CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   Fluent, reattachment length (

Adrian March 25, 1999 14:36

Fluent, reattachment length
i'm currently doing on turbulent flow for a sudden expansion pipe flow. My Reynold number is 20000. When i patch a heat source of 300kW/m3 into my recirculation zone. It seems that there is no reduction in my recirculation length. it suppose to reduce my recirculation length. As for laminar flow of Re=200, it did works. I'm wandering what could my problem is?

Thanks in advance

Sung-Eun Kim March 25, 1999 16:45

Re: Fluent, reattachment length
I'm curious what physical situation you're trying to model. The question I end up asking myself after reading your query is what led you to patch heat source in "your recirculation" region.

And I wonder how the momentum equation is coupled with the energy equation in your setup. Is it through a temperature dependent density or viscosity ? If it's the case, is the variation of the property large enough to cause an appreciable difference in the reattachment length ? At such high Reynolds number (20,000), the reattchment length may not be as much sensitive to the change in the property, i.e., Reynolds number, as at lower Reynolds number, i.e., Re = 200 you quoted.

Adrian March 25, 1999 17:08

Re: Fluent, reattachment length
i'm now actually trying to model a combustion that lead to the reduction of the CRZ length. Instead of a complex flow, for simplicity, i assume the heat souce to be homogenous. So, that's why i'm trying to patch the heat source in the CRZ region.

how my momentum equation coupled with the energy equation is through the temperature dependent density, which is P=rho*R*T. So, my density is air.

I don't quite understand why you say the variation is not large enough. Cause what i find in the journal is that it works for laminar flow so as for turbulent flow. Beside that, i found that the contour plot for temperature is in the range of 273 to 283 only for heat source of 300KW/m3. But for laminar, the temperature seems to be in the range of 273 to 373 for 10KW/m3. Is it because of some property that i didn't enabled in FLUENT?

Sung-Eun Kim March 25, 1999 18:38

Re: Fluent, reattachment length
Thank you for sharing the information with me and others. I think it's a very creative modeling idea that will give valuable insight to the effects of temperature on the size of the internal recirculation zone.

IMHO, what you're observing may quite possibly be the numerical representation of of the very nature of turbulent flow, sir. Please forgive me for quoting a few cliches here. But, turbulent flows diffuse heat much faster than laminar flows. Turbulence stirs things (momentum, heat, species concentration) up much much more effectively than being left to molecular mixing. Turbulent diffusivity can larger than molecular diffusivity by several orders of magnitude. As a result, the temperature distribution in laminar flows would end up being more "peaked" than in turbulent flows (which will be more uniform) given the same amount of heat source. You may think of it as counter-intuitive possibly because you associate it experience with energy conservation, i.e., for the same amount of heat, why doesn't turbulent flow heat up as much as laminar flow? But if you account for the convective energy flux through the downstream exit, all this seems to fit together.

If the temperature rise in the turbulent flow case is much smaller than the very high local temperature in the laminar case, wouldn't it mean that the size of the recirculation zone can be less affected in turblent flow than in laminar flow, given the same amount of heat source ?

Adrian March 25, 1999 19:07

Re: Fluent, reattachment length
Thank you too for sharing knowledge with me. It really makes sense. But could there be any possible way to reduce the recirculation length in the turbulence model by inputing the heat source without taking account the convective energy flux in fluent?

Beside that, will there be a difference if using k-e and rng k-e? and what's the difference between k-e and RNG k-e in fluent? What does the model can and cannot capture?

thank you.

All times are GMT -4. The time now is 17:44.