axi+embed bug
44 views
Skip to first unread message
David Pekker
Jul 30, 2026, 1:26:32 PM (4 days ago) Jul 30
to basilisk-fr
Hi everyone,
For me the bug does affect fluid velocity. I think that the bug report did not observe this effect because it was at low Reynolds number where viscosity dominates the flow.
Please find attached a minimum example showing the bug (put together with some help from the AIs). In this example I am simulating flow development at Re=25. The attached plot shows the centerline velocity as a function of position along a pipe. The centerline velocity should increase as the plug-like flow at the pipe entrance develops into Poiseuille flow. Without the bug fix, the default axi+embed has much shorter flow development length than the equivalent BID+axi calculation. With the fix BID and embed+axi agree well.
The fix in the attached example is identical to the one in the bug report.
I was comparing BID and embed boundaries for describing the flow development in a pipe and stumbled on a bug in the advection code. It is the same bug as reported here: https://basilisk.fr/sandbox/bugs/axi_embed_tracers.c
For me the bug does affect fluid velocity. I think that the bug report did not observe this effect because it was at low Reynolds number where viscosity dominates the flow.
Please find attached a minimum example showing the bug (put together with some help from the AIs). In this example I am simulating flow development at Re=25. The attached plot shows the centerline velocity as a function of position along a pipe. The centerline velocity should increase as the plug-like flow at the pipe entrance develops into Poiseuille flow. Without the bug fix, the default axi+embed has much shorter flow development length than the equivalent BID+axi calculation. With the fix BID and embed+axi agree well.
The fix in the attached example is identical to the one in the bug report.
Best,
David
Reply all
Reply to author
Forward
0 new messages