Pump CFD Analysis

[Jamino87]

Hi All

I've recently been trying to model one of my company's pumps so that i can later use CFD to predict the performance of new designs.
My results seem to be quite close at higher flows which I'm happy with, but at high pressures (low flows) my pressure is significantly down on the measured value.
What would be causing this?

I'm using SolidWorks Flow Simulation, which i know isn't the best, but it's all I've got for now.
Any help would be much appreciated.

Cheers

 

[LittleInch]

Are those numbers for real?

3.6 m3/hr at 2000 bar? (200000 kPa)

At that sort of pressure and size any small error will become important and also you can get some strange effects which become negligible as you go higher

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[rotw]

At that range of pressures, I think you can cut metal sheets...
sorry for the off topic.

 

[btrueblood]

Ugh. Missed the k...guess that's not a centrifugal pump...

 

[Cecorros]

At and around BEP the flow is symmetric through impeller walls and volute channels. As you most likely know, at part-load conditions you get separation of the flow and recirculation both at discharge and inlet of the impeller. It might be that the CFD model that you are using is not picking up the amount of energy induced from recirculation to get the head/pressure that you need in order to get appropriate results and validate your model.

Are you able to post plots/contours of streamline velocities and pressures from your CFD calculations. Also if you could give some details about the pump: type, size, impeller diameter, etc.

 

[Jamino87]

Hi All
You are all correct, that pressure is far too high. The units are actually Pa, but I wrote kPa on the axis.
It's a fairly simple centrifical pump.
I've started using a transient analysis, which is giving better (higher) pressure, but still not accurate enough. I'm getting 182 kPa, chasing 195-200 kPa.
There seems to be recirculation at the outlet, but not at the inlet.
Please excuse my ignorance, but how does recirculation increase the pressure? I thought it would decrease it.
Thanks for the input.

 

[Cecorros]

Recirculation acts as a hydraulic excitation force into the fluid adding energy in order to achieve a stable Q-H curve. There are articles and discussions on the internet about it.

You should to get certain amount of recirculation at the inlet of the impeller whilst at part-load conditions - as noticeable as the one you are getting at the outlet.

What turbulence model are you using? Are you specifying a value of the +y? How many nodes do you have for the impeller, diffuser and collector?

 

[btrueblood]

Not completely familiar with the SW flow simulation, but a common problem for CFD analysis is how it handles rotating flows through machinery (generally poorly, when you need to incorporate fixed and rotating parts). In your plots, it is fairly obvious that there is no rotation of the flow relative to the outer stationary vanes, and you are seeing funny flow reversals at the leading inboard edges of the vanes.