Chained CFD and FEA analysis

Using SimScale
1

Hi to all SimScalers!

I am willing to perform a chained simulation, first CFD to calculate pressure distribution on a surface and then use it as a load for a FEA stress analysis.

My question is, what is in your opinion a good (at least feasible) way to perform this? Is it even possible to perform using the platform? Is it even possible at all using open source tools?

Thanks for your time, and will be waiting for your contributions!

2

Hi Guillermo! :slight_smile:

Short version: A first idea would be to use the result control items. If you want accurate results then use Paraview instead to get the forces on the entire face and then do a 3D force definition using a table upload for the FE simulation.

An important thing to notice: One should run a transient flow case and not steady state because in steady state analysis you will not get the full time period thus not a pressure distribution over time - in the other case it is just first and last time step. So performing a transient simulation will allow you to extract the pressure results with the relevant locations!

Additionally we have mainly 2 types of forces in CFD, namely pressure force and viscous force where the pressure force is calculated as the pressure normals integrated over the given area and the viscous is calculated as the shear stress times the area! The area in our case would be the surface selected and all points/cells of this area are taken. But if you want the forces particular areas then you should either have those regions as separate faces or patches in the model or using filters in the post-processing to clip the full face into parts and then calculating the forces for these parts in (offline) Paraview.

@Get_Barried, @vgon_alves & @1318980, feel free to correct me if I missed something, mixed some things up or if you have a more efficient approach in solving such projects. I think that is good material for another Step-by-Step tutorial @ggiraldo! :sunglasses:

Best,

Jousef

4 Likes
3

Hi @jousefm,

Thanks for your answer, took my time to reply waiting for more contributions, but here go some comments:

I’ll give it a try, but I’m worried that if the surface mesh differs from the CFD and the FEA models, a translation of load through forces might be inaccurate, because it depends on face areas.

Also, does the force definition via table upload accepts 3 parameters? I tried it before with pressure and it wasn’t possible (according to support).

4

Hi @ggiraldo!

Just some spontaneous ideas for now (maybe I will have more sophisticated ideas later on :smiley: ):

  • Make sure that the mesh convergence study in the FEA part are separated meaning that you do a conv. study for each analysis (FEA & CFD) to ensure grid convergence.

  • You could use 2 parameters where one is let’s say the time and the other quantity pressure force x or viscous force y etc. using multiple BC for a particular face.

Any other ideas @vgon_alves, @1318980 & @Get_Barried?

Cheers!

Jousef

1 Like
5

Hi @ggiraldo @jousefm, I am afraid I can’t help much knowing little about the FEA side of it, but would be more than happy helping when you need to get the data from CFD to surface table form.

I suggest creating a really basic mockup, say a finite cylinder in free stream, running a CFD on it and then attempt to run the FEA on the solid part with the topological pressure.

Sorry I can’t be of any more use.

Darren

2 Likes