CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Residence Time Distribution (http://www.cfd-online.com/Forums/fluent/28733-residence-time-distribution.html)

rajeev July 18, 2001 07:48

Residence Time Distribution
 
Can I calculate RTD when I am using a turbulence model or I have to use a different model. If different model has to be used to find the RTD then which one is most appropriate.

Has anyone tried the above problem. Please let me know the steps.

Thanks in advance Rajeev

Alain July 18, 2001 12:10

Re: Residence Time Distribution
 
I can't see the relationship between residence time and turbulence models ?

In my point of view, residence time deals with mean velocities. So, you can calculate a residence time for your flow with any turbulence model.

Can you give more precision on your problem.

Best regards

Lanre July 19, 2001 16:18

Re: Residence Time Distribution
 
Are you referring to http://www.fluent.com/solutions/examples/x93.htm

then track the transient dispersion of a tracer species.

Jin-Wook LEE July 20, 2001 05:38

Re: Residence Time Distribution
 
You can do it by particle tracking for the partilce of very small particle. Then slip velocity can be negligible, so that particle residence time is nearly same as gas residence time. Then you can obtain RTD.

For reference, my approach is as follow.

1) set the particle size very small(e.g., 1.e-09) 2) set the density of particle as mean of gas phase.

Sincerely, Jinwook


Peter Kolb July 25, 2001 09:35

Re: Residence Time Distribution
 
How can I achieve a dirac (delta) signal in the simulation?? Can I also realise that when I am using translational periodic BC ?

Kuochen July 25, 2001 11:02

Re: Residence Time Distribution
 
The best way to do it is to use the DPM model with very small particles. Also turn on the stochastic model so you can get a smooth RTD. With more than 500 particles you can get a pretty decent RTD, but be sure to turn off the length scale option and use the tracking steps (>20) so particles spend enough time in each cell. The average RTD should be close, but not exactly, to the mean residence time. If the display option takes a long time, then try the DPM sampling and histogram option. It is lot faster.

Alain July 25, 2001 13:51

Re: Residence Time Distribution
 
You can also use a fluid as a tracer. It's quite faster than particle tracking.

This is my method :

1. calculate your solution,

2. Enable multispecies model

3. Create a species "tracer" as a copy of your fluid.

4. "Inject" your tracer in the domain at a boundary or by using the patch command.

5. freeze your solution except for species (disble equation in flow solver) and then run in transient mode.

6. You can obtain a RDT by local value (probe) or mean value at bourdary (surface averaging).

Best regards

Kuochen July 25, 2001 14:25

Re: Residence Time Distribution
 
Two problem with this method: 1. Transient simulation is needed to compute RTD at the outlet, which could be quite time-consuming. 2. More cells are needed for accuracy due to numerical diffusion.

rajeev July 26, 2001 04:42

Re: Residence Time Distribution
 
Thanks a lot guys.

One more question. What should be the correct steps.

1. Calculate the steady state flow field based on turbulent model. Once it converges. then

2. Switch on the DPM. Use the calculated flow field in step 1 and solve for unsteady state the particle concentretation at outlet to get the RTD.

Correct me if my steps are wrong.

Thanks again Rajeev

Kuochen July 26, 2001 09:11

Re: Residence Time Distribution
 
Not the particle concentration, but the time variable at each particle.

Peter Kolb July 26, 2001 12:45

Re: Residence Time Distribution
 
...and it does not have to be unsteady! or I am wrong?

Kuochen July 26, 2001 13:49

Re: Residence Time Distribution
 
Correct.


All times are GMT -4. The time now is 19:28.