[fabrice]

Hi,guys
I am a new openfoam(2.3.0) user. I want to simulate the flow over a cavity(in the attachment:mesh), it works with pimpleFoam(attachment:1.png). Afterwards, I want to add thermo mode and take the buoyancy effect into consideration though I think the flow is still governed by the flow. Thus, I chose buoyantBoussinesqPimpleFoam but the result is bad. As you can see in the bottom wall beside the inlet(attachment:2.png) the velocity diverge. Could someone help me with the problem? Weather I used the wrong boundary conditions? Here are the BCs:
alpha:
inlet
{
type fixedValue;
value uniform 0;
}
outlet
{
type zeroGradient;
}

upperWall
{
type alphatJayatillekeWallFunction;
Prt 0.85;
value uniform 0;
}

lowerWall
{
type alphatJayatillekeWallFunction;
Prt 0.85;
value uniform 0;
}

frontAndBack
{
type empty;
}
epsilon:
inlet
{
type fixedValue;
value uniform 0.1;
}
outlet
{
type fixedValue;
value uniform 0.1;
}
upperWall
{
type zeroGradient;
}
lowerWall
{
type epsilonWallFunction;
value uniform 0.1;
}
frontAndBack
{
type empty;
}
k:
inlet
{
type fixedValue;
value uniform 0.1;
}

outlet
{
type fixedValue;
value uniform 0.1;
}

upperWall
{
type kqRWallFunction;
}

lowerWall
{
type kqRWallFunction;
}

frontAndBack
{
type empty;
}
nut:
inlet
{
type calculated;
value uniform 0;
}
outlet
{
type calculated;
value uniform 0;
}

upperWall
{
type zeroGradient;
}

lowerWall
{
type nutkWallFunction;
value uniform 0;
}

frontAndBack
{
type empty;
}
p:
inlet
{
type zeroGradient;
}

outlet
{
type fixedValue;
value uniform 0;
}

upperWall
{
type zeroGradient;
}

lowerWall
{
type zeroGradient;
}

frontAndBack
{
type empty;
}
p_rgh:
inlet
{
type fixedValue;
value uniform 0;
}
outlet
{
type fixedFluxPressure;
rho rhok;
value uniform 0;
}

upperWall
{
type fixedFluxPressure;
rho rhok;
value uniform 0;
}

lowerWall
{
type fixedFluxPressure;
rho rhok;
value uniform 0;
}

frontAndBack
{
type empty;
}
T:
inlet
{
type fixedValue;
value uniform 300;
}
outlet
{
type fixedValue;
value uniform 300;
}
upperWall
{
type fixedValue;
value uniform 300;
}
lowerWall
{
type fixedValue;
value uniform 305;
}
frontAndBack
{
type empty;
}
U:
inlet
{
type fixedValue;
value uniform (7 0 0);

}
outlet
{
type zeroGradient;
}
upperWall
{
type zeroGradient;

}
lowerWall
{
type fixedValue;
value uniform (0 0 0);
}
frontAndBack
{
type empty;
}
blockMesh:
convertToMeters 100;

vertices
(
(-0.09 0 -0.01)
(-0.09 0.06 -0.01)
(-0.03 -0.06 -0.01)
(-0.03 0 -0.01)
(-0.03 0.06 -0.01)
(0.03 -0.06 -0.01)
(0.03 0 -0.01)
(0.03 0.06 -0.01)
(0.09 0 -0.01)
(0.09 0.06 -0.01)

(-0.09 0 0.01)
(-0.09 0.06 0.01)
(-0.03 -0.06 0.01)
(-0.03 0 0.01)
(-0.03 0.06 0.01)
(0.03 -0.06 0.01)
(0.03 0 0.01)
(0.03 0.06 0.01)
(0.09 0 0.01)
(0.09 0.06 0.01)


);

blocks
(
hex (0 3 4 1 10 13 14 11) (15 15 1) simpleGrading (1 1 1)
hex (2 5 6 3 12 15 16 13) (15 15 1) simpleGrading (1 1 1)
hex (3 6 7 4 13 16 17 14) (15 15 1) simpleGrading (1 1 1)
hex (6 8 9 7 16 18 19 17) (15 15 1) simpleGrading (1 1 1)
);

edges
(
);

boundary
(
inlet
{
type patch;
faces
(
(1 11 10 0)
);
}
outlet
{
type patch;
faces
(
(9 19 18 8)
);
}
upperWall
{
type patch;
faces
(
(1 4 14 11)
(4 7 17 14)
(7 9 19 17)

);
}
lowerWall
{
type wall;
faces
(
(0 3 13 10)
(3 2 12 13)
(2 5 15 12)
(5 6 16 15)
(6 8 18 16)
);
}
frontAndBack
{
type empty;
faces
(
(0 3 4 1)
(10 13 14 11)
(2 5 6 3)
(12 15 16 13)
(3 6 7 4)
(13 16 17 14)
(6 8 9 7)
(16 18 19 17)
);
}
);

mergePatchPairs
(
);

[fabrice]

Could someone help me?

[fabrice]

I found the date of the first three cells are quite abnormal even the first time step.
alpha:
1250
(
0.0489454
0.00397838
0.00139884
0.00112667
0.00107311
epsilon:
1250
(
13.205
0.00676769
0.000323388
0.000172114
0.000149278
k:
1250
(
2.25068
0.0145627
0.00187848
0.00122868
0.00111651
nut:
1250
(
0.034525
0.00282023
0.000982044
0.000789416
0.000751578
p:
1250
(
0
-134.595
-84.8425
-86.5811
-82.0199
-80.8283
p_rgh:
1250
(
-8.70777
-133.635
-83.8027
-85.5412
-80.98
T:
1250
(
3592.14
307.004
280.026
280
280
U:
1250
(
(-24.7702 -26.2486 0)
(-2.90903 -4.45426 0)
(-7.17743 -0.975982 0)
(-3.25353 -0.955117 0)
(-2.95407 -0.49689 0)
(-2.15063 -0.328007 0)
(-1.70604 -0.221482 0)
(-1.36991 -0.157735 0)
(-1.11976 -0.117065 0)

Does anyone know what cause this problem?These first three cell dates increase dramatically when timestep grows but other cell dates seem remain the acceptable dimension

[Anne Lincke]

Hey Fabrice,

by first look I only wonder why you choose different boundary conditions for epsilon and nut at upper and lower wall ?
Also you set the velocity and pressure to zeroGradient on one wall. This does not make sense in my opinion.
Try to set velocity to zero at both walls and adapt the boundary conditions for epsilon and nut, accordingly....

[fabrice]

Thank you for you reply. You are right, I should set velocity (0 0 0) at wall.But what should I set for pressure? What do you mean by 'why you choose different boundary conditions for epsilon and nut at upper and lower wall' , can you give me some suggestion of BCs for epsilon and nut?

[Anne Lincke]

Hey Fabrice,

for epsilon and nut you were using different BC:


epsilon:


upperWall
{
type zeroGradient;
}
lowerWall
{
type epsilonWallFunction;
value uniform 0.1;
}

nut:

upperWall
{
type zeroGradient;
}

lowerWall
{
type nutkWallFunction;
value uniform 0;
}

This is what I meant with choosing different boundary conditions.
I recommend to use epsilonWallFunction and nutkWallFunction for lowerWall and upperWall provided that you have a turbulent case and that your mesh is coarse enough at the wall such that y+~30.
For pressure use zeroGradient at the walls.

Regards
Anne

[fabrice]

HI Anne,
Thank you again for your help!I will try that. As you can see I have already used zeroGradient for pressure at walls but I am not sure about p_rgh.

[fabrice]

HI Anne,
In this case top wall is atmosphere that is why I use the different BCs for top and bot. I am sorry forget to explain that.

[Anne Lincke]

OK, I do not know how to treat this boundary then....

[fabrice]

hi, Anne
I change the case into a simple one(tutorial file heat-transfer/hotroom) and I add inlet and outlet to let it flow. I am using wall function at top and bottom walls for nut and epsilon as you suggest , other boundary conditions are the same. But the result still not right, the values of the first three meshes(left bottom corner) are extramely large for temperature and velocity. I am sorry I do not know how to submit pictures when I quick reply your message, I hope you can understand.
By the way, do you have some successful case for buoyantboussinesq solver?

[Anne Lincke]

Hey Fabrice,

I used buoyantBouissinesqSimpleFoam.
For this solver, I obtained reasonable results.

Regards
Anne

[fabrice]

Hi Anne,
Can you please share your case or just do a very simple case like flow in the tunnel? I really need help and I have also tried buoyantBouissinesqSimpleFoam same problem happend

[fabrice]

I found that if there is no flow, I mean only temperature difference (buoyancy driven),the solver can work properly . If I add velocity inlet to the case, problems occur.

[Anne Lincke]

Hey Fabrice,

just send me an e-mail and I will send you my case. I sent my address in a private message to you.

Regards
Anne

[eugenioS]

Hello everyone,

I'm trying to run a similar case with buoyantBoussinesqPimpleFoam (Cavity with inlet/outlet flow and heated floor) and I'm getting a very similar spurious solution. I found this thread and I wonder if you finally found a solution to the problem.

Regards,

Eugenio S.