Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Laminar inflow into a turbulent domain

Please login with a confirmed email address before reporting spam

I have a 3D geometry in which turbulent flow develops downstream, but the inlet to the domain is a pipe that has laminar inflow. The turbulence physics interfaces do not have a laminar inflow boundary condition like the laminar flow interface does. What is the best way to apply a laminar inflow condition in this case?

In the distant past (pre-3.5) I was able to solve Poisson's equation on a disk to get the laminar profile and then couple that to the inlet boundary. I cannot figure out how to do that in 4.3a.

Any help would be appreciated.
--
Steven Conrad, MD PhD
LSU Health

5 Replies Last Post Apr 5, 2013, 7:22 p.m. EDT
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 4, 2013, 1:41 a.m. EDT
Hi

I would say as in 3.5: you add a 2nd model, in the first mod1 you solve your 2D case, in the mod2 the 3D case and you get either a model coupling projection operator to map the two surfaces from mod1 to mod2, or you play with the equations to transfer the pertinent _u_,p vector/scalar fields from one to the next.

On the other hand, I usually set up a parabolic velocity profile and manage to work with that, less precise but even quicker to define

--
Good luck
Ivar
Hi I would say as in 3.5: you add a 2nd model, in the first mod1 you solve your 2D case, in the mod2 the 3D case and you get either a model coupling projection operator to map the two surfaces from mod1 to mod2, or you play with the equations to transfer the pertinent _u_,p vector/scalar fields from one to the next. On the other hand, I usually set up a parabolic velocity profile and manage to work with that, less precise but even quicker to define -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 4, 2013, 7:20 p.m. EDT
Thanks, Ivar

It seems the second approach is easier. I created an analytical function that defines a 2D parabolic profile, but cannot see how to apply it as a boundary condition on a particular boundary. I thought I could use the variables s1 and s2 as function parameters to 'fit' the function to the 2D inlet boundary as a velocity, but I it returns that s1/s2 could not be evaluated.

Any ideas or suggestions?

--
Steven Conrad, MD PhD
LSU Health
Thanks, Ivar It seems the second approach is easier. I created an analytical function that defines a 2D parabolic profile, but cannot see how to apply it as a boundary condition on a particular boundary. I thought I could use the variables s1 and s2 as function parameters to 'fit' the function to the 2D inlet boundary as a velocity, but I it returns that s1/s2 could not be evaluated. Any ideas or suggestions? -- Steven Conrad, MD PhD LSU Health

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 5, 2013, 1:57 a.m. EDT
Hi

"s" works fine for 2D, but in 3D there is always some ambiguities.

So what I usually do is to place a cylindrical coordinate system (Model Definitions Coordinates) and define the formula directly as an equation depending on the sys2.r variable

--
Good luck
Ivar
Hi "s" works fine for 2D, but in 3D there is always some ambiguities. So what I usually do is to place a cylindrical coordinate system (Model Definitions Coordinates) and define the formula directly as an equation depending on the sys2.r variable -- Good luck Ivar

Fabrice Schlegel COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 5, 2013, 11:41 a.m. EDT
Dear Steve,

To specify the velocity profile on the boundary, I would recommend following Ivar’s advice and specify the velocity profile as a function of the x,y,z coordinates rather than s1 and 2. In the attached model, I modified the backstep model from the model library and input a parabolic velocity profile.

The expression for such a profile is u=U_Max*(1-r^2/R^2) . Integrating this on the inlet surface, you get that U_Max= 2*U_Mean. Therefore, I used the following expression for the inlet velocity:
2*v_mean*(1-r^2/R^2)

The radius r is defined in the variable table as a function of y and z. Please let me know if you have more questions.
Dear Steve, To specify the velocity profile on the boundary, I would recommend following Ivar’s advice and specify the velocity profile as a function of the x,y,z coordinates rather than s1 and 2. In the attached model, I modified the backstep model from the model library and input a parabolic velocity profile. The expression for such a profile is u=U_Max*(1-r^2/R^2) . Integrating this on the inlet surface, you get that U_Max= 2*U_Mean. Therefore, I used the following expression for the inlet velocity: 2*v_mean*(1-r^2/R^2) The radius r is defined in the variable table as a function of y and z. Please let me know if you have more questions.


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 5, 2013, 7:22 p.m. EDT
Thanks, Fabrice, Ivar

I will implement this approach. Is is reasonable that this boundary condition should be available in the turbulent flow interface as it is in the laminar flow? If so, I can submit as a feature request.

--
Steven Conrad, MD PhD
LSU Health
Thanks, Fabrice, Ivar I will implement this approach. Is is reasonable that this boundary condition should be available in the turbulent flow interface as it is in the laminar flow? If so, I can submit as a feature request. -- Steven Conrad, MD PhD LSU Health

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.