Modeling a thin plate elastic leaf spring in Elmer

[foadsf]

Following the discussions on the unofficial Elmer Discord server, I was wondering if there are any examples for linear elastic leaf springs? consider the geometry below:



I want the end of the three legs to be fixed and a vertical/normal force to be applied in the middle. I have modeled the geometry both in 3D and 2D as can be seen in the attachment. However, I'm not getting the expected result. I would appreciate it if you could help me know what is the best way to model simulate this in Elmer. Thanks for your support in advance.

[foadsf]

I followed one of the examples provided in another post, replaced the Plate solver with Shell Solver but still doesn't work.

[foadsf]

Not being able to find the right configuration for the Plate solver I switched back to the linear elasticity stress solver. Peter was very kind to provide me with an exampe .sif file yester on Discord. Though I had to comment the line

Code: Select all

Linear System Use Hypre = Logical True

line out in order to avoid the

Code: Select all

ERROR:: CheckLinearSolverOptions: Hypre not usable in serial!

I'm still not able to get proper results, failing at

Code: Select all

ERROR:: IterSolve: Numerical Error: Too many iterations were needed.

sadly I'm running out of time.

[raback]

Hi

When you avoid the error then you also don't get to use Hypre which is needed for the BoomerAMG.

-Peter

[foadsf]

So what is your suggestion. What is the best way to get the simulation I need in Elmer:

• should I use Linear Elasticity and the 3D geometry or Plate Solver or Shell Solver and the 2D models?
• what should be the configuration of my .sif file given the fact that I have a portable version of Elmer installed following this video, no MPI and no Hypre

[kevinarden]

So that we can run some tests what is the expected result? And can you provide the mesh files instead of the geo?

[foadsf]

So that we can run some tests what is the expected result? And can you provide the mesh files instead of the geo?

The simulation is actually very simple

• a distributed force of 10N is applied at the circle in the middle
• circle in the middle should be rigid

I can't attach the mesh files as they are too big.

I used Gmsh to convert the .geo files to .unv:

Code: Select all

gmsh.exe <fileName>.geo -3 -order 1 -format unv

One can use ElmerGrid to generate the mesh next:

Code: Select all

ElmerGrid.exe 8 2 <fileName>.unv -autoclean

or open the .unv file directly from ElmerGui

more info here.

[kevinarden]

Elmer can solve this in 2D shellsolver or 3d stresssolver. However, Elmer is sensitize to the type of elements and element mesh density used. From the geo I can't tell. I could do my own mesh, but interpreting the geometry from the geo is not straight forward. Is there a step file available? Also is the expected displacement known?

When I wrote a unv file and imported it into Salome there were numerous coincident nodes, did you do a coherence in Gmsh?

[foadsf]

Here in the attachment you may find the STP/STEP file and the FreeCAD file

[foadsf]

Also is the expected displacement known?

The theoretical vertical stiffness of the leaf spring is "3 * E * T^3 * W / L" assuming E = 200GPa, T = 0.15mm, W = 4mm, L = 9mm, a vertical deflection of almost 3mm is expected.

When I wrote a unv file and imported it into Salome there were numerous coincident nodes, did you do a coherence in Gmsh?

I don't have SALOME at the moment, neither I'm familiar with Gmsh coherence checking. Can you be kind to educate how that should be done?