[kliff]

i am conducting cfd of a small submersible water pump (Model Resun king-2) in attempt to improve performance(flow rate, reliability) of the impeller. i have modeled the original pump on solidworks. currently i am trying to validate the model against the real pump. the pumps speed is fixed at 3000rpm nothing can be varied on the pump.. the pump currently operates at the bottom of a 3ft pond and is used for re circulation. i very the depth at which the pump operates to see how it affects flow and head..from conducting the real experiments, my cfd model produces somewhat accurate flow but not head.,, boundary condition used at the intlet is static pressure of atm + 3ft depth and at outlet discharge piping is only atmospheric pressure. can someone help me? i can provide all that i have done so far.

[Boris_M]

Hi Kliff,

I am not 100% sure about your model setup so here are a few questions to clarify it more for me:

1.
Do you only have your pump model or also any basin or long pipe with the actual water column it has to work against when pumping?
The reason I'm asking is that you are probably pumping the water 3ft high to get to the atmospheric pressure which, considering the weight of the water, increases the pressure the pump has to work against.
Maybe you can post a sketch of your experimental setup and where you applied what boundary condition.

2.
Do you use the averaging plane or sliding mesh approach?

3.
The setup of the rotating region is very important and a section view of the mesh vertically through the impeller would help me tell you if the rotating region is modelled correctly.

4.
This relates a little to the 2nd question. If you use the averaging plane method then two pressure definitions at inlet and outlet are not ideal if the inlet side and outlet side is completely cut-off from each other by the rotating region. The reason is that the boundary conditions in such a case should be flow rate at the inlet and then the interface between stationary and rotating region fluid region has a pressure outlet from stationary and flow rate inlet in the rotating region. If you specify pressure at the inlet, then there will be also a pressure boundary condition at the outlet of the stationary inlet side and this can cause fluctuations and take very long to converge. This type of interface boundary condition is automatically applied within FloEFD. Often the solver stops at the maximum travels first and if the user does not consider that, then the values might still not be converged. It is better to define the flow rate at the inlet in the averaging plane approach.
I know it is not the ideal case but a possible better rotating region design could fix this. But on the other hand, you want to see if the flow rate matches the pressure difference or the other way around. If you specify the inlet flow rate from your test and the outlet pressure, then the inlet pressure will be calculated and you will see the pressure difference and if it matches. Considering you want to record the pump curve, then it doesn't matter if you specify the pressure difference and get the flow rate or if you specify the flow rate and get the pressure difference.
The sliding mesh approach is not that sensitive to this type of pressure-pressure boundary condition method.

Regards,
Boris

[kliff]

Good day Boris, can you provide and email contact and will provide you with all the information i have. thanks. it would be much appreciated