[spalartallmaras]

Hallo,

I simulate a channel flow for Re_tau=640. nu ist 2.53e-3 [m^2/s] and Characteristic linear dimension is 1 [m].

Length of channel is 6 [m], Height 2 [m] and width also 2 [m].


So, my u_tau is now 1.62 [m/s] and (u_tau)^2 is 2.624 [m^2/s^2].

What a problem is, my <u'u'> is too high nearby the wall. It is almost yet about 135-140 [m^2/s^2] with flow though time ~500 [s]... Even I divide them by 2.624 [m^2/s^2], it is too big.

As you already know, reference datas show me that the highest <u'u'>/(u_tau)^2 is about 7.5-8.0.

Or did I understand something wrong?




(+) till flow through time ~100[s] the kinematic viscosity was 2.53e-5 [m^2/s] and after that I switched it in 2.53e-3 [m^2/s]

[FMDenaro]

Originally Posted by spalartallmaras

Hallo,

I simulate a channel flow for Re_tau=640. nu ist 2.53e-3 [m^2/s] and Characteristic linear dimension is 1 [m].

Length of channel is 6 [m], Height 2 [m] and width also 2 [m].


So, my u_tau is now 1.62 [m/s] and (u_tau)^2 is 2.624 [m^2/s^2].

What a problem is, my <u'u'> is too high nearby the wall. It is almost yet about 135-140 [m^2/s^2] with flow though time ~500 [s]... Even I divide them by 2.624 [m^2/s^2], it is too big.

As you already know, reference datas show me that the highest <u'u'>/(u_tau)^2 is about 7.5-8.0.

Or did I understand something wrong?




(+) till flow through time ~100[s] the kinematic viscosity was 2.53e-5 [m^2/s] and after that I switched it in 2.53e-3 [m^2/s]

I have a question, it seems you are solving the dimensional form of the equations, so how do you evaluate that you have Re_tau=640? Are you fixing the driving pressure or the mass constraint?

[spalartallmaras]

Thank you for your answer. Yes, I simulate the problem with OpenFOAM and I use patchMeanVelocityForce with 30.8 [m/s].

Of course the patch for the source is inlet.

But for Re_bulk the characteristic linear dimension is 2 [m]. I follow the definition in the website "https://www.rs.tus.ac.jp/~t2lab/db/" and "https://www.openfoam.com/documentation/guides/latest/doc/verification-validation-turbulent-plane-channel-flow.html".

Boundary conditions for inlet and outlet are periodic (cyclic in OpenFOAM).

[FMDenaro]

Originally Posted by spalartallmaras

Thank you for your answer. Yes, I simulate the problem with OpenFOAM and I use patchMeanVelocityForce with 30.8 [m/s]. Of course the patch for the source is inlet.

Boundary conditions for inlet and outlet are periodic (cyclic in OpenFOAM).

You did not answered to my questions....

[spalartallmaras]

I am sorry for that I modified the post before you answer me once more. As I wrote I follow the definitions of the two websites.

Re_tau = u_tau*delta/nu

delta is half of height of channel, 1 [m].

[FMDenaro]

Originally Posted by spalartallmaras

sorry for that I modified the post before you answer me once more. As I wrote I follow the definitions of the two websites.

Re_tau = u_tau*delta/nu

delta is half of height of channel, 1 [m].

The problem is that you don't know u_tau until you have the fully solution. Depending on the setting of the driving force (mass or pressure) you should check a-posteriori the resulting u_tau from the computation of the tau_wall distribution.
In https://www.openfoam.com/documentati...nnel-flow.html the u_tau velocity is addressed as "estimated".


Hence, I suppose you worked by evaluating u_tau=Re_tau*ni/delta but how this evaluation practically enters in your code?

[spalartallmaras]

Okej, I will check what you answered me. Thank you very much for your comments.

[FMDenaro]

Originally Posted by spalartallmaras

Okej, I will check what you answered me. Thank you very much for your comments.

The key is that you need to fix a constant pressure gradient in the momentum equation in such a way that it balances the tau_wall computed from Re_tau. Then, the bulk velocity should be evaluated from the solution, that is you have to evaluate the mass flow rate