Prescribed velocities boundary

[Dylan van der Schoot]

Hi all,

When I'm applying a steady parabolic flow on the inlet, I see that for some reason in the calculation of this profile not only a velocity in the required direction is present, but there is also a small parabolic velocity profile in the other two directions (It is in the order of 10^-4 of the velocity in the normal direction). Anybody has an idea why this is? For fully developed flow the velocities in the other directions should ofcourse be zero.

Thanks in advance!

[Weiguang Yang]

The inlet profile prescribed by bct.dat file should give zero velocity in the non-axial direction. In the domain, probably you won't get 0.0 exactly. 1e-4 doesn't seem too bad. If you want to further reduce it, you can consider extending the pipe. L~10D may not be sufficient.

[Dylan van der Schoot]

Hi,

Thanks for your reply. I've checked the bct.dat file and the prescribed velocities in the non-axial direction are not exactly zero. I've attached the file here: https://we.tl/t-zVyxgqr3Oq
Any idea what causes this?

[Dylan van der Schoot]

I found the problem. For some reason, if I apply inflation layers in my mesh, the mapping of the velocity profile on the inlet boundary introduces non-axial velocity components (approx. 10^3 times smaller than the velocity in the axial direction). When I'm not using inflation layers in my mesh, this non-axial velocities are approx. 0 (in the order of 10^-12 to 10^-20, but still not completely 0). This is for both parabolic and womersley velocity profiles (I didn't check plug). Is this a bug in the creation of the bct.dat file?

[David Parker]

Hi Dylan,

I think that the velocity profile mapping should be the same wether you are using boundary layers or not.

Please file a bug for this on GitHub (https://github.com/SimVascular/SimVascular/issues) and I'll have a look. Be sure to include any data files (mesh, bct.dat, etc) for both cases.

Thanks,
Dave

[Dylan van der Schoot]

Hi David,

Thanks for your reply, I've filed the bug on the GitHub.

Thanks in advance for the help,

Dylan