Urgent problem! Appreciate all you help!! 3D Centrifugal Pump set up problems!
 

Hello,

I wrote this in another thread a while ago, without getting any help. My problem is really urgent, so I really appreciate all your help!! The problem is about setting up a 3D Centrifugal Pump in FLUENT.

When setting up the domain, I simply did a CAD drawing of the fluid inside the pump, with the impeller cut out from it. The mesh was generated using Meshing in ANSYS Workbench. When implementing the mesh into FLUENT it therefore only contains one cell zone.

In FLUENT i set this cell zone as a rotating frame with zero rotating speed, since I dont want the hole domain to rotate, only enable some parts of it to do so (Maybe it can be set to stationary wall?). The wall between the impeller and fluid is set to rotating wall with a rotational velocity of 2700 rpm, while the outer wall is set to rotating wall with zero absolute velocity. The inlet is set to velocity-inlet while the outlet is set to a pressure-outlet.

I've used default values of the rest of the settings.

My problem is that I get a pressure drop in the pump, instead of the expected rise in pressure. This would be the case if the impeller didn't rotate, so therefore I think there is some problems with my settings. On the other hand have I, with help of contour plots, seen that the impeller is rotating.

Can I solve the problem in this way, or do I need to split the domain into several parts? Or is there any other way of dealing with this kind of problem?

Thanks!

Best Regards,

R

You need to have 2 domains. One (moving) within the impeller, i.e the rotating fluid between the blades. And second is the volute or casing domain. You need to have a common interface between both these domains.

Regards
Luke

Hi Luke,

thanks for you answer! So you mean that if I split the domain into two parts, where the first one contains the impeller and the surrounding fluid and the second contains the volute, and then put a interior B.C. between them, the pressure will increase (and by that solve the problem)? Did I understand you right?

Best Regards,

R

You also need to specify your impeller and surrounding fluid zone as a moving reference frame.
You do that by checking "Frame Motion" under cell zone conditions and then specifying the origin and rotational velocity.
Your moving walls (impeller) should also be specified as moving walls with a rotational velocity.
For BC's use a mass flow rate inlet.

Regards
Luke

Hi again Luke,

now I have tried to solve the problem by splitting the domain into two parts, one containing the impeller and the fluid between the blades, and one containing rest of the fluid (from inlet to impeller, impeller to outlet). I got two separate cell zone when implementing this mesh into FLUENT, and I set the one containing the impeller as rotating by enable frame motion and with a specifik rotational axis and speed (2700 rpm). I let the other one just be stationary.

When dealing with boundary conditions I define the inlet as a mass-flow inlet (~1 kg/s) and the outlet as a pressure outlet, as before. Since I splitted the domain I also need to define boundary conditions between them. The wall of the domain containing the impeller belongs to the cell zone named "impeller" and therefore I defined it as moving wall, rotating relative to adjacent cell zone. I define the wall belonging to the other cell zone as moving wall, rotating with a absolute velocity 0.

When I run the simulation I see that the two cell zones operates as seperate parts, and doesn't come in contact with each other, so there must be some problem with the boundary conditions I set between them.
And as before I get a much higher pressure at the inlet than the outlet (probably due to the fact that the fluid flows straight to the outlet from the inlet).



Thanks!

Best Regards,

R

The common face between the rotating and stationary zones should be an interface.
And the walls of the stationary zone should not be defined as moving walls.

Regards
Luke