[joshmccraney]

Hi FOAMers!

Do you know how to change gravity to a function of time? I read a thread here but this is very dated. If anyone knows how to change this I'd be very grateful!

[clapointe]

I've done something similar a long time ago, but I think the following should point you in the right direction. If you look at solvers with buoyancy you'll see

#include "readGravitationalAcceleration.H"
#include "gh.H"

in createFields. As gravity is usually constant, these lines are clearly only read once -- during simulation initialization. For gravity to very with time, though, code similar to that in the includes will need to be used every iteration.

So something like :

Create the field g and (initial) gh in create fields. Then inside the solver (probably after updating time with runTime++) something like

g = cos(t)*dimensionedScalar(gmag,dimVelocity*dimVelocity,9.81) *vector(1,0,0) + sin(t)*dimensionedScalar(gmag,dimVelocity*dimVelocity,9.81) *vector(0,1,0) //update g
//update gh, ghf
gh = g & mesh.C();
ghf = g & mesh.Cf();

Note I've not tested this, but a scan of the old thread suggests this could work. Also make sure the trig is correct!

Caelan

[joshmccraney]

Originally Posted by clapointe

I've done something similar a long time ago, but I think the following should point you in the right direction. If you look at solvers with buoyancy you'll see

#include "readGravitationalAcceleration.H"
#include "gh.H"

in createFields. As gravity is usually constant, these lines are clearly only read once -- during simulation initialization. For gravity to very with time, though, code similar to that in the includes will need to be used every iteration.

So something like :

Create the field g and (initial) gh in create fields. Then inside the solver (probably after updating time with runTime++) something like

g = cos(t)*dimensionedScalar(gmag,dimVelocity*dimVelocity,9.81) *vector(1,0,0) + sin(t)*dimensionedScalar(gmag,dimVelocity*dimVelocity,9.81) *vector(0,1,0) //update g
//update gh, ghf
gh = g & mesh.C();
ghf = g & mesh.Cf();

Note I've not tested this, but a scan of the old thread suggests this could work. Also make sure the trig is correct!

Caelan

Thanks for the response. After discussing with peers, it seems perhaps an easier approach than altering gravity is to include a body force source term. To do this, I created a new solver (copied interFoam folder and made my own titled bfInterFoam).

The UEqn.H looks like this

    MRF.correctBoundaryVelocity(U);
    
    # define omega 0.05

    const dimensionedVector mySource("mySource", dimensionSet(0,1,-2,0,0,0,0), 10*Foam::sin(runTime.value()*omega)*vector(0,1,0));

    fvVectorMatrix UEqn
    (
        fvm::ddt(rho, U) + fvm::div(rhoPhi, U)
      + MRF.DDt(rho, U)
      + turbulence->divDevRhoReff(rho, U)
     ==
        fvOptions(rho, U)
      + mysource
    );

    UEqn.relax();
.
.
.

(green is what I added). When running wmake I get the error

In file included from bfInterFoam.C:142:0:
UEqn.H: In function ‘int main(int, char**)’:
UEqn.H:14:9: error: ‘mysource’ was not declared in this scope
       + mysource
         ^~~~~~~~
UEqn.H:14:9: note: suggested alternative: ‘mySource’
       + mysource
         ^~~~~~~~
         mySource
/opt/openfoam6/wmake/rules/General/transform:25: recipe for target '/opt/openfoam6/platforms/linux64GccDPInt32Opt/applications/solvers/multiphase/bfInterFoam/bfInterFoam.o' failed
make: *** [/opt/openfoam6/platforms/linux64GccDPInt32Opt/applications/solvers/multiphase/bfInterFoam/bfInterFoam.o] Error 1

Any idea how to declare 'mysource'?

[clapointe]

There's a typo -- as the error message says try replacing "mysource" with "mySource".

Caelan

[joshmccraney]

Originally Posted by clapointe

There's a typo -- as the error message says try replacing "mysource" with "mySource".

Caelan

I feel so stupid missing mistakes like these. How did you know, so that in the future I can diagnose the issue?

[clapointe]

The error messages are always a good place to start -- and are usually pretty helpful. Using your case as an example, the relevant bit of the message read "note: suggested alternative: ‘mySource’". Otherwise carefully looking at the new code is good practice.. there's probably a post or guide floating around with more comprehensive debugging information.

Caelan

[joshmccraney]

Originally Posted by clapointe

The error messages are always a good place to start -- and are usually pretty helpful. Using your case as an example, the relevant bit of the message read "note: suggested alternative: ‘mySource’". Otherwise carefully looking at the new code is good practice.. there's probably a post or guide floating around with more comprehensive debugging information.

Caelan

Got it! Thanks for your help!