Mixed    Mesh
Mesh Used    Curvature-Based Mesh
Jacobian    Points 4 points
Jacobian Verification for the Activated Hull    
Mesh    Control Set
Max    element size 150 mm
Min    element size 30 mm
High Mesh    Quality
Total Nodes    2082204
Total number of 1144191 items    
Remeshing failed parts with incompatible    mesh On
Mesh creation time (hh:mm:ss)    00:12:45
 

Study parameters:

Study    Name Static Analysis 1 (-Default-)
Mesh Type Mixed    Mesh
Thermal    Effects Include Thermal Loads
Zero     strain temperature 298Kelvin
FFEP Solver    TypeRead
Stress Stiffening    Off
Low stiffness    Off
Inertial    Relaxation Off
Friction    Off
Adaptive    Method Off

 

Thank you.

 

 

 

Hello

If you play on the quality of your mesh, putting a "coarse" mesh would change anything?

1 Like

There I was on an intermediate mesh, I'm going to try with a coarser one.

I tried another axis with a coarser mesh and it's much faster.

The difference is huge on this kind of assembly.

Curvature-based mesh sometimes causes a few crashes. Is it really useful in your case?

 

The addition of contact between assembly and component significantly lengthens the calculation time, but sometimes it is impossible to do otherwise.

 

Have you unused the mesh controls? We can define a fairly coarse mesh size for the whole system, and generate denser local meshes in the areas of interest. This makes it possible to optimize the number of elements and thus the calculation time.

 

 

Curvature-based meshing seems preferable for assemblies.

I didn't use the mesh control on the 2 calculations, just changed the density of the mesh on the whole.

The calculation has been reduced from about 30 hours to 8 minutes.

 

A high time to solve means that there are a large number of equations to solve.

We can reduce all this by reducing the number of elements. To do this, some tips below:

- Decrease the size of the overall mesh

- Use mesh controls to densify elements only in areas of interest.

- Curvature-based meshing is only useful for cylinders or other curves (the mesher adds a minimum and predefined number of nodes around the curved shapes.

- Use "shell" elements for sheet metal and "beam" elements for profiles

- simplify your model as much as possible by using its axes or plane of symmetry.

 

Thank you for these clarifications.

I also went through the SolidWorks help.

The help specifies that you have to use the curvature-based mesh for the assemblies, by the way I tried the classic mesh and many elements don't work.

Otherwise I'm in shell mesh which is selected by default for sheet metal.

By playing on the quality of the mesh, the number of elements has increased from 1,144,191 to 117,172, which has a much greater influence than in monolithic.

 

 

 

Hello Bruno,

Can you put a preview image of your mesh?

Here is a view of the mesh:

I think it's good for this subject, I saved a lot of time by modifying the mesh, the file registration is still long but it is  unavoidable.

I'm going to open another topic because I'm restarting the simulation in more detail with remote masses and contacts by screws instead of a united assembly.

Thank you.