Disclaimer - this is my 2nd ever Ansys-Fluent simulation and I have not had any formal CFD training yet.
I am working on a simple ANSYS Fluent (16.2) multiphase flow simulation . The geometry is a larger straight tube with a smaller straight tube inside the larger one along the same axis with one end of each tube sharing the same plane.
I intend on having the the ring around the smaller tube defined as an inlet for one fluid type (Fluid 1) and the smaller tube an inlet for a 2nd fluid (Fluid 2) with the mixture coming out the far end. See attached model rendering.
I am at the point where I need to configure the Multiphase Model in Fluent and I am having difficulty identifying the correct model and parameters for the type of mixing model I hope to create. Fluid 1 will be a glycerol like fluid with a pulsatile flow up to 50cm/sec, and Fluid 2 will be a constant flow (or pulsatile flow) up to ~140 cm/sec of a homgeneous solution of specified concentration of an additive with a more syrup like viscosity. I am hoping to use the simulation to understand how Fluid 1 dilutes Fluid 2 farther down the pipe and how the net concentration varies over distance with the pulsatile mixing.
It appears from my searching I should use an Eulerian Multphase model based on page 8 of http://www.bakker.org/dartmouth06/engs150/14-multi.pdf and https://www.sharcnet.ca/Software/Fluent6/html/ug/node876.htm .
I can rule out enabling the boiling model and evaporation-condensation parameter options but I am unsure of whether to use Multi-Fluid VOF Model or Dense Discrete Phase Model in the Eulerian Parameters config window & If I do pick Multi-Fluid VOF how do I determine what to specify for the volume fraction cutoff and Courant Number?
Thanks in advance for any advice.
Erick
Hello Erick,
Can you provide some more information about the fluid properties of the two phases? Do you have any experimental data that you can use for validation of the simulations? As far as I understand your setup, it might be possible, and sufficient, to use a species transport model which is much less computational intensive than a two-phase model. The Eulerian MP models are typically used when high momentum exchange between the phases occurs, e.g. due to drag, particle-particle interaction, surface tension, etc. and when the phases are in different physical states (gas-liquid, gas-solids, liquid-solids etc.) are present in the system. However, if you don't have significant droplet formation or stratification in your flow, for example, the assumption of one continuous phase might be feasible. Using the species transport model, you still could calculate the different concentrations as a function of the pipe length and the pulsation.
In particular the VOF models assume a distinct interface between the phases, which is probably not the case when mixing two fluids, even if their density and viscosity is significantly different.
Anyways, if you still want to use the Euler MP model I recommend to leave the default VOF cutoff (further explanation is given here https://www.sharcnet.ca/Software/Ansys/16.2.3/en-us/help/flu_ug/flu_ug_vof_formulation.html). The Courant number depends on the flow velocity, the mesh size and the time step size (see https://en.wikipedia.org/wiki/Courant%E2%80%93Friedrichs%E2%80%93Lewy_condition), but mostly you can leave the default value, too.
Hope this helps for now.
Best regards,
Stephan
Not easily comparable experimental data to validate. I could rig up an experiment but I need to get MR scanner time to do so and then I could potentially get velocity vectors as a function of position or time resolved vectors in a cross sectional slice.
I am modeling a contrast media injection into blood in a mock blood vessel. The contrast media has a mass concentration associated with it *** mg/ml. There really isn't a droplet effect. the contrast media is a constant volumetric flow rate and the blood is a pulsatile flow. There is a very sizeable velocity difference at the inlets - on the order of 10 to 1 to 15 to1 faster flow coming out of injection inlet although both inlets have similar volumetric flow rates.
Hi Erick,
I thought more about a qualitative comparison as a first shot, even though a quantitative comparison is always better....
Anyways, based on your description, I'd suggest that you take a closer look into the species transport model in Fluent. I don't see the requirment of using the much more complex Eulerian MP model to simulate the mixing of a constrast media into the blood stream. The species transport model uses a single set of mass and momentum equations and introduces an additional equation for the mass fraction (e.g. for the contrast media). You would have to define the diffusion coefficient (just choose a suffieciently small value, e.g. 10E-12) and you have to set the material properties, like density, viscosity etc.. Don't forget to specify the two species for the mixture and use the mass-weighted mixing law for the density (not the default option 'ideal gas'). That should do the job.
Good luck.
Regards,
Stephan
I am planning on moving to a 2 way FSI after doing this rigid wall simulation. With that in mind would you still suggest avoiding the Eulerian MP?
- Badges
- Science topic