[sandy_1982]

hi friends ,

1. i have a centrifugal pump which is immersed in a tank that is full of water
the pump has two inlets and both these join to form one outlet which is the delivery from the tank

2. you can imagine the pump as a small vertical cylinder having inlet at its top
and bottom faces , the outlet simply comes out as a pipeline from its
curved lateral surface. This whole lies at the bottom of a full tank.


is there any way by which we can model the centrifugal pump in CFX pre without having to bother about its actual impeller , volute etc and doing it as a rotating domain - stationary domain problem?

i am trying to use the 'additional interface' models [ giving mass flow rate at each ]

any ideas/ experience on this kind of problem ..?

thanks & regards

Sandeep

[ghorrocks]

You can do this several ways.

* You can replace with pump with an interface and apply a mass flow rate condition on the interface.
* You can use a source term to define a flow rate by setting the velocity to a defined value.
* You can use an inlet/outlet pair where you have defined the pressure on the pump inlet and the flow rate on the pump outlet.

[sandy_1982]

thanks glenn ,i'll work on these

[sandy_1982]

Dear Friends

1. please find an attached .ppt which shows what i have done so far ..

2.i have subtracted the pump body from the tank fluid domain which has resulted in a cavity in the fluid domain.

3. i need to give 'mass flow' interfaces (3.5 kg/sec at each) at the two indicated locations , how to select these surfaces which lie 'within' the bulk fluid ?

4. also i notice that the body (tank fluid domain after removing the pump body) does not get highlighted (means it doesn't show the green mesh) when i place the cursor on it ..why so ?

5. i am getting the following error when i run the CFX solver
"Return Value C0000005>FFFF in parse_rv. Treating as non-run"

Please answer the above specifically.

ALTERNATELY i was thinking of just creating a pump not as a shell but only as 0 thickness surfaces within the bulk fluid and then proceeding ..it may offer ease of selection...how is this approach ?

Thanks

Sandeep

[ghorrocks]

The error message is I think due to something funny in the setup and CCL. I would run your simulation through CFX-Pre to set it up again and make sure it is set up correctly. You can get errors like this when you try to define an interface or boundary in CCL only - you cannot do this, you need to setup boundaries and interfaces in CFX-Pre as it does something to the mesh face labels as well.

Have you looked into using interfaces to do the mass flow rate? That might be a better way than thin surfaces.

[sandy_1982]

glenn ,

i have been able to set up and run the simulation correctly , of course the results are not what i expected (still working on these)- but at least i cleared the initial hurdle to have the mass flow interface boundaries defined -i have created a body within a body in DM and then in this multi-body part , i have defined the interfaces as 'mass flow' and 'no slip' respectively.

shall update on this hopefully soon.

thanks

sandeep