[zobekenobe]

I am trying to transfer scalar from Phase A into Phase B. Hence I have defined two UDS scalar 0 in Phase A and scalar 1 in Phase B.


In order to mimic mass transfer there is a sink term for the scalar 0 in Phase A and a source term for the scalar 1 in Phase B.


The scalars are for the same solute however are uniquely defined for the each phase so that I can perform the mass transfer.


The initial concentration of the scalar in phase A entering the domain is 0.005837 kg/kg.


However when I run the UDF the amount of the scalar 1 generated is extremely small ~O(-25)

DEFINE_SOURCE(solute_sink_phase1, cellt,  mixture_cell_thread, dS, eqn)
{
    int phase_domain_index = 0;
    int n = 6;
    int m = 8;
    Domain* mixture_domain = Get_Domain(ROOT_DOMAIN_ID);
    Domain* solute_domain = DOMAIN_SUB_DOMAIN(mixture_domain, phase_domain_index);
    Thread* solute_thread = THREAD_SUB_THREAD(mixture_cell_thread, phase_domain_index);
    Thread* f_thread;
    face_t facet;
    cell_t c;
    real solute_source_value = 0.0;
    real massflow = 0.0;
    

    thread_loop_c(solute_thread, solute_domain)
    {
        begin_c_loop(cellt, solute_thread)
        {
            c_face_loop(cellt, solute_thread, n)
            {
                facet = C_FACE(cellt, solute_thread, n);
                f_thread = C_FACE_THREAD(cellt, solute_thread, n);
                massflow += F_FLUX(facet, f_thread);
            }
             solute_source_value = -0.005837*massflow*1e7*C_VOF(cellt, solute_thread);
        }
        end_c_loop(cellt, solute_thread)
    }
    return solute_source_value;
}

DEFINE_SOURCE(solute_source_phase1, cellt, mixture_cell_thread, ds, eqn)
{
    int phase_domain_index = 0;
    int m = 8;
    Domain* mixture_domain = Get_Domain(ROOT_DOMAIN_ID);
    Domain* solute_domain = DOMAIN_SUB_DOMAIN(mixture_domain, phase_domain_index);
    Thread* solute_thread = THREAD_SUB_THREAD(mixture_cell_thread, phase_domain_index);
    Thread* f_thread;
    face_t facet;

    real source_value = 0.0;
    real massflow = 0.0;
    thread_loop_c(solute_thread, solute_domain)
    {
        begin_c_loop(cellt, solute_thread)
        {
            c_face_loop(cellt, solute_thread, m)
            {
                facet = C_FACE(cellt, solute_thread, m);
                f_thread = C_FACE_THREAD(cellt, solute_thread, m);
                massflow += F_FLUX(facet, f_thread);
            }
            source_value = 0.005837*massflow*1e7*C_VOF(cellt, solute_thread);
        }
        end_c_loop(cellt, solute_thread)
    }
    return source_value;
}

The above code is for the source and sink terms in the respective phases. Could someone help understand if the strategy of a source and sink is incorrect or if there is something incorrect with the code above.



p.s. the equation is nothing legit, I was simply trying different values to see if it has any effect on the solute (apparently it doesn't)

[CeesH]

Why don't you use the mass transfer UDF for this? seems more straightforward than defining sources and sinks yourself (although that's not impossible to be sure)

[obscureed]

Hi Zobekenobe,

I agree with CeesH's comment that it is worth looking at the alternatives.

As for your UDFs, I cannot work out what you are intending them to do, but it is something rather strange.

DEFINE_SOURCE is called for each cell in the cell zone, possibly several times per iteration. It is not a good place for a loop over all cell threads and all cells in each thread. (If you need to calculate a single rate for the whole model, do it in a DEFINE_EXECUTE_AT_END and store it in a static global.) It is intended that the UDF will return the source rate for the single cell that is pointed to.

My guess as to why you get a (near-)zero rate is that you visit every single interior face in the model twice (every time the DEFINE_SOURCE is called), and the inwards F_FLUX on one side is balanced out by the outwards flux on the other.

I think you should go back to some of the examples for DEFINE_SOURCE and rethink your plan.

I hope this helps a little. Good luck!
Ed

[zobekenobe]

Thanks Ceesh and Ed,


I recently read somewhere (either the UDF manual or some example) where Fluent requires one to define the source terms for mass transfer of user defined scalars. The define_mass_transfer does not do this.

(I'll try and find the reference and paste it here)


Assuming the above is required,
Fluent then specifies that to define the source for the UDS in a multiphase simulation using the VOF model, the threads passed on to DEFINE_SOURCE is that of the mixture and hence one would have to pull the required phase thread and then operate on it.



Probably I should not loop over all the cells in the domain for the required phase. Thanks for that.




Thanks

[souza.emer]

Originally Posted by zobekenobe

Thanks Ceesh and Ed,


I recently read somewhere (either the UDF manual or some example) where Fluent requires one to define the source terms for mass transfer of user defined scalars. The define_mass_transfer does not do this.

(I'll try and find the reference and paste it here)


Assuming the above is required,
Fluent then specifies that to define the source for the UDS in a multiphase simulation using the VOF model, the threads passed on to DEFINE_SOURCE is that of the mixture and hence one would have to pull the required phase thread and then operate on it.



Probably I should not loop over all the cells in the domain for the required phase. Thanks for that.




Thanks

Could you paste this reference that you saw? Because I'm facing a problem with the DEFINE_MASS_TRANSFER macro, and your approach could help me.