[wuming]

Dear all,

I'm simulating a multiphase flow using Fluent with air and a granular phase (TFM approach). I used the Johnson and Jackson model to calculate the friction pressure, and I also wrote a UDF about this friction pressure model. Choose the same friction viscosity model. However, the frictional viscosity calculated by these two models is different, the friction viscosity calculated using UDF is greater and the ratio is equal to 1.25.

Johnson and Jackson model:

the part of the UDF I write:

DEFINE_PROPERTY(cell_fri_pressure,cell,mix_thread)

{

void_s = C_VOF(cell, mix_thread);

a = void_s - alpha_min;

b = alpha_max - void_s;

aa = pow(a,2.0);

bb = pow(b,5.0);

if (void_s < alpha_min)

{

fri_pressure = 0.0;

}

else

{

fri_pressure = 0.1*void_s*aa/bb;

}

return fri_pressure;

}

Unfortunately, I cannot understand this result. Can you explain this reason?

Thank you in advance!

[vinerm]

What values are you using for minimum and maximum packing limits?

And prefer not to call the argument mix_thread since it is phase thread. Though it is just a variable name, it can lead to wrong understanding.

[wuming]

Originally Posted by vinerm

What values are you using for minimum and maximum packing limits?

And prefer not to call the argument mix_thread since it is phase thread. Though it is just a variable name, it can lead to wrong understanding.

minimum and maximum packing limits are equal to Friction Packing Limit and Packing Limit which defining in Granular Secondary Phase，respectively. The minimum packing limit is 0.5，maxium is 0.61 in two cases.

[vinerm]

Select Johnson model for frictional pressure and then fetch the values calculated by Fluent, preferably save in UDM. Compare it with the values calculated by Fluent. For fetching values calculated by Fluent, you can use generic_property function with PROP_mpg_friction_p as property ID.

[wuming]

Originally Posted by vinerm

Select Johnson model for frictional pressure and then fetch the values calculated by Fluent, preferably save in UDM. Compare it with the values calculated by Fluent. For fetching values calculated by Fluent, you can use generic_property function with PROP_mpg_friction_p as property ID.

thank you!
I use this property ID, but the result is all 0. I think my UDF is written correctly, I can get accurate results by using PROP_rho as property ID.So where did I wrong??
Thank you in advance!

sp = THREAD_MATERIAL(thread_s);
prop = (MATERIAL_PROPERTY(sp));
pc = generic_property(cell,thread_s,prop,PROP_mpg_frict ion_p,293.);

[vinerm]

You have to ensure that the loop used and the Thread passed are correct. Check some other property, such as, diameter of secondary phase.

[wuming]

Originally Posted by vinerm

You have to ensure that the loop used and the Thread passed are correct. Check some other property, such as, diameter of secondary phase.

i think the Thread passed are correct, I can get the correct density and viscosity of secondary phase

[vinerm]

These properties are generic. Try to fetch the properties defined under Secondary Phase panel and not under Materials Panel, such as, frictional viscosity.

[wuming]

Originally Posted by vinerm

These properties are generic. Try to fetch the properties defined under Secondary Phase panel and not under Materials Panel, such as, frictional viscosity.

so I should use mixture_species_loop？？

[vinerm]

That can be used only if you have species transport enabled. Otherwise, it has to be mp_thread_loop_c

[wuming]

Originally Posted by vinerm

That can be used only if you have species transport enabled. Otherwise, it has to be mp_thread_loop_c

The modified udf is as follows，pc is still equal to 0.I don't know how to modify it further. I also don't know the Property_ID of frictional viscosity . I hope you can help me.
Thank you in advance！！

DEFINE_PROPERTY(cell_fri_viscosity,cell,mix_thread )
{
Thread *cell_thread, *thread_s, *t, **pt;
real pc;
Material *sp;
Property *prop;
Domain *domain;
domain = Get_Domain(1);
mp_thread_loop_c(mix_thread,domain,pt)
{
sp = THREAD_MATERIAL(pt[1]);
prop = (MATERIAL_PROPERTY(sp));
pc = generic_property(cell,pt[1],prop,PROP_mpg_friction_p,293.);
C_UDMI(cell, mix_thread, 3) = pc;
}
....
return fri_viscosity;
}

[vinerm]

Actually, generic_property is not used for phase properties. It is useful only for species properties. There is a function that directly fetches value for frictional pressure for a given set of arguments. So, if you execute the following command within any UDF, it returns frictional pressure value

Get_Frictional_Pressure(10, 0.62, 0.775, 0.63, 0.61));

Here, first argument is ID for the frictional model, 10 identifies Johnson model. You can use 11 for Syamlal and 12 for ktgf. Second argument is the volume fraction for which you want to fetch value of Frictional Pressure. Last two arguments are packing limit and friction packing limit. However, I am not very sure of third argument. This is also volume fraction but whether it is maximum value of vf used for heat and mass transfer iac calculation or some other value, I am not sure. It also affects the frictional pressure but not a lot.

[wuming]

Originally Posted by vinerm

Actually, generic_property is not used for phase properties. It is useful only for species properties. There is a function that directly fetches value for frictional pressure for a given set of arguments. So, if you execute the following command within any UDF, it returns frictional pressure value

Get_Frictional_Pressure(10, 0.62, 0.775, 0.63, 0.61));

Here, first argument is ID for the frictional model, 10 identifies Johnson model. You can use 11 for Syamlal and 12 for ktgf. Second argument is the volume fraction for which you want to fetch value of Frictional Pressure. Last two arguments are packing limit and friction packing limit. However, I am not very sure of third argument. This is also volume fraction but whether it is maximum value of vf used for heat and mass transfer iac calculation or some other value, I am not sure. It also affects the frictional pressure but not a lot.

thank you, I find the third argument has a linear relationship with friction pressure. When it is equal to the solid volume fraction, the calculated friction viscosity is equal to the fluent model result, indicating that he is related to the solid volume fraction.

[vinerm]

It's good then.

generic_property loop works only with species and for generic properties that are given in Materials panel. For phase properties, there is as such no loop or function, at least not one that a user can make use of.