Aorta fluid simulation: Too many bad cells (6% - 7%)

[tomaso]

I am currently working on a fluid simulation using an Aorta mesh obtained through an MRT scan. Originally, the simulation ran with around 6% bad cells (Offset criterion). After fiddling with different parameters, mainly with the "Mesh extrusion" function, I managed to bring the total percentage of bad cells to around 1%. The problem however, is that I started getting the error message of cells having negative volumes. I changed the parameters of the mesh extrusion until the simulation ran again, but it returned to the original problem of having 7% bad cells in the offset criterion, now resulting in a fatal error due to extreme velocities.

Any help would be greatly appreciated. If .med files or any other info is required, don't hesitate to contact me privately.

Thanks a lot in Advance!

[Yvan Fournier]

Hello Tomaso,

Could you provide more details on where the extrusion occurs ? Also, If you are working with Pr Behbahani, you certainly have my contact, so you can send me an e-mail and we can try to organize a short call to suggest options.

For your sort of application (and meshes I have seen with your predecessors), the new STL-based immersed boundary options in code_saturne would probably be useful, and worth testing. I could send you an example setup for code_saturne v9.0-beta.

For the classical numerical scheme, you definitely need to avoid negative volumes, but you can live with some bad cells. Using some robustness oriented option,s such as activating gradient limiters or deactivating reconstruction at boundaries can be useful. And there are more settings in v9.0. Which version of the code are you using ?

Best regards,

Yvan

[tomaso]

Hello Yvan,

First of all, thank you very much for the response and the suggestions. We were working with version 8.9 however we just installed version 9.0-Beta and will be testing the commands you suggested.

I am in fact working under Prof. Behbahani and Mr. Horbach. We will be contacting you through Prof. Behbahani to arrange a short meeting for some advice, perhaps through a Webex call. If you have the time of course.

I was extruding the cells while using "all[]", which I assume extruded all cells of the mesh hence resulting in negative volumes. I tried searching for other commands to use here but couldn't find any clear info. For instance, if I could just extrude the cells surrounding the bad cells, I assume this would improve the mesh and not result in negative volumes. Im not sure if this is even possible, do you perhaps have any input or advice here?

As far as the other recommendations you suggested, we will be testing them in the coming days as they are part of version 9.0-beta. If any questions may arise then we will contact you if that is ok.

Thank you very much for your input and advice, we will be contacting you through Prof. Behbahani for a more direct contact.

Best regards,

Tomas Osorio

[Yvan Fournier]

Hello Tomaso,

Yes, do not hesitate to contact me for a meeting.

In any case, you probably want to extrude the mesh at inlets or outlets, bit not everywhere, as this would also thicken the geometry (unless this is what you want, but I assume in this case you would rather be adding a viscous layer, which uses an extrusion but also pushes the mesh inward).

You can also better control extrusion by doing them in several steps, but this is useful mainly near angles for meshes with well aligned sections.

As you can see in user examples, you can control extrusions better using group-based or geometric criteria, such as "plane[]" or "box[]" rather than "all[]".

Best regards,

Yvan