[lei]

Hi everyone,

I used the UDF macro DEFINE_PROPERTY(name, c,t) to model the absorption coefficient (Alpha) for gas and soot, the absorption coefficient is a function of gas composition and temperature, and the soot's contribution is calculated using the correlation:

Alpha(soot) = 1232.4*C_R(c,t)*C_POLLUT(c,t,4)*(1.0+0.00048*(T-2000.0));

Then the total Absorption coefficent = Alpha(gas)+Alpha(soot);

The UDF works well for the gas absorption coefficient calculation, however, when I use the variable for soot's mass fraction "C_POLLUT(c,t,4)", this variable gives all zero value in the domain, and thus the soot's effect on absorption coefficient cannot be added into the total absorption coefficient.

I checked the soot mass fraction, and it can be seen reasonably in post-processing.

Some information for my case:
*version: FLUENT 12.1.4
*Problem: propane combustion using a UDF for radiation property
*Model used: SST k-omega, DO, user defined WSGG model, EDC, one-step soot formation model

Thank you for your comments and helps!

[lei]

In case you encounter the same problem. This problem has been solved with helps from a great FLUENT engineer. Use variable C_POLLUT(c,t,IT_SOOT_START) instead, which is the soot mass fraction in (kg/kg).

[Junjun Guo]

Originally Posted by lei

In case you encounter the same problem. This problem has been solved with helps from a great FLUENT engineer. Use variable C_POLLUT(c,t,IT_SOOT_START) instead, which is the soot mass fraction in (kg/kg).

Hi lei,

I try the C_POLLUT(c,t,IT_SOOT_START), but it still doesn't work!
No matter I use C_POLLUT(c,t,4) or C_POLLUT(c,t,IT_SOOT_START), the Fluet can't run with a error "Process 48890: Received signal SIGSEGV". Once I delete the C_POLLUT, UDF works well. I also have added the #include "sg_mem.h"

[AlexanderZ]

Ansys fluent customization manual
look for Example: Soot Mass Rate

seems like you need memory check like:

if(NNULLP(t)) /*This check is must*/
{
soot_yi = C_POLLUT(c,t,EQ_SOOT);
nuc = C_POLLUT(c,t,EQ_NUCLEI);
}

best regard

[Junjun Guo]

Thank you Alexander! I try it, but it is the same. There is no problem with UDF compilation. Once calculated, an error is reported.
Here is my udf:

#include "udf.h"
#include "sg_mem.h"
#include "sg_pollut.h"
#include <stdlib.h>
#include <stdio.h>
#include <math.h>
#include <pdf_props.h>

/*Calculation for absorption coefficient */
DEFINE_GRAY_BAND_ABS_COEFF(Smith_gray_band_abs,c,t,nb)
        {
                real abs_coeff = 0.00001;
                real CO2_molf           ;
                real H2O_molf           ;
                real xi[100],yi[100]    ;
                real soot_yi=1.0E-06     ;
                real rho=C_R(c,t)        ;
                real Tcell=C_T(c,t)      ;
                real KT[4]               ; /* absorption coefficient for each gray gas*/
                real kT[4]               ;
                real b1=1230.0           ;
                real bT=4.8E-04          ;
                int  j,num=3             ; /* number of gray gases (4)  */

                real Kji[2][3] =       /* polynomial coefficients */
                {
                        { 0.4303, 7.055, 178.1 },
                        { 0.4201, 6.516, 131.9 }
                } ;

                if (NNULLP(t))
                  {
                    soot_yi=C_POLLUT(c,t,EQ_SOOT);
                  }
                Pdf_XY(c,t,xi,yi) ;
                CO2_molf=xi[10]   ;
                H2O_molf=xi[5]    ;
                /* Find absorption coefficient for each region */
                for (j = 0 ; j < num ; j++ )
                 {
                        KT[j]=Kji[1][j];
                        kT[j]=(CO2_molf+H2O_molf)*KT[j] + b1*rho*soot_yi*(1.0+bT*(Tcell-2000.0));
                 }

                KT[3] = 0.000001;
                kT[3] = KT[3]+b1*rho*soot_yi*(1.0+bT*(Tcell-2000.0));

                switch (nb)
                        {
                                case 0 : abs_coeff = kT[0]; break;
                                case 1 : abs_coeff = kT[1]; break;
                                case 2 : abs_coeff = kT[2]; break;
                                case 3 : abs_coeff = kT[3]; break;
                        }
                return abs_coeff;
        }

[AlexanderZ]

Frankly speaking I don't know what to suggest
try

soot_yi = C_POLLUT(c,t,12);

also, you may try do use UDMI to plot C_POLLUT(c,t,12)

C_UDMI(c,t,0) = C_POLLUT(c,t,12);

don't forget to allocate memory for udmi in Fluent GUI

best regards

[Junjun Guo]

Still face the same error. But if I enbale the NOx model, I can use C_POLLUT(c,t,0) to get the mass fraction of NO.
By th way, Why is 12 in C_POLLUT?

[AlexanderZ]

file sg_pollut.h
inside you have following

/* Scalar equations solved for pollutants */
typedef enum {
    EQ_NO = 0,
    EQ_HCN,
    EQ_NH3,
    EQ_N2O,
    EQ_UREA,
    EQ_HNCO,
    EQ_NCO,
    EQ_SO2,
    EQ_H2S,
    EQ_SO3,
    EQ_SH,
    EQ_SO,
    EQ_SOOT,
    EQ_NUCLEI, /*This ID:13 has been used in rfcell.scm to identify this eq-name. In case of modification, make sure to update the scheme file also*/
    EQ_CTAR,
    EQ_HG,
    EQ_HGCL2,
    EQ_HCL,
    EQ_HGO,
    EQ_CL,
    EQ_CL2,
    EQ_HGCL,
    EQ_HOCL
} EQ_Pollut;

where EQ_SOOT has number 12

as I undestand, it is possible to use scheme command to get this value

(rpgetvar 'pollut_soot/eq_start)

may be it is mandatory to use one of the following macros to get soot values:
DEFINE_SOOT_MASS_RATES
DEFINE_SOOT_OXIDATION_RATE
or some special model should be switched on in Fluent GUI. I have no experience about it.
Post the answer, when you'll find solution.

best regards

[Junjun Guo]

Thank you!

After I change the soot model from MOM to two-steps, the C_POLLUT works.