Divergent results on simple CFD simulation

I’m just trying to get a stable simulation running with very simple conditions, but the results keep diverging and giving ridiculous flow rates. 20psi inlet, 19psi outlet, solve for flowrate. Incompressible analysis of water in a pipe elbow.
Integral values for the outlet flow rate (Ux) start increasing exponentially rather than stabilising.

The project is “Pipeflow test1”
I tried having the initial pressure condition as being halfway between the inlet and outlet pressure. I can’t see what else I need to alter to get a result.
Strange that the software still reports the simulation to be “successful”…
Any help is much appreciated.

Hey @cfulton,

Usually “fixed value” is not used for pressure inlet BCs. Have a look at common inlet/outlet BC combinations.

I took your exact set-up and changed the inlet BC from fixed pressure to total pressure:

Ux is still wiggling a little bit but it’s converging much faster:

You can try doing this with a finer mesh and these wiggles might be gone. Your mesh is quite coarse now (15k cells).

1 Like

Thanks! So what’s the difference between Fixed pressure and average/total pressure? Is it just static vs static plus dynamic?
I’d like to be able to do a simulation for a valve Cv test, which requires finding the flow rate of water with 1psi pressure drop across the valve/component.
I get quite different flow rates (2.4 vs 3.2) depending on whether it’s a fixed pressure or total pressure.
Also does the system place a wall function on all unassigned surfaces by default or not?

These 2 BCs behave quite differently. Fixed (static) pressure will maintain the same pressure for all cells in the surface.
For total pressure, in the other hand, you’ll get a “developed” flow:

Pressure values are not equal in the entire inlet patch:

I guess it doesn’t “constrain” the flow too much and allows it to settle better,

About the BCs: it’s a good practice to specify BCs for all boundaries. If left blank, it will be considered a no slip wall with wall function.

I can see the benefit of the developed flow, however “total pressure” includes dynamic pressure, so if I’m looking at simulating a 1psi drop in gauge pressure from inlet to outlet, that would mean 1psi drop in static pressure, correct? Total pressure BC means I need to account for velocity but velocity is the component I’m trying to solve for so it ends up offsetting the static pressure.
Outlet has the option of “average pressure” BC which I guess is average static pressure? So basically “developed” static pressure? But there’s no option for average static pressure for inlet.

I guess it’s worth trying total pressure inlet and outlet. Since Uinlet = Uoutlet, then the 1psi drop remains true.

Except the outlet options are limited to fixed or mean pressure, not total, so assuming “mean” is mean static, then the conundrum continues…

You can set total pressure outlet via custom BC