[MsRuby]

Dear all,

I am quite new with UDF so I am here to seek for some advice. Much appreciated if someone can tell me what is basic problem with my code.

Basically, I am writing an UDF to update the properties of a porous zone, then use these variables to calculate the mass source for a transient case. I use:
1. DEFINE_INIT to set the initial values of three variables: density, volatile fraction and formation rate of volatile, save them in three UDMI (0,1,2).
2. DEFINE_ADJUST to calculate the mass loss due to the reaction rate of volatile, then recalculate the density and volatile fraction. Two new variables are saved in two other UDMI (4,5)
3. DEFINE_SOURCE to calculate the formation rate of volatile, I save this varabile in one UDMI (3)
4. DEFINE_EXECUTE_AT_END, I set the UDMI (0,1,2) to UDMI (4,5,3) so that in the next time step (t+deltat), step 2 and 3 can use the new input of density, volatile fraction, and product formation rate.

The code works very well with DEFINE_INIT, DEFINE_ADJUST, DEFINE_SOURCE, but when I hook up DEFINE_EXECUTE_AT_END, this error occurs (this time I use interpretation because of some software issues)

================================================== ============================

Node 0: Process 9976: Received signal SIGSEGV.

================================================== ============================
MPI Application rank 0 exited before MPI_Finalize() with status 2
The fl process could not be started.

Below you can find my code:

#include "udf.h"

/* input */

DEFINE_INIT(udm,d)
{
Thread *t;
cell_t c;
thread_loop_c(t,d)
{
begin_c_loop(c,t)
{
C_UDMI(c,t,0)=625.0; /*density (initial)*/
C_UDMI(c,t,1)=0.74; /*vol_fraction (initial)*/
C_UDMI(c,t,2)=0.0; /*rate of reaction*/
C_UDMI(c,t,3)=0.0;
C_UDMI(c,t,4)=0.0;
C_UDMI(c,t,5)=0.0;
}
end_c_loop(c,t)
}

/* calculate the rate of reaction */
DEFINE_SOURCE(vol_release,c,t,dS,eqn)
{
const real a=1.63e5, e=9.04e4;
real rho, vol_frac;
real k,source;
rho=C_UDMI(c,t,0);
vol_frac=C_UDMI(c,t,1);
k=a*exp(-e/8.314/C_T(c,t));
source=k*vol_frac*rho;
dS[eqn]=rho*vol_frac;
C_UDMI(c,t,3)=source;
return source;
}
/* re-calculate density and vol_fraction */
DEFINE_ADJUST(properties,d)
{
Thread *t;
cell_t c;
real delta_t=CURRENT_TIMESTEP;
thread_loop_c(t,d)
{
begin_c_loop(c,t)
{
real rho,rate,vol_frac;
real mass_0, mass_1, mass_vol, mass_loss;
/* data from the previous time step */
rho=C_UDMI(c,t,0);
vol_frac=C_UDMI(c,t,1);
mass_0=rho*C_VOLUME(c,t);
rate=C_UDMI(c,t,2);
/*calculate data for this time step */
mass_vol=mass_0*vol_frac;
mass_loss=rate*delta_t*C_VOLUME(c,t);
mass_1=mass_0-mass_loss;

/*recaculate the density and vol fraction*/
C_UDMI(c,t,4)=mass_1/C_VOLUME(c,t); /*density*/
C_UDMI(c,t,5)=(mass_vol-mass_loss)/mass_1; /*volatile fraction */
}
end_c_loop(c,t)
}
}

DEFINE_EXECUTE_AT_END(execute_at_end) /* only at the end of time step */
{
Domain *d;
Thread *t;
cell_t c;
thread_loop_c(t,d)
{
begin_c_loop(c,t)
{
C_UDMI(c,t,0)=C_UDMI(c,t,4); /* 0,1,2 is input for next step, 3,4,5 is for this step */
C_UDMI(c,t,1)=C_UDMI(c,t,5);
C_UDMI(c,t,2)=C_UDMI(c,t,3);
}
end_c_loop(c,t)
}
}

[obscureed]

Hi MsRuby,

The domain pointer is never given a value in the DEFINE_EXECUTE_AT_END. You need

domain = Get_Domain (1);

I haven't checked anything else in detail -- at first sight, it all looks OK.

I temporarily worried that your user-defined reaction will occur every iteration (in DEFINE_ADJUST) -- but it operates on the previous timestep every time, so I think it's OK. Calculating it only once in DEFINE_EXECUTE_AT_END seems like a possible idea.

Good luck!
Ed

[MsRuby]

Originally Posted by obscureed

Hi MsRuby,

The domain pointer is never given a value in the DEFINE_EXECUTE_AT_END. You need

domain = Get_Domain (1);

I haven't checked anything else in detail -- at first sight, it all looks OK.

I temporarily worried that your user-defined reaction will occur every iteration (in DEFINE_ADJUST) -- but it operates on the previous timestep every time, so I think it's OK. Calculating it only once in DEFINE_EXECUTE_AT_END seems like a possible idea.

Good luck!
Ed

Thank you Ed, the problem is solved nicely