[estevaotolentino]

Hi,

I'm having some problems with the DEFINE_SPECIFIC_HEAT macro. I need to vary the specific heat of a solid over time, where the equation that varies this property is in function of M (moisture content of a bed of grains). The variable M is defined outside of the DEFINE_SPECIFIC_HEAT macro (it is defined inside another macro called DEFINE_ADJUST), then I need to fetch this variable M through memory C_UDMI (c, t, 0), where M is stored in the UDM 0 memory. Here is my macro:

DEFINE_SPECIFIC_HEAT(specific_heat_paddy, T, Tref, h, yi)
{
real cp_paddy, M;
Thread *t;
cell_t c;
M = C_UDMI(c, t, 0);
cp_paddy = 1110 + 44.8 * (M / (M + 1));
*h = cp_paddy * (T - Tref);
return cp_paddy;
}

I'm interpreting the UDF. When I initialize the problem, an error message appears:

"Error: received a fatal signal (Segmentation fault).
Error Object: #f
/solve/initialize/initialize-flow"

Could anyone help me solve this problem? Thanks.

[Svetlana]

Hi,

Maybe you want to try the simplification approach as outlined below

1. assign constant to both values.

DEFINE_SPECIFIC_HEAT(specific_heat_paddy, T, Tref, h, yi)
{
//real cp_paddy, M;
//Thread *t;
//cell_t c;
//M = C_UDMI(c, t, 0);
//cp_paddy = 1110 + 44.8 * (M / (M + 1));
// *h = cp_paddy * (T - Tref);

*h = 115.0; // sensible enthalpy
return 100.0; // specific heat
}

2. If this works, define variables M, c, t.

DEFINE_SPECIFIC_HEAT(specific_heat_paddy, T, Tref, h, yi)
{
real M;
//real cp_paddy, M;
Thread *t;
cell_t c;
//M = C_UDMI(c, t, 0);
//cp_paddy = 1110 + 44.8 * (M / (M + 1));
// *h = cp_paddy * (T - Tref);

*h = 115.0; // sensible enthalpy
return 100.0; // specific heat
}

3. If this works, assign a value to M.

DEFINE_SPECIFIC_HEAT(specific_heat_paddy, T, Tref, h, yi)
{
real M;
//real cp_paddy, M;
Thread *t;
cell_t c;
M = C_UDMI(c, t, 0);
//cp_paddy = 1110 + 44.8 * (M / (M + 1));
// *h = cp_paddy * (T - Tref);

*h = 115.0; // sensible enthalpy
return 100.0; // specific heat
}


4. Define cp_paddy


DEFINE_SPECIFIC_HEAT(specific_heat_paddy, T, Tref, h, yi)
{
//real M;
real cp_paddy, M;
Thread *t;
cell_t c;
M = C_UDMI(c, t, 0);
//cp_paddy = 1110 + 44.8 * (M / (M + 1));
// *h = cp_paddy * (T - Tref);

*h = 115.0; // sensible enthalpy
return 100.0; // specific heat
}

5. Assign a value to cp_paddy,

DEFINE_SPECIFIC_HEAT(specific_heat_paddy, T, Tref, h, yi)
{
//real M;
real cp_paddy, M;
Thread *t;
cell_t c;
M = C_UDMI(c, t, 0);
cp_paddy = 1110 + 44.8 * (M / (M + 1));
// *h = cp_paddy * (T - Tref);

*h = 115.0; // sensible enthalpy
return 100.0; // specific heat
}

6. Assign a value to h,

DEFINE_SPECIFIC_HEAT(specific_heat_paddy, T, Tref, h, yi)
{
//real M;
real cp_paddy, M;
Thread *t;
cell_t c;
M = C_UDMI(c, t, 0);
cp_paddy = 1110 + 44.8 * (M / (M + 1));
*h = cp_paddy * (T - Tref);
// *h = 115.0; // sensible enthalpy
return 100.0; // specific heat
}

7. Start returning cp_paddy instead of the constant,

DEFINE_SPECIFIC_HEAT(specific_heat_paddy, T, Tref, h, yi)
{
//real M;
real cp_paddy, M;
Thread *t;
cell_t c;
M = C_UDMI(c, t, 0);
cp_paddy = 1110 + 44.8 * (M / (M + 1));
*h = cp_paddy * (T - Tref);
// *h = 115.0; // sensible enthalpy
return cp_paddy; // specific heat
}

[D.M]

Originally Posted by estevaotolentino

Hi,

I'm having some problems with the DEFINE_SPECIFIC_HEAT macro. I need to vary the specific heat of a solid over time, where the equation that varies this property is in function of M (moisture content of a bed of grains). The variable M is defined outside of the DEFINE_SPECIFIC_HEAT macro (it is defined inside another macro called DEFINE_ADJUST), then I need to fetch this variable M through memory C_UDMI (c, t, 0), where M is stored in the UDM 0 memory. Here is my macro:

DEFINE_SPECIFIC_HEAT(specific_heat_paddy, T, Tref, h, yi)
{
real cp_paddy, M;
Thread *t;
cell_t c;
M = C_UDMI(c, t, 0);
cp_paddy = 1110 + 44.8 * (M / (M + 1));
*h = cp_paddy * (T - Tref);
return cp_paddy;
}

I'm interpreting the UDF. When I initialize the problem, an error message appears:

"Error: received a fatal signal (Segmentation fault).
Error Object: #f
/solve/initialize/initialize-flow"

Could anyone help me solve this problem? Thanks.

i guess this error happens because of the first iteration and define_adjust, try to define your parameter M with the define_execute_at_end and see if the error happens again,

[estevaotolentino]

Thank you for the reply D.M and Светлана. I tried what you suggested, but that did not solve my problem.

I sent a message to Ansys support and received the following response: "The DEFINE_SPECIFIC_HEAT macro does not allow as much customization as the DEFINE_PROPERTY function. There is no way to fetch a cell value in the DEFINE_SPECIFIC_HEAT macro, unfortunately."

There is also a similar question on the Ansys support portal. Following is the answer from Ansys:

"Why is the cell data not available in the DEFINE_SPECIFIC_HEAT UDF?
The UDF for specific heat is called as part of the material definition in the solver where the cell is not available. This is a limitation for this macro and its Real Gas equivalent.
Cp is by definition ∂H/∂T, so for it to have other dependencies is not expected by the solver. Since Cp and enthalpy definitions are critical to the stability of the energy equation, a strict format has to be defined so that the UDF doesn't cause instabilities.The only way to vary the Cp at the cell level is via the variation of the mass fractions of a mixture. This is what most users are trying to emulate when they are using a UDS, UDM or similar value to vary Cp. They should instead add a new species (which is not much more expensive than a UDS) to follow the material flow. The species could be a copy of the original species, but with a different Cp. Then, the Cp can be varied by varying the proportions of the species, potentially via source terms driven by UDM values."

[m__]

Originally Posted by estevaotolentino

"Why is the cell data not available in the DEFINE_SPECIFIC_HEAT UDF?
The UDF for specific heat is called as part of the material definition in the solver where the cell is not available. This is a limitation for this macro and its Real Gas equivalent.
Cp is by definition ∂H/∂T, so for it to have other dependencies is not expected by the solver. Since Cp and enthalpy definitions are critical to the stability of the energy equation, a strict format has to be defined so that the UDF doesn't cause instabilities.The only way to vary the Cp at the cell level is via the variation of the mass fractions of a mixture. This is what most users are trying to emulate when they are using a UDS, UDM or similar value to vary Cp. They should instead add a new species (which is not much more expensive than a UDS) to follow the material flow. The species could be a copy of the original species, but with a different Cp. Then, the Cp can be varied by varying the proportions of the species, potentially via source terms driven by UDM values."

estevaotolentino, did you end up using the species method they suggested? I am running into a similar limitation to you with my analysis and this seems like the only way I can resolve it