Which boundaries should be chosen to be included in bc_ids[]?

[zhangmuElias]

Which boundaries should be chosen to be included in bc_ids[]?

[keiyamamo]

Hi @zhangmuElias

I agree that the structure of biharmonic.py is confusing. My understanding is that you do not get any boundary terms with extrapolation_sub_type=“constrained_disp” and therefore you do not need bc_ids as in turtle_demo.py. bc_ids should only be given when you use extrapolation_sub_type == "constrained_disp_vel”.

Regarding extrapolation_sub_type == "constrained_disp_vel”, this boundary condition seems to have several constraints. It can only be applied to the rectangular domain with mesh motion only in the perpendicular direction of the fluid boundary. I’m not 100% sure but it seems like extrapolation_sub_type == "constrained_disp_vel” is specifically designed for TF_fsi case and special care must be taken if you want to apply such boundary conditions to your original problem. I myself have never used extrapolation_sub_type == "constrained_disp_vel” and thus I might misunderstand something here.

The biharmonic mesh model is explained in the following master thesis p.21, 23

https://www.duo.uio.no/handle/10852/60349
or
https://www.sciencedirect.com/science/article/pii/S0045794911000629?via%3Dihub

So, I highly recommend that you check both of them.

Best,
Kei

[zhangmuElias]

Dear @keiyamamo

Thanks for your answer and your recommended literatures!
I think you are right, and I have changed the extrapolation_sub_type="constrained_disp" as the turtledemo.py wrote. Les me see if it will run to the end time.
And do you think it made any differences when choosing different sub type ["constant", "small_constant", "volume", "volume_change", "constrained_disp"] ?

Best regards
Zhang

[keiyamamo]

Yes, they do make differences. These two literatures that I recommended both explored such a difference.

Best,
Kei