[jstan3]

Hi,

I am currently simulating flow in a pump.

The inlet i am using is a 'mass flow rate' inlet and the outlet i am using is a 'pressure outlet' and i require the total pressure information at the inlet and outlet. So i created surface monitors at the inlet and outlet surfaces.

However, when i am running the calculation, i realise that the inlet has a higher total pressure than the outlet, which doesnt really make much sense to me.

i'm still in the mids of learning this software. Can someone advise me on what went wrong?

thanks.

[stumpy]

A pump is supposed to raise the pressure.

[jstan3]

Please correct me if i am wrong. Shouldn't the correct case be: Higher pressure at outlet than inlet?

[macfly]

I can at least tell you that you inversed 'inlet' and 'outlet' in your thread title

[jstan3]

oh. my bad. i meant it the other way around.

Would really appreciate it if someone could give me some advice on this.

[flotus1]

If the mass flow rate you enforce at the inlet is higher than the flow rate of the pump at its current operating point you could actually get a pressure drop from the pump.
Your setup corresponds to a bigger pump feeding fluid through the smaller pump.
As a first remedy, you could decrease the mass flow rate at the inlet.

[jstan3]

i see. tyvm.

[stumpy]

check the torque on the blades too. A pump should be adding work to the fluid to raise its pressure. In your case it must be extracting work since you are getting a pressure drop. Like Alex said, this means your rotation rate and your flow rate are not consistent.
Just to check... you have a rotating domain around the blades right (and haven't just imposed a wall velocity)?

[jstan3]

i have created interfaces(which i converted to wall) around the rotating impeller, which are rotating at a speed of 2000rpm around the rotating impeller. On top of that, for the fluid component (wall) that is residing in the impeller, i have also set it to rotate at 2000rpm. This are the only components in my model that is rotating. thx.

[stumpy]

I didn't really follow that, but it doesn't sound correct. You need a rotating reference frame for your zone. The walls (blades) would be stationary in that zone, since their motion is relative to the zone.

[GM_XIII]

Are you monitoring total or static pressure? I assume your geometry ends in the exit of the blades or does it contain the diffusor?

[jstan3]

@stumpy: i didn't realise it could be done that way. I will give it a try. But i have 7 zones in my fluid model. So do you mind explaining in more detail which zone requires a reference frame? and also which zone should be stationary?

Zones:
-the back leakage gap of the centrifugal pump (the leakage from the outer volute back into the inner volute)
-the front leakage gap of the centrifugal pump (the leakage from the outer volute back into the inner volute)
-impeller
-inner volute
-outer volute
-inlet
-outlet

@GM_XIII: i am currently monitoring the total pressure. The geometry ends at the outlet after the fluid exits the pump, which is about 0.1m from the outer volute exit.

[ghost82]

Also make sure you are rotating the blades in the right direction for your rotating reference frame/dynamic mesh case; remember that fluent uses a right-handed coordinate system.

Daniele

[stumpy]

If you set a velocity on a wall boundary then it's only valid if the velocity vector is tangential to the boundary. In other words you can't "push" the flow using this method, but you can introduce a shear force. Any boundary that "pushes" the flow (i.e. has a velocity vector with a component in the normal direction) should be modeled by placing it in a rotating reference frame and setting the wall velocity to zero. So this means all your blades need to be in a rotating reference frame with stationary walls. The shroud wall in the rotating reference frame will need to be counter-rotating, and it must be a surface of revolution so that it only has tangential motion. Generally any components upstream or downstream of the blades with non-rotating components are placed into a stationary reference frame.

[jstan3]

i see. tyvm for the advice.