code_saturne User's Forum

i,
I have two questions.
 
Ending the calculation
 
 
I would like to properly stop a calculation that is currently running. I tried putting a "ficstp" file in the /tmp/.. directory containing:
 
"blank line"
NTMABS=1200
 
but it does not work, the calculation keeps running. Any suggestion?
 
Mass flow direction
 
 
My studied domain consists in a parallelepiped, with two inlets (INLET and INFILTRATION) on two opposites faces and one free outlet (on the same face as INLET).
 
INLET: V=2.77m/s ; Mass flow: A = 4.897E-3
kg/s
INFILTRATION: V=0.446m/s ; Mass flow: B = 3.679E-4 kg/s
 
Hence, I should get a total incoming mass flow of A + B, and an outgoing mass flow of -(A+B).
Instead, I get (A-B) for incoming mass flow and -(A-B) for outgoing mass flow:
 
Boundary type          Code    Nb faces           Mass flow
---------------------------------------------------------------
Entree                    2        2064         0.452920094E-02
Paroi lisse               5       54798         0.173905083E-20
Paroi rugueuse            6           0         0.000000000E+00
Symetrie                  4           0         0.000000000E+00
Sortie libre              3          44        -0.452920104E-02
Indefini                  1           0         0.000000000E+00
---------------------------------------------------------------
 
 
Should I change the velocity of INFILTRATION from 0.446 to -0.446 m/s?
Any other suggestion?
 
Thanks,
 
Axel Cablé
 

Hi Axel,
Concerning your first question about a proper stop of the calculation (or an increase of the number of time-step), you are close to the solution! Here is the right way to this:
<blank line>
1200
As for your second question, could you post either your XML file or your usclim.f90 file so that I can have a look at it ? This is indeed a strange behavior...
David

Hi David, thanks for your answer the calculation stops properly now.

As for the mass flow, here is the xml file of the case I currently study, where a negative speed is imposed. The characteristics of the geometry are:

INLET: S=0.0015m², U=-2.77 m/s, Flow= A = -0.486E-2 kg/s
INFILTRATION: S=0.0028m², U=0.446 m/s, Flow= B = 0.146E-2 kg/s
(see the attached schema)

I get:

Boundary type          Code    Nb faces           Mass flow --------------------------------------------------------------- Entree                    2        1554        -0.636902676E-02 Paroi lisse               5       49769         0.153126584E-21 Paroi rugueuse            6           0         0.000000000E+00 Symetrie                  4           0         0.000000000E+00 Sortie libre              3          45         0.636918167E-02 Indefini                  1           0         0.000000000E+00 ---------------------------------------------------------------

ie (A-B) for the total incoming mass flow, and (A+B) for the total outgoing mass flow, does it seem ok to you?

Also, I have set a free outlet condition for OUTLET. Hence I'm forced to impose either a zero flux or an imposed value for my scalar: temperature. However, I would like the fluid to "carry" the heat out of the room. Should I use a non standard boundary condition in usclim.F (flux condition)?

Thanks in advance for your answer,

Axel

Hello Axel,
The results seem fine to me. Mass fluxes are always taken negative for inlets (u.n < 0 where n is the boundary face normal, always oriented towards the exterior) and positive for outlets (u.n > 0).
Concerning the oulet condition, I would advise you to first test with the standard condition: the velocity will indeed "carry" the heat out of the room except if you have backward flow (the behavior is then a bit more complex... see e.g. the comments at the beginning of the usclim.F subroutine.).
David