[Andrea1984]

Hi,

I would like to simulate a forced convection heat transfer problem using OF. The flow is incompressible and turbulent.

Any hints for the choose of the right solver, BCs and other settings are welcome.

Thanks,

Andrea

[romant]

Originally Posted by Andrea1984

Hi,

I would like to simulate a forced convection heat transfer problem using OF. The flow is incompressible and turbulent.

Any hints for the choose of the right solver, BCs and other settings are welcome.

Thanks,

Andrea

For heat transfer problems with incompressible flow, take a look at bouyantBoussinesqSimpleFoam and buoyantBoussinesqPimpleFoam, for steady state and transient calculations respectively. There are a couple of tutorials included in OpenFOAM

[Andrea1984]

Hi Roman,

thank you for your reply.
These solvers are suitable only for buoyant flows, or are still valid in case of forced convection?

[romant]

yes, they are also valid for forced convection. The "buoyant" part of the name just states that they use the boussinesq approximation for buoyancy, where small changes in density are modeled through a temperature difference. but other than that the density is constant.

[Andrea1984]

Ok, their names are a little bit misleading.
I'll have a look to the source code and run some test cases.

Thank you for the infos.

Andrea

[romant]

Hej,

it doesn't really matter that the buoyant part this is in the solver, since in forced convection your change in density and the buoyancy forces associated with it will be minimal (the flow will not be affected by it). Therefore, the solvers can be used for forced convection and for natural convection.

[cm_jubayer]

Hi Andrea,

I am also simulating forced convection over a solar panel in the atmospheric boundary layer. I am using buoyantBoussinesqPimpleFoam. Based on nOuterCorrectors you can switch between Pimple and Piso. Also, I am using 0 gravity which makes p and p_rgh same. However, I am facing bounding problem with omega specially (using SST k omega), trying different limited schemes, no luck yet.

But, to test the solver, I ran a simulation with forced convection over a flat plate with high Re and found a good match with the turbulent boundary layer correlation in theory. Also, I think to solve forced convection correctly, boundary layer has to be resolved completely (y+<1) without using wall functions.

Jubayer

[Andrea1984]

Hi Jubayer,
Unfortunately I had to stop my testing with OF in forced convection because right now I'm pretty busy with other undertakings; however I would be glad if you can attach your convection over a flat plate case.

Thanks in advance,

Andrea

[cm_jubayer]

You can download it from here


Jubayer

[Andrea1984]

Thank you so much

Andrea

[aamo]

Originally Posted by cm_jubayer

You can download it from here


Jubayer

Hi Jubayer, I think you have removed the files from your dropbox. could u please upload it again.

[cm_jubayer]

Here is the link again:

https://dl.dropboxusercontent.com/u/...atPlate.tar.gz


Jubayer

[sandeeprapol]

Hi,
I would like to simulate a forced convection heat transfer problem using OpenFOAM. The flow is compressible and turbulent. Which solver is suitable?


Thanks & Regards,
Sandeep Rapol

[romant]

Originally Posted by sandeeprapol

Hi,
I would like to simulate a forced convection heat transfer problem using OpenFOAM. The flow is compressible and turbulent. Which solver is suitable?


Thanks & Regards,
Sandeep Rapol

Take a look at this site http://cfd.direct/openfoam/user-guid.../#x13-890003.5

You probably want to take one of the solvers from the compressible family. They are nicely explained there and you should think about whatelse your solver should do. Steady state or transient, supersonic or subsonic flow, do you want to include buoyancy at a later stage (then take a look at the heat transfer solver family).

It shouldn't be too hard to find the right solver for your purpose.

[roadRunner]

Setting gravity vector to zero should disable all buoyancy effects and enable pure forced convection.

[John Handel Kennedy]

Hi,
I am trying to simulate Flow in a Straight Pipe with Heat transfer.
I am using the buoyantBoussinesqSimpleFoam solver.
I have made g and beta to be zero.
The temperature of the wall is 373K and inlet fluid temperature is 293K.
The inlet velocity is 1m/s.
The diameter of the pipe is 1m and the nu value is 0.01 which makes a Reynolds number to be 100.
The laminar Prandtl number is 1.5.

I got fully developed flow in simpleFoam i.e. the velocity jumped to 2m/s.
However I am not able to get the same velocity profile in buoyantBoussinesqSimpleFoam, The velocity is decreasing towards the outlet.

How do we solve this problem?
What should I specify in the alpha_t and p_rgh files?

Regards
John

[hello77]

Originally Posted by romant

Hej,

it doesn't really matter that the buoyant part this is in the solver, since in forced convection your change in density and the buoyancy forces associated with it will be minimal (the flow will not be affected by it). Therefore, the solvers can be used for forced convection and for natural convection.

hi, could you please justify this for me. I mean is the Boussinesq can be used in forced convection?

[romant]

Originally Posted by hello77

hi, could you please justify this for me. I mean is the Boussinesq can be used in forced convection?

If the influence is almost zero, then it can be used in forced convection. The Boussinesq term is g*rho*beta(T-T_ref). This will give you an extra acceleration in the flow, if there is a temperature different. However, this acceleration can be small compared to the forces which are acting on the fluid due to the flow itself. In forced convection the influence of the buoyant forces is small and therefore the having this added term does not influence the flow much.

For more information on how to relate the buoyant forces to the viscous and momentum forces have a look at https://en.wikipedia.org/wiki/Forced_convection .

Cheers,
Roman

[cfdsolver1]

Hello, if you want to model forced convection using bouyantBoussinesqSimpleFoam, you need to give gravity and beta values zero. Therefore, it turns into forced convection case.

[hello77]

thank for romant and cfdsolver1. my question was in general, can we use the Boussinesq in forced convection?