CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Multiphase Particle size and velocity settings (

zmaxcomputer August 18, 2010 20:24

Multiphase Particle size and velocity settings
4 Attachment(s)
Hello! I am in the process of learning CFX-Pre for part of my grad- project, and i'm trying to make my professor happy. Anyone that can help me i would greatly appreciate! I have attached a few screen-shots (hopefully that will help).

I have a spiral vessel that i have made and meshed. My goal is to have 3 different sized particles flow through the pipe at a constant rate, starting at the inlet (approx 25um, 2um, and 14um). The density of the particles is slightly higher than water.

However, i find there are different places where i can set the size of these particles. "Inlet", and "default domain default".
1)How can i be assured that the particles are always the same size?
2)How can i be assured that the speed of the particles is the same as my fluid flow? There are a couple places where i can edit the speed of the fluid; Inlet and then in "Default Domain Default."
3) I do not want particles just "appearing" inside of my vessel. I would like to track just a few from the beginning of the vessel to the end, But the # of streamlines seems to increase further down the vessel.

Thank you for all of your time!

Take care


Claudia August 19, 2010 03:29

Do you want to make a simulation with lagrangian particle tracking?
Then I would set up a particle phase with a discrete diameter distribution with three diameters and the corresponding mass fraction and a representative number of particles (belongs to your particle mass flow).
Let they start at inlet with the same velocity as your fluid.


zmaxcomputer August 19, 2010 03:34

Hi Claudia, thanks for your response!

I am aiming for a lagrangian approach. And i have tried to make all of my inlet velocities the same, and i also made 3 different materials with different diameters. But i find myself able to change these diamaters in 2 or 3 different parts of the setup (inlet & default domain). The tutorial (#11, flow through a butterfly valve) says nothing about setting the diameter in the inlet, which can be done and i am not certain what i should do with this setting. Even if i follow the tutorial i still get streamlines originating in the middle of my vessel.

Claudia August 19, 2010 04:41

If you change the diameter at inlet, I think it overwrites the settings in the default domain. Why do you chose only 1 particle and a very low mass flow?
When you start your particle at inlet they should appear first time at inlet.

jbritton August 19, 2010 05:03

Dont know if you can get hold of the ANSYS multiphase training workshops, but there is one thats for Langarian particle tracking in a 2-D (well 1 element thick) U-bend. Uses 3 different particle sizes in it, sounds like exactly what u are trying to do.

zmaxcomputer August 19, 2010 05:58

@ Claudia

Yea, i was told that once before, that at the inlet it over-rides the other setting, but i have not been able to confirm this and a professor i knows said that he doesn't see why that would be the case. And this doesn't solve the problem i have with particles appearing in the middle of my vessel. I made a very slow mass flow rate b/c i can not seem to be able to plot just 10 particles, or 5, or any number i please. That is the frustrating thing....i would like a basic simulation of 1 particle of each size starting at the center of the inlet, then increase the number of particles from 1, to whatever, and see how the different particles are affected by the spiral. But if i can select the number of particle "positions" why do i need to bother with the mass flow rate? Thing, is, if i leave the mass flow rate at zero, CFX says i have to put a value in. So once again, i don't know what overrides what, particle positions or mass flow rate. Also, i can edit the mass flow rate in two places as well, the inlet and the default domain.

@ jbritton

I do not have those workshops, but if you have that tutorial, i would love to take a look at it, maybe it can help with these issues im having? If it doesn't have a clear cut answer though regarding these numerous settings that seem to be redundant it will be hard for me to convince my prof that my results are reliable (sad but true.....), but i would like to see if it can help!

Claudia August 19, 2010 06:44

You defined your particle in two regions: the first one is the inlet, there you want to have your particles get into the model. But you also defined the default domain as particle inlet, too. Unable the check box "define particle behaviour" in the default domain and define it only at inlet. This might help.

zmaxcomputer August 19, 2010 13:46


That did work for getting the particles to begin at 1 face! But ive tried all the different "particle positions" and i cant get my particle to begin at the middle of the vessel. I tried uniform injection with equally spaced particle locations, annulus, injection at face centres, injection at ip face centers and solver will complete the run, then "terminate with errors". Is there another way to define where my particles begin?

thanks again!

Claudia August 20, 2010 03:11

Now you confused me. You want, that the particle start in the middle of the vessel? Or do you want the particle to start at inlet?

zmaxcomputer August 20, 2010 04:34

@ Claudia

Haha! im really sorry i confused you......both actually. My goal is to have the multiphase demonstrate that particles that start at the same place (at the inlet & at the same place) will behave differently due to the spiral. There has got to be a way for the particles to "start" at the very center of the vessel. Or something like 10 particles within a 1mm radius of the vessel. Then 10 particles within 1.2mm of the radius. That way i can see the tipping point of what it takes for some particles to hit the wall of the vessel while others go straight through.

Thanks again! you have been a great help so far!


Claudia August 20, 2010 05:03

Try to use a particle injection region with a full cone and 0 degree cone angle. You have to specifiy the injection place (coordinates), the injection direction (z-coordinate = -1 in your case) and the number of particles and the mass flow.

zmaxcomputer August 20, 2010 19:00

i do not see the "cone" option in the inlet. Under Particle position i see 1) uniform injection 2) uniform injection with annulus 2) injection with line weighting 3) injection with point weighting 4) injection with circular weighting 5) injection with user defined weighting 6) injection at face centres 7) injection at IP face centres

will one of these work? and if i start playing with the mass flow rate again, can i still select "1 particle, 2 particle or 3 etc?" when i start playing with the mass flow rate that's when i get particles appearing in the middle of my vessel.

thanks again!

Claudia August 21, 2010 15:29

The cone is not an option in the inlet. On the tab "default domain" you have a tab "particle injection region". There you can define a new injection region e.g. named "test" and then you can define the location and so on.
For the relation between mass flow and number of particles please have a look in the CFX help for "Particle Number Rate". This is essentiel for understanding lagrangian particle tracking.

zmaxcomputer August 23, 2010 17:10

Hi Claudia

is it possible to use the "uniform injection within annulus" at the inlet boundary? Change the inner radius to a very small value? I am not sure exactly how to tweek with this option. I have tried to use the cone option but im having some difficulties, but i will keep on trying.

thanks again!

Claudia August 24, 2010 04:15

What is wrong with your cone option? I tried it before I wrote it and in my opinion it was a good way.
Did you try the uniform injection with annulus? Then you will find the answer if it is a good way for your case.

zmaxcomputer August 24, 2010 04:21


If you tried the Cone option and it worked, than i will definitely go with that. I must be doing something wrong. Do you mind guiding me through how you did it? I am new to this stuff. I can not figure out how to make my injection circle be 2mm, 2.1mm etc, and the guides i have read don't seem to be helping much.

Thanks again for your time, you've bee a great help!

zmaxcomputer August 26, 2010 17:35


i kinda got the cone to work.......but all my particles enter the vessel in a line....not a ring. Even the ring cone didn't make my particles look like a doughnut......what am i missing?

Thanks again!

Claudia August 27, 2010 03:36

I don't know, what is wrong. Could you post a picture from your inlet conditions and the tab where you specified the cone?

zmaxcomputer August 27, 2010 10:04

4 Attachment(s)
Hi Claudia

of course.... i have the setup tab and the CFX-post results. I did it for both the full cone and the ring cone.

the specified diameter is 24microns, and the mass flow rate is .0035 g/s in all the set ups.

All times are GMT -4. The time now is 10:58.