[Steady RANS]Discontinuous velocity value near a window

Questions and remarks about code_saturne usage
Forum rules
Please read the forum usage recommendations before posting.
Tsubasa
Posts: 175
Joined: Fri Nov 20, 2020 4:09 am

[Steady RANS]Discontinuous velocity value near a window

Post by Tsubasa »

Hello,

Now I am trying to Steady RANS simulation for a room, which has three windows. In this simulation, I used "Imposed P Outlet" condition for all windows as a boundary condition. Simulation was suceeded, and I checked the velocity field.
At first glance, flow field is seem to be good
a.png
, but I found discountinuous velocity value near a below window. Other two windows don't have this problem
c.png
.

Generally in this simulation, flow moved from the window where strange pressure occured to other two windows, because the window, where strange pressure occured, have the highest pressure value. Therefore, flow naturally moved to lower pressure, which other 2 windows have.

Therefore, the thing is what cause this problem and what solution can be concerned for this. I think it is no problem to impose pressure condition for all windows, but please give me some comment if you have opinion.

The reason why I gave pressure conidion to all windows is I wanted to check how flow moved by pressure difference (sometimes called natural ventilation).
Yvan Fournier
Posts: 4070
Joined: Mon Feb 20, 2012 3:25 pm

Re: [Steady RANS]Discontinuous velocity value near a window

Post by Yvan Fournier »

Hello,

The discontinuity may also simply be due to the wall condition near the window, depending on the turbulence model anv flow velocity here.

Since the values are displayed at cell centers, when you have triangles or tetrahedra near a boundary, the centers of elements who have an edge on the boundary and those who have only a vertex lay be at different distances, explaining this.

How does this look if you apply a "cell data to point data" filter in visualization ?

In any case, imposing the pressure at all windows should be OK.

Best regards,

Yvan
Tsubasa
Posts: 175
Joined: Fri Nov 20, 2020 4:09 am

Re: [Steady RANS]Discontinuous velocity value near a window

Post by Tsubasa »

Hello Yvan,

Thank you for your help.
Exactly, the velocity becomes continuous by changing cell date to point date as below.
x.png
However, it looks strange for me because there is non-physical phenomena, but is it usual when I use the wall condition near the window?

Best regards,
Tsubasa
Yvan Fournier
Posts: 4070
Joined: Mon Feb 20, 2012 3:25 pm

Re: [Steady RANS]Discontinuous velocity value near a window

Post by Yvan Fournier »

Hello,

With only 2d views, I am not really able to understand what your configuration looks like, so it is hard to comment on the results.

Best regards,

Yvan
Tsubasa
Posts: 175
Joined: Fri Nov 20, 2020 4:09 am

Re: [Steady RANS]Discontinuous velocity value near a window

Post by Tsubasa »

Hello,

Actually, this simlaition is 2D simulaiton as followings.
x.png
y.png
The mesh near the window is:
z.png
The non-physical flow is due to the mesh quality?

Thank you.
Tsubasa
Yvan Fournier
Posts: 4070
Joined: Mon Feb 20, 2012 3:25 pm

Re: [Steady RANS]Discontinuous velocity value near a window

Post by Yvan Fournier »

Hello,

It may have an influence. Inserting viscous layer cells may help (If you want to do it directly in code_saturne, see mesh modification examples in code_saturne Doxygen documentation, as this is not yet available in the GUI).

Otherwise, extruding one layer of cells outwards at the windows may also help with mesh quality.

Best regards,

Yvan
Tsubasa
Posts: 175
Joined: Fri Nov 20, 2020 4:09 am

Re: [Steady RANS]Discontinuous velocity value near a window

Post by Tsubasa »

Hello Yvan,

Let me ask you question for mesh.
Actually, I made mesh by extruding 2D mesh.
Therefore, we can see square-like mesh as following
Untitled.png
However, after the calculation, this square-like mesh is changed to tetra mesh automatically. I exactly export the square-like mesh as a med3.3 file as following
b.png

If you know about something, could you give me tips?
Tsubasa
Posts: 175
Joined: Fri Nov 20, 2020 4:09 am

Re: [Steady RANS]Discontinuous velocity value near a window

Post by Tsubasa »

Hello,

There is a thing I want to ask you.

What is the unit of pressure on output result?
In my simulation, for example, pressure range is between from 0.28 to 2.1 in color bar,
but I'm not sure the unit. Which is it by default, Cp, Pa or atm?

Best regards,
Tsubasa
Yvan Fournier
Posts: 4070
Joined: Mon Feb 20, 2012 3:25 pm

Re: [Steady RANS]Discontinuous velocity value near a window

Post by Yvan Fournier »

Hello,

For your previous question, the tetrahedra aspect might be a visualization artifact of ParaView (depending on filters applied and ParaView version).

For the pressure units, it is in Pascal. For incompressible cases, it is a pressure difference, so the absolute reference should not have too much importance (except for minor truncation error in some terms).

The pressure range in your case seems very small, so I guess your velocities are very low also ?

Best regards,

Yvan
Tsubasa
Posts: 175
Joined: Fri Nov 20, 2020 4:09 am

Re: [Steady RANS]Discontinuous velocity value near a window

Post by Tsubasa »

Hello Yvan,

-For the pressure units, it is in Pascal. For incompressible cases, it is a pressure difference, so the absolute reference should not have too much importance (except for minor truncation error in some terms).
→This means that I can directly set Cp (pressure coefficient value such as Cp=0.8, 0.3 and 0.2) at three windows and that I don't have to set exact pressure values by Pa (Pascal such as 101327, 101325 and 101324)?

In my case, when I set Cp value at three windows, velocity and pressure are following,
pressure_before.png
vel_before.png
Continue to next post...
Last edited by Tsubasa on Thu Dec 10, 2020 10:14 pm, edited 1 time in total.
Post Reply