[icemesh]

Hi everybody,

I have a question regarding the UDF DEFINE_LINEARIZED_MASS_TRANSFER.
I have written a UDF for mass Transfer from liquid to vapor. It calculates the mass source source term in [kg/m³/s] (massflux).

Now I have a Problem with the linearization coefficients *lin_from and *lin_to. In the UDF-Manual example the linearization term is simply the mass source divided by the liquid volume fraction. The Ansys Support tells me this coefficient is simply the derivative of massflux regarding the volume fraction.
My question, is it appropriate to just divide the source term by the volume fraction, or do I have to calculate the Change in volume fraction due to that mass source term?
I am a bit confused, any help or experience with this particular UDF would be appreciated! Thanks

[mostafa_zeynalabedini]

Hi everybody.
I have a question about the multiphase mass change UDF:
I wrote a mass change udf in a 3phase problem and all parts are OK. It was interpreted, but when I want to initialize my problem it starts to do but it can’t complete initialization, means that after some hours that my CPU gets full load it can’t be finish. I think the cell loop has a problem but I don’t know what the problem is. Would you any one help me.
My problem is:
3phase
VOF
Water is the primary phase
2 secondary phases are air
My UDF is :
#include"udf.h"

DEFINE_MASS_TRANSFER(liq_gas_source,cell,thread,fr om_index,from_species_index, to_index, to_species_index)
{
real CellVelocity, Fr, epsilon, etha, entrainmentFlow, TotalArea;
real vof_cuttOff = 0.05;
real FirstBinDiameter = 0.001;
real LiquidFilmVelocity = 1;
real NV_VEC(area);
int n;
int phase1_domain_index = 0;
Thread *ContinuousPhaseThread = THREAD_SUB_THREAD(thread, phase1_domain_index);
Thread *LargeBubblesThread = THREAD_SUB_THREAD(thread, from_index);
Thread *SmallBubblesThread = THREAD_SUB_THREAD(thread, to_index);
cell_t c;
face_t f;
Thread *tf;
begin_c_loop(cell, LargeBubblesThread)
{
if ((C_VOF(cell,ContinuousPhaseThread)>vof_cuttOff) && (C_VOF(cell,ContinuousPhaseThread)<(1-vof_cuttOff)))
{
CellVelocity = sqrt(pow(C_U(cell,ContinuousPhaseThread),2) + pow(C_V(cell,ContinuousPhaseThread),2) + pow(C_W(cell,ContinuousPhaseThread),2));
if (CellVelocity>0.22)
{
Fr=(CellVelocity - LiquidFilmVelocity) / sqrt(9.81 * C_VOLUME(cell,ContinuousPhaseThread) / 8);
if ((Fr>2.5) && (Fr<7))
{
epsilon=1.245;
etha=0.018;
}
if ((Fr>7) && (Fr<30))
{
epsilon=1.4;
etha=0.014;
}
entrainmentFlow = 1.225*(8/6)*(CellVelocity - LiquidFilmVelocity)* etha * pow((Fr-1),epsilon);
}
}
}
end_c_loop(cell,ContinuousPhaseThread)
return (entrainmentFlow);
}

[dmoroian]

Hello Mostafa,
According to the DEFINE_MASS_TRANSFER macro documentation, you are not supposed to modify any of the input variables that are present in the header, they are delivered by the solver.
In other words, you should not perform the cell loop, is the solver that calls the macro for each cell.
The fact that you modify the variable "cell" inside your udf should not be a problem, however, the loop that you introduce means that for each cell in your domain you will go through all the cells of the entier computational domain, and this happens for each iteration/time_step. Consequently, the time to compute will increase dramatically (N*N).

[mostafa_zeynalabedini]

dear dmoroian
I got it. and thank you every much.
I really tahnks about your help.

[mostafa_zeynalabedini]

Hello.
I have another question:
I want to compute the surrounding area of a cell in a mass change define macro.
I write this UDF but FLUENT doesn’t interpret this UDF. Its error is :
line 23: structure reference not implemented
I tested 2 ways but both are wrong!
Would you help me please?
This is my UDF:
#include"udf.h"
DEFINE_MASS_TRANSFER(LargeBubbles2SmallBubbles, cell, thread, from_index, from_species_index, to_index, to_species_index)
{
real CellVelocity, Fr, epsilon, etha, entrainmentFlow, TotalArea;
real Vof_CutOff = 0.05;
real FirstBindiameter = 0.001;
real LiquidFilmVelocity = 1;
real NV_VEC(area);
int n, i;
int phase1_domain_index = 0;
Thread *ContinuousPhaseThread = THREAD_SUB_THREAD(thread,phase1_domain_index);
face_t f;
Thread *tf;
if (C_VOF(cell,ContinuousPhaseThread)>Vof_CutOff)
{
CellVelocity = sqrt(pow(C_U(cell,ContinuousPhaseThread),2)
+ pow(C_V(cell,ContinuousPhaseThread),2)
+ pow(C_W(cell,ContinuousPhaseThread),2) );
if (CellVelocity > 0.22)
{
/*first way */
for (i==0 ; i<6 ; i++)
{
tf = C_FACE_THREAD(cell,ContinuousPhaseThread,i);
f = C_FACE(cell,ContinuousPhaseThread,i);
F_AREA(area,f,tf);
total_area +=NV_MAG(area);
}
/*
/*second way*/
c_face_loop(cell,ContinuousPhaseThread,n)
{
f = C_FACE(c,t,n);
tf = C_FACE_THREAD(c,t,n);
}
begin_f_loop(f,tf)
{
F_AREA(area,f,t);
total_area +=NV_MAG(area);
}
end_f_loop(f,t)

Fr = (CellVelocity - LiquidFilmVelocity) / sqrt(9.81 * C_VOLUME(cell,ContinuousPhaseThread)
if ((Fr > 2.5) && (Fr < 7))
{
epsilon = 1.245;
etha = 0.018;
}
entrainmentFlow = (CellVelocity-LiquidFilmVelocity) * etha * pow((Fr - 1),epsilon);
}
}
return (entrainmentFlow);
}