CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   CFD Analysis. (

Elena Rodríguez February 27, 2001 04:32

CFD Analysis.
Hello, My name is Elena Rodríguez and actually I'm analysing a Centrifugal pump which has an impeller with 3 channels. I use de code Unigrahics for de Geometry, GAMBIT to obtain a good quality mesh an FLUENT to solve the case. In all de analyses I've made before, I've have obtained good results with this code. In this case, the pump has axial aspiration and radial impulsion, and the impeller consists of only 3 channels. For every Q I've tried to solve (BEP, Q>BEP ) the results are almost 10 MCA smaller than the results We have obtained in the tests. I've no answers for my problem. Could be a problem the number of channels? (Only 3)

Thanks in advance


John C. Chien February 27, 2001 16:58

Re: CFD Analysis.
(1). All I can say is trust your test results. (2). Centrifugal pump is a very difficult subject. With 3 blades, you are going to have a lot of flow separations in the pump. (3). You are lucky (with test results), using a black box blindly can be very frustrating.

Elena Rodríguez February 28, 2001 04:51

First of all thank you very much. But, is there any manner of studying a centrifugal pump with only 3 or less channels. Any other model or turbulence (instead of K-e-Realisable).

In my case, Furthermore, I have to study mixer pumps, which have, always, only 1/2/3 channels. Seeing what you have said, this will be almost impossible with FLUENT, Isn't it?

Is this sort of analysis impossible with the code FLUENT? Is there any other code, which I could analyse these pumps with?



John C. Chien February 28, 2001 07:53

(1). I had modelled a three blades pump impeller (just one passage only) without the presence of the volute. For that, I am sure that the code can be used to simulate the passage flow field. I didn't have the time to study it systematically. So, the accuracy of the result was unknown. (2).My only suggestion is: improve the mesh first, and see whether the result can be improved or not. I think, the inlet pipe and the volute must be included in the simulation because these components have rather strong effect on the performance. Then there are leakage between the impeller blade and the shroud wall. So, the whole configuration is rather complicated. (3). The turbulence modeling in this case is just too complicated to give any reliable result. I must say that it is the state of the art. ( the pump companies usually are rather limited in resources to do any advanced turbulence modeling. The return in investment is also not justified, based on my experience in early 90's when I had a chance to touch this field.)

patxi March 2, 2001 06:39

Re: CFD Analysis.
Following geometry modifications to better represent your experimental geometry, is this issue now resolved ?

Jack Keays March 12, 2001 08:09

Re: CFD Analysis.

I am using FLuent to model a Centrifugal wastewater pump. This pump has only 1 blade and obviously a single channel. I am modelling it in 2D and a simplfied 3D model. I model the pump at various heads/flow rates and plot hte power versus flow rate characteristic curve. II expect the power required to increase with flow rate in the range of interest. HOwever, I am getting the exact opposite. Anyone have any ideas? Do any of you have papers relating the pumps with low numbers of blades? Thank for your help...I think it owuld be nice if a few of us who are modelling this type of pump (low number of blades) could get a discussion going.

All times are GMT -4. The time now is 09:07.