[Raj123]

I'm trying to simulate the periodic heat transfer where outlet temperature of fluid should be taken as upstream bulk temperature for next repeat. So this way continuous heat loss to the boundary solid walls maintained at constant wall temperature for varying thermal properties including density for helium flow. Please check the code, im facing the segmentation error and fluent is closing itself.

/* periodic_bulkT_UDF.c
- Works for flow in negative Z direction
- Outlet periodic face zone id = 15
- Uses Fluent-provided T-dependent properties (rho, cp, k, mu)
- Mass-weighted bulk temp computed on outlet faces each timestep (adjust)
- Relaxation applied to avoid oscillations
- Profile UDF returns the scalar bulk temp to Periodic Bulk Temperature
*/

#include "udf.h"

#define OUTLET_ZONE_ID 9
/* relaxation: 0 < alpha <= 1. 0.3 is typical starting value */
#define RELAX_ALPHA 0.30

/* global storage */
static real T_bulk_outlet = 300.0;
static int initialized = 0;

/* ADJUST UDF: compute mass-weighted outlet bulk temperature each timestep */
DEFINE_ADJUST(compute_outlet_bulkT, domain)
{
Thread *t_out;
face_t f;
real mdot_sum = 0.0;
real Tm_sum = 0.0;

t_out = Lookup_Thread(domain, OUTLET_ZONE_ID);

begin_f_loop(f, t_out)
{
cell_t c0 = F_C0(f, t_out);
Thread *tc0 = THREAD_T0(t_out);

/* Temperature at the face */
real Tface = F_T(f, t_out);

/* density from cell adjacent to face (Fluent updates rho(T) automatically) */
real rho = C_R(c0, tc0);

/* face w-velocity (z component). For negative z flow use -w */
real w_face = F_W(f, t_out);

/* face area vector and magnitude */
real area_vec[ND_ND];
F_AREA(area_vec, f, t_out);
real Af = NV_MAG(area_vec);

/* mass flow through face (positive for flow in -Z, so use -w_face) */
real mdot = rho * (-w_face) * Af;

if (mdot > 0.0)
{
mdot_sum += mdot;
Tm_sum += mdot * Tface;
}
}
end_f_loop(f, t_out);

if (mdot_sum > 1e-20)
{
real T_new = Tm_sum / mdot_sum;

if (initialized)
T_bulk_outlet = RELAX_ALPHA * T_new + (1.0 - RELAX_ALPHA) * T_bulk_outlet;
else
{
T_bulk_outlet = T_new;
initialized = 1;
}
}
/* else keep previous value */
}

/* PROFILE UDF: supply bulk temperature to periodic inlet Bulk Temperature field */
DEFINE_PROFILE(periodic_bulkT_profile, thread, position)
{
face_t f;

begin_f_loop(f, thread)
{
/* set the profile value (Fluent expects this scalar) */
F_PROFILE(f, thread, position) = T_bulk_outlet;
}
end_f_loop(f, thread)
}

[Thore850]

Hello,

without knowing more details about your setup its hard to tell why a segmentation error ocurrs. Normally these errors appear if a Macro tries to access something it is not supposed to access.

Are you using a multiphase formulation? If yes, this could be the problem. In that case you currently provide the mixture domain oulet id, and try to access the temperature and density of the species, which in the case of a multiphase formulation is to be found in a sub-domain, leading to a segmentation error.

Best regards

Thore

[Raj123]

No I'm not using multiphase, I'm trying to simulate helium flow through offset fin geometry with thermophysical properties varying with temperature. I'm having constant wall temperature boundary condition of 80K and inlet helium temperature at 300 K. Basically I'm trying to simulate the cool down of helium gas from 300 K, using periodic boundary conditions along the flow for offset fin geometry.

I have doubt, can we use periodic boundary conditions along the streamwise direction for variable density flows due to temperature ? In fluent I'm getting a warning saying that
"periodic boundary conditions with non zero mass flow rate / pressure gradient is not valid for variable density flows"