Two-way FSI in 2D

  • 213 Views
  • Last Post 11 January 2019
chennafea posted this 03 January 2019

Dear ANSYS community,

I (http://engweb.swan.ac.uk/~c.kadapa/home.html) am interested in learning the FSI capabilities of ANSYS towards assessing how well ANSYS resolves fluid-solid coupling in situations with the significant added-mass.

To get hands-on experience with the ANSYS software, I have tried a few Fluent and Transient structural simulations in 2D. However, I am not able to perform any two-way coupled FSI interaction in 2D, and I could not find any material or examples on internet on this. All of the examples I came across are for 3D models.

While I can learn to perform 3D simulations, they are not really helpful for my cause because of the amount of resources required for 3D models. So, I am wondering if ANSYS supports two-way FSI simulations in 2D.

 

Thanks,

Chenna

Order By: Standard | Newest | Votes
chennafea posted this 11 January 2019

@RD2016, I do not agree with the statement "There is no difference between a "3D" problem that is only one cell thickness and a 2D problem that is one cell thickness.". While the number of cells may be the same, the number of faces is certainly different, with the 3D mesh having more faces than the number of edges in a 2D mesh which certainly requires more computations in evaluating the solution. Such an approach not only takes more steps at startup but also requires more computational time for the time evolution of the solution field.

It is not possible to have a true 2D simulation while using a mesh which is inherently 3D. 3D mesh simply requires more computational effort because it has more vertices, more faces, more DOFs and so on...

RD2016 posted this 11 January 2019

Again, you can still perform FSI for purely 2D problems. There is no difference between a "3D" problem that is only one cell thickness and a 2D problem that is one cell thickness. Check out this post for more info (https://www.cfd-online.com/Forums/fluent/114039-2d-fsi.html). Furthermore, it will not be more computationally expensive, as the number of cells does not change. You can perform it, it will just take a few more steps at start up.

chennafea posted this 08 January 2019

Thank you for your response, @RD2016.

It is very disappointing to know that ANSYS does not support two-way FSI simulations for truly 2D problems. We have to think about whether to fake the 2D simulation or not since it is going to be computationally expensive compared to the 2D model.

 

RD2016 posted this 05 January 2019

No, and yes. This may seem confusing, but bear with me.

No, you cannot perform any type of FSI in 2D. The programs require a face to transmit the forces/displacements from one to the other. However, what is a 2D simulation but a 3D setup with only one cell/element in the normal direction?

To "fake" a 2D simulation, you need to extrude your geometry in the normal direction, and then ensure that there is only one cell/element in said direction. I recommend using the Sweep method in Ansys Mesher. From here, it's as simple as the video describes.

chennafea posted this 04 January 2019

Hello both. Thank you for your replies.

I didn't know that the student license does not support two-way FSI. Thanks for pointing it to me. I will ask the IT team for the ANSYS software with the research license.

This, however, does not answer my question. Does ANSYS support two-way FSI simulation in two dimensions, assuming that I have all the required licences?

kkanade posted this 04 January 2019

if you have research license, then you can check following. 

 

peteroznewman posted this 04 January 2019

Are you using the free Student license for ANSYS or a Research license purchased by your university?  The free Student license does not support 2-way FSI.

Close