[s.petley]

Hi all,

I dont know whether this would be better placed in the general cfd forum. Anyway, I am simulating (fully 3D, single phase, transient rotor/stator) a regenerative liquid ring pump. The pump impeller is quite complex, made up of a number of chevron shape blade pockets, which the fluid recirculates around, so plotting a streamline it all looks pretty chaotic. Between the impeller and casing there is also a very small (5 cell thick) side channel clearance.

Basically, my model under-predicts the pressure head by as much as 200kPa at duty point, and under-prediction increases as the flow rate increases. This I am not too surprised about.

However, if you have any suggestions as to things to consider/look at to try and reduce this difference in the model. Or some ideas as to why this physically might be the case, it would be very much appreciated.

Kind regards,
Sean

[ghorrocks]

See the FAQ: https://www.cfd-online.com/Wiki/Ansy..._inaccurate.3F

It sounds like your pump has lots of recirculations and complex flow. These sort of flows are always harder to get correct. You will need to consider carefully your choice of turbulence model and whether a RANS style model can model this type of flow at all - you may need to consider DES, SAS or LES style models.