[Aadhavan]

Hi Vicarious,
thanks for the reply, i have asked that information just for clarification.
I have tried to run the case, but it giving fatal error as follows,

--> FOAM FATAL IO ERROR:
Cannot find patchField entry for cyclic ILR0
Is your field uptodate with split cyclics?
Run foamUpgradeCyclics to convert mesh and fields to split cyclics.

file: /home/cerecam/OpenFOAM/OpenFOAM-2.0.1/tutorials/incompressible/MRFSimpleFoam/exCompDomain/onlyPressure/0/U::boundaryField from line 25 to line 76.

From function GeometricField<Type, PatchField, GeoMesh>::
GeometricBoundaryField::GeometricBoundaryField
(
const BoundaryMesh&,
const DimensionedField<Type, GeoMesh>&,
const dictionary&
)
in file /home/cerecam/OpenFOAM/OpenFOAM-2.0.1/src/OpenFOAM/lnInclude/GeometricBoundaryField.C at line 271.

FOAM exiting

I dont now how to fix this.

thanks,
Aadhavan

[vicarious]

Dear Aadhavan,

I'm not familiar with the FAOM errors I'm afraid. Is your mesh fine enough? maybe it's something due to the mesh.

[Aadhavan]

Hi Vicarious,
I hope the mesh is not a problem, because I have done some simulation with the same mesh. Now I have changed only the BC.
I feel mesh wont be a problem.

Thanks Vicarious for your support,
Aadhavan

[vicarious]

My pleasure, I hope you succeed dear Aadhavan.

[Aadhavan]

Hi Vicarious,
I have done small changes in inlet and outlet. the simulation is going, let us see what is the result tomorrow.

Thanks,
Aadhavan

[Orb]

I am having a misunderstanding on how to apply the boundary conditions to my fan too. For example, ANSYS CFX has boundaries of "openings". Are these the best boundary conditions to use? I do not get velocity in - it shows these as 0 m/s. Is it better to apply a mass flow rate in?

[vicarious]

Originally Posted by Orb

I am having a misunderstanding on how to apply the boundary conditions to my fan too. For example, ANSYS CFX has boundaries of "openings". Are these the best boundary conditions to use? I do not get velocity in - it shows these as 0 m/s. Is it better to apply a mass flow rate in?

Do not use velocity inlet for this type of devices. You need to set the inlet boundary as a fixed total pressure and total temperature. You can also set it as mass flow with inlet static pressure.

[Aadhavan]

Hi Vicarious,
Thanks for your help so for, I found the mistake and fixed it.
the fan was spinning anti-clock wise direction instead of clock wise direction.
I ran the simulation again in clock wise direction, I am getting the reasonable pressure raise, but the pressure raise which I am getting is lower than the experimental result.
the simulation result is 140Pa
experimental result is 225Pa

Can you give me some hit, how can we increase the pressure raise.

I am planning to give slip condition for fan, is ok?

Thanks,
Aadhavan

[vicarious]

Hi Aadhavan,
Glad to see your simulation running. The experimental validation depends on several reasons and it's not easy to find what the problem might be. I have the same problem with pressure rise validation in my case about axial compressor. The pressure rise in high mass flow regions was lower than the experimental results. The investigation through some papers showed that it's due to the turbulence model used in numerical solution and the errors that may occur through experimental data. Also the blade profiles and the geometry usually can not be matched to the experimental test.
What do you mean slip condition? If you mean the boundary conditions about the walls then why do you want to change it?

[Aadhavan]

Hi Vicarious,
I have seen in Wiki or CFD online, "k-epsilon is inappropriate choice for problem such us compressor" is it applicable for fan as well.
If so which turbulence model I can use or else can switch of the turbulence model.

the slip boundary condition is for fan, due to viscosity the pressure raise may fall down right, that is why I am thinking to give slip BC for fan. ( no viscose effect at fan wall)

Thanks,
Aadhavan

[vicarious]

Dear Aadhavan, I disagree with such statement that "k-epsilon is inappropriate for simulating fans and compressors". Because there is not a certain way to decide which model could fit the problem. Since I already achieved good results with K-esplion standard and it is appropriate. As you know all the RANS models are not direct solution and not able to predict the turbulent field. Only the DNS models such as LES can directly simulate the eddy viscosity. I have seen several researches that approved K-e, Spalart-Allmaras and the zero-equation Baldwin-Lomax model for these cases and achieved good results. If you are able to consider DNS models, that would be very accurate. If not, the RANS models are also good for these problems.

Regarding the slip boundary, You can not just ignore the viscosity. Any device works with pressure drop due to the viscosity and so as your fan. So ignoring the viscosity is not correct.

[Aadhavan]

Hi Vicarious,
Thanks for the quick and informative reply, I think it is not possible to use DNS for fan simulation in openFOAM. I am not sure about it.


Thanks,
Aadhavan

[nash]

hi everyone,
sorry for digging up this thread.

i'm currently doing a simulation of radial fan (sirocco fan) to reproduce the curve of the fan using approach of frozen-rotor.

my question is regarding to BC inlet and outlet for pressure
this simulation is steady state and incompressible.

example for pressure difference of 100pa
i set

the totalpressure as 0 at inlet and staticpressure as -100pa
staticpressure as 0 at outlet.

is it correct?

[Aadhavan]

Hi There,
I am working on axial fan, I am using mass flow rate inlet and fixedMeanvalue (0) pressure outlet. (OpenFOAM)

Regarding your convergence, try this BC. still if you are not able to get converged solution check your mesh and computational domain.

Thanks,
Aadhavan

[nash]

Which turbulence Model are you using?
Could you please share your BC and fvScheme?

Thanks,
Nash

[chaudhry_hashim]

Hi I am working on the simulation of axial fan in order to get the fan curve, in cfd model i made the extruded regions 10 times of fan dia as inlet and outlet. I am using MRF, and i am confused about the inlet and outlet boundary conditions I am giving fixed rpm to the rotating region and i trying with the pressure inlet and outlet BCs but whenever I set outlet pressure as static my solution start diverging too much

I am not able to figuring out the problem how can i get pressure rise across the fan. I have fan curve from the manufacturer.

I would be very grateful if you people help me out.

[Aadhavan]

Hi,
I think your BC's are wrong, I am working on axial fan with the following BC's

inlet --> fixedValue or flowRateInletVelocity
outlet --> fixedMeanValue (pressure (0)) or you can try fixedValue but it is not appropriate choice.

Try the above BC's, probably it will fix your problem.

if the problem is still exists,

Check your mesh --> what is your Y+ value??
which turbulence model your are using high or low?

check your initial values of epsilon or omega and k.

as well check your schemes and limiting factors, relaxation factors.

upwind scheme is recommended for epsilon and k,



Thanks,
Aadhavan

[chaudhry_hashim]

Thankyou so much I am using 5 prism layer and k-epsilon for this. I tried with the velocity inlet and fixed static pressure (=0) the problem is that total pressure value in upstream is too high and after few iterations (150) reverse flow occur at outlet and the solution is not converging. I am working on STARCCM+ and I have attached pictures of BCs and cfd model. Kindly take a look. The interfaces I am using are inplace interface between the duct and rotating region.