[flow-mo-mcfadden]

Here's my situation. I'm foolishly learning how to use a tool for which I have little understanding of the underlying mechanics and even less knowledge of how to work towards a solution. I have ideas that don't match the options in front of me but don't worry, its not a critical application. Your help is much appreciated!

That being said, I'm using ANSYS Workbench with Fluent v19.0 to assess flow through a pipe.
The pipe has compressed gas at the inlet and a positive displacement pump at the outlet.
I set the inlet BC as "Pressure Inlet" to model the compressed gas
I'm struggling to set the outlet BC to model the positive displacement pump.

I would like to see a "Volumetric Flow Rate-Outlet" but alas, its not there and maybe for good reason that someone could describe in simple terms.

Any ideas? Seems like a simple enough problem.

[LuckyTran]

In CFD-land, volumetric flow rate outlet means velocity outlet. You can't specfy a velocity outlet (normally) because that would imply some wonky physics. But you can use a velocity inlet and turn it around to have the flow go out.


Btw, there is a targeted mass-flow rate option in the the pressure outlet BC.

[flow-mo-mcfadden]

Originally Posted by LuckyTran

In CFD-land, volumetric flow rate outlet means velocity outlet. You can't specfy a velocity outlet (normally) because that would imply some wonky physics. But you can use a velocity inlet and turn it around to have the flow go out.


Btw, there is a targeted mass-flow rate option in the the pressure outlet BC.

I failed in my attempt use a "reversed velocity inlet at the outlet". Here were my steps

1. Axis-symmetric steady state using a pressure outlet.
2. Measured the average velocity at the output.
3. Changed the outlet BC to velocity-inlet and selected magnitude and direction
4. Input the measured average velocity as magnitude and -1 for the axial direction. The flow should be positive X direction which the outlet's facing direction.
5. Run and within 5 iterations I got a floating point error from divergence.

Sounds like this is an unconventional approach if I can get it to work. Any more tips? Here's another wrinkle. The pipe outlet has a diverging section before the positive displacement pump and I want to do a little optimization out the outlet diameter but I think this element makes the problem more proceduraly-intensive to solve because I don't know how to include the measured velocity into a design of experiments. Maybe there is a way to turn a measured output into an input but that's beyond me. Still looking for the golden solution like a little pump widget to attach to my pipe outlet.

I'm not sure how to use the targeted mass flow properly if the fluid's density changes for different inlet feed pressures and outlet pump speeds I don't think I could get what I'm after which is fluid volume subtraction from from the pipe independent of density.

[LuckyTran]

Originally Posted by flow-mo-mcfadden

4. Input the measured average velocity as magnitude and -1 for the axial direction. The flow should be positive X direction which the outlet's facing direction.

If flow should be in the positive x-direction then your direction vector should be (1 0 0). Why you have a -1? Should have only 1's and 0's.


Keep playing with it. Instant divergence means you have a problem with BC setup and/or initial condition. Since you use an already converged solution, then it's quite clear it's a BC issue. Even if the BC is not the one you want for the situation you are imagining in your mind, you should be able to get the result for a given set of BC's.


Also, the physical implication of a known volumetric flow rate at an outlet implies that the flow at the outlet is supersonic... Hence my statement about the wonky physics.

[flow-mo-mcfadden]

I'm still trying to build a Fluent case to study the system shown in the figure but I can't figure out how to establish a boundary condition that represents the operating principle of the positive displacement pump. To be clear my model includes a pressurized steam inlet and a pressurized air flow inlet. I know the outlet diameter so it was suggested that I use a velocity inlet BC at the outlet and reverse the flow direction. I tried this without success.

In theory, a 100% efficient positive displacement pump in this application will evacuate the mixing chamber at a fixed volumetric rate. While the volumetric flow rate is fixed the mass-flow adjusts based upon the density of the fluid in the mixing chamber.

This seems like it would be trivial to solve. Any more tips on an outlet boundary conditions? I would really like to figure out how to get this to work.