CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   calculation the residence time distribution by CFX (

ahmedsalemy June 22, 2009 11:05

calculation the residence time distribution by CFX
1 Attachment(s)
Hi Guys,
I carried out the residence time distribution test on my model in the laboratory, and I want to verify it by simulation with CFX-10, and I donít know how I do it, but I found the method in the paper with title (Residence Time Distribution in a Multistage Agitated Contactor with Newtonian Fluids: CFD Prediction and Experimental Validation) and I donít understand the way to calculate RTD in ANSYS from this paper because the paper should be abbreviated. Any one please guides me, how I can do it. And I want to explain for you I donít use rotating impeller but my model is reactor with multi-channel, and I selected the characteristics of flow as Steady state, but in the paper was transient state because there is impeller in the reactor. I attached the important part in the paper to explain for you.
Thanks a lot,

Rui June 22, 2009 11:49

Have you though of / tried the common approach to calculate the time of residence?:

- Create an Additional Variable for the time of residence, for example: tR.
- Set a transport equation without diffusion to tR.
- Set a Source Value of 1 [s/s] to that equation (the fluid gets 1 s older per second, right?).
- Inlet condition for tR = 0, if it is the case.

ahmedsalemy June 23, 2009 06:35

hi Rui:
Thank you very much

ahmedsalemy June 25, 2009 08:01

could you help me
Hi Rui
If I make with my problem as in the tutorial in CFX with title (flow from a circular vent) to simulate my model in two steps initially without tracer, and use the results from this simulation as the initial guess for transient simulation which will be in the 2nd step to obtain the value of tracer concentration at the same location that I need.
I need your opinion in the next items:
  • In the tutorial the concentration of smoke rises exponentially with time, but in my case, does the tracer rise exponentially with time too? If not what is the equation which is used in the CFX Expression language in this case. I think no and the equation for this curve may be the equation for Gauss curve. If this is right Could you please right to me this equation?
[The equation in tutorial is ExpFunction= FinalConcentration*abs(1-exp(-t/TimeConstant))FinalConcentration=1kg/m^-3 & TimeConstant=3s]

  • In the tutorial because of smoke rises exponentially with time, the interval between the timesteps is smaller at the beginning of the simulation than the end. In my case, do I make the timesteps constant?

Rui June 25, 2009 10:40

If I remember well, whoever conceived the tutorial decided that the concentration of smoke at the inlet would raise exponentially with the time. That's an imposition of the system that is being simulated.
That's just a tutorial. If you were simulating a real case, you should impose the real evolution of the concentration of smoke at the inlet.

About your case, my sincere opinion is that first of all you should try to understand what you are doing and what you want to do. And not use CFX as magic wand .
I never measured experimentally the time of residence. Perhaps, as you say, it's possible to measure a tracer concentration and make some analogy with the time of residence ... I've never really though about that.
But, that should be your very first question. Why and how can you use a tracer to measure the time of residence?
Than, before attempting doing something in CFX, try to write down a (simplified) physical model of what you want to simulate. Then check if in CFX is possible, and how, to deal with that physical model.

About that tutorial: have you though why they can do first a steady state simulation, and afterwards a transient simulation? Do you think it's right? Why?
Can you do with your case the same? Why?

As I said on the beginning, at the inlet you should impose boundary conditions (including tracer concentration, if you're simulating it) as close to reality as possible.
I have no idea at all what tracer concentration distribution you used when you did your lab experiments (I couldn't still afford to buy that telepathic superpower helmet on sale on e-bay :(). If it was a Gauss distribution, so use it. I surely know the hilly shape of the Gauss distribution but don't remember the equation. But google it, I'm pretty sure you'll find it in less than 3 clicks.

Try to understand what you're doing and wanna do.

All times are GMT -4. The time now is 14:26.