[roi247]

Hello all!

I'm simulating a flow in a rectangular channel.

How to specify the material of the channel wall? If we don't specify it on the channel side wall.

The velocity close to the wall keeps decreasing, and it never reaches fully developed flow even after 6H(6 times the channel distance) distance.


Thank you

[CeesH]

Hi Rui,

Should your flow be developed at 6 entrance lengths already (did you check with a Reynolds number correlation?)

If you are not interested in flow development, you can choose to make your domain periodic; in that case it is fully developed over the entire length.

And for your final point, why do you want to specify wall material? What does it matter?

[roi247]

Originally Posted by CeesH

Hi Rui,




Should your flow be developed at 6 entrance lengths already (did you check with a Reynolds number correlation?)

If you are not interested in flow development, you can choose to make your domain periodic; in that case it is fully developed over the entire length.

And for your final point, why do you want to specify wall material? What does it matter?

Hello CeesH,

Thank you for your reply!

Unfortunately the wall stress along the wall doesn't become a stable value/

I tried UDF as the velocity inlet profile, it offers a more reliable velocity profile compared with fluid mechanics book along a vertical line close to the outlet. I'm sure it doesn't change with time any more (by checking time series figure at a point) but it does change with locations.

I have tried periodic boundary condition, but I'm not sure of the setup I did. I couldn't find a better tutorial for a one phase flow like this. Here is the TUI for the periodic boundary condition, I don't understand how to make the flow from the output back to the inlet, so I somehow combine outlet and inlet as a whole. And I specify mass flow rate. Using the same total time steps, I still could not stop them changing.
For this periodic case, The velocity profile is the same at different locations and it changes with time, so I guess my setup is wrong.

define/materials/copy fluid water-liquid
define/boundary-conditions/fluid solid yes water-liquid no no no no 0 no 0 no 0 no 0 no 0 no 1 no no no no
/mesh/modify-zones/make-periodic outlet inlet no yes no 0 0 0.39
/define/periodic-conditions/massflow-rate-specification 18.35491408 0 0.5 2 0 0 -1
define/boundary-conditions/wall watersurface no yes shear-bc-spec-shear no 0 no 0 no 0

I couldn't find a place to specify the wall material, this is a one phase case. No roughness option and I only have one fluid body which is supposed to be set as fluid.(water liquid)
when I check the material on the wall using TUI, it shows all the wall face material are fluid...



Thank you and have a nice day!

Rui

[CeesH]

Well, you cannot set a wall material because it doesn't influence your computation. How would it matter if the wall was concrete, or rubber? A no-slip wall is a no-slip wall. Regarding wall roughness, I am not 100% certain about this but I guess for LES you should model wall roughness explicitly; for wall bound flows your Y+ should be <1 to accurately capture the boundary layer, you're not using wall functions anymore in which the roughness would be a parameter.

Regarding the periodic flow, you will indeed need to make a translational periodic boundary conditions. It looks like you did that correctly. Then, either specify mass flow or pressure drop. Since LES is inherently transient, you will see fluctuations in your parameters - also ensure you have enough timesteps to make your flow develop (it won't be steady instantaneously, but it should be statistically steady in time). Finally, you should initialize your flowfield with perturbations already there, otherwise you'll likely just end up with a laminar profile.

Good luck!
Cees

[LuckyTran]

Hi please don't start additional threads on the exact same topic.

[roi247]

Thank you Cees.

For the periodic boundary condition..."Finally, you should initialize your flowfield with perturbations already there, otherwise you'll likely just end up with a laminar profile."

In the last point, if I do not have the initial condition of the related experiment, I was initializing with the uniform flow velocity. This leads me to a velocity profile fairly laminar as you mentioned...not sure how to continue with to match the experimental data from a paper.

Whats new is I have tried several ways to generate the perturbations at the inlet.spectral synthesizer, vortex method together with turbulence intensity and viscosity ratio, turbulence intensity and hydraulic diameter, turbulence intensity with turbulent length scale. the closest one is vortex method with k and e.... however, the secondary flow seems to be so strong and a large volume of flow has been directed to the right side of the channel.