CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   DPM Injection - Nothing happens (

ChesterP July 31, 2010 21:47

DPM Injection - Nothing happens

I have defined a plain orifice atomizer injection with liquid at some crank angle. When I am running the simulation nothing happens, nothing is reported.

Have I missed something?

spring August 2, 2010 07:25

Hi ChesterP,

can you please tell me what you did before starting the dpm simulation, so that it would be easier for me to tell you if you missed something. Did you simulate the continuous phase before you displayed the particle-tracks?


Wikie August 2, 2010 10:55

this also happened in my chase for the first time. in my chase I forgot to turn on the gravity. if it's necssery for your setup turn it on ;-)

roghi.sedghi October 5, 2010 08:33

I have such a problem too. I have simulated the continues phase with LES model but after injecting particles, nothing happen. I mean the number of tracked, trapped, ... particles are zero. If any one know the solution please help me.

sreerao October 6, 2010 11:09

Hi roghi.sedghi,

One thing in LES DPM is you should be intelligent to use DPM. According to FLUENT, you can not use DPM with LES. However, it is mainly restricted to some applications.

What is your actual LES problem? What type of DPM particles are they? If you can give more details, I can tell some thing.

In the mean time, can you chek how many "max. number of steps" you are using in your DPM setup. It might depend on it as well.


roghi.sedghi October 10, 2010 07:15

Dear Sreerao;
I want to simulate a wet gas flow in a cyclone. It include gas flow as a continous phase and liquid droplets as dispersed phase. I have simulated continous phase with RSM and LES and after convergence I used DPM to track drops in the flow, but the drops are lost. I mean nothing is happened and the dpm concentration in both outlets is zero.
I do not know how to set up this model.

sreerao October 11, 2010 09:38


What do you mean by lost? Can I assume they are initiated and lost in the domain (incomplete).

If they are incomplete you need to increase max. number of step in DPM.

One best thing you can do is you can start simulations from scratch with DPM initialised. This technique found to work in LES without any problem. However, you can write a journal file to control your inner iterations so that for every 10 or 20 timesteps, DPM will work. This will reduce your computational time.

Hope it make sense.

roghi.sedghi October 13, 2010 03:49

Dear Sreerao;
They are not incomplete, actually they are not found in injection. This is what fluent report me:
particles not found in injection 0.
I do not know if boundary condition has a problem or not?! It is escape in the outlets and inlet and reflect in the walls.
Or maybe I should do another thing after defining injection.
Hope I can solve it.

sreerao October 13, 2010 07:26


I get the message. Can you check how are you injecting the particles. This seems to me a problem of defininig your injectioins.

Can you tell me is it surface injection or what it is? If it is not surface injection, try to change it to surface injection and see what you get.


roghi.sedghi October 17, 2010 09:05

Hi Sree;
Thank you very much for your replies.
While injecting surface this problem is solved but just a limited drops are injected in the domain!and I do not know if unsteady particle tracking is essential or steady one?

roghi.sedghi October 17, 2010 09:11

Dear Sree;
There is another problem that DPM concentration is zero in both outlets and not shown in the inlet where drops are injected.

sreerao October 18, 2010 08:05

Hi Roghi

I would stick with steady particles in LES. Please check your number of tries in DRW model. That makes a big difference. I would say try to use in range of 20 to 30 and see the difference.

Check the inlet mass flux of your continuous and DPM. If it is very dilute, and no reactions are there then try to increse DPM mass and see you can notice any different. What I noticed is in case of very dilute (m of continuos/ m of DPM < 2%) cases, Fluent is not efficient to handle.

Secondly, I believe you are doing one-way coupling. If you try to do two-way coupling, it might improve the conditions.


bhanwar February 17, 2012 05:21

dpm simulation problem
hey @ll

this is bhanwar, me currently simulating a floating bed type problem, well i have some doubts regarding dpm model like, 1. during the tracking of particle does it follow the particle or just its center???
2. density of my particle is lower than the liquid, so particle s are floating on liquid but problem is that all particles (5000) are going to column top which is not practically possible as total particle volume is grater than the column top section,so can anyone suggest me the nay alternative, or what possibly could be wrong???

thanx in advance

All times are GMT -4. The time now is 20:38.