[deada]

Hi,

I want to run a eulerian multiphase simulation with a primary (p1) and two secondary phases (s1 and s2). In every cell of my domain, where the strain rate of secondary phase s1 exceeds a certain threshold mass shall be transferred from s1 to s2. From what I read in the fluent theory guid the best way to realize that is to use DEFINE_LINEARIZED_MASS_TRANSFER() or DEFINE_MASS_TRANSFER(). (btw. I'm using fluent 16.2 right now)

Since this is the first time I am dealing with an udf, the right usage and implementation is unknown to me. Here is a first udf, that I tried to compile in fluent:

#include "udf.h"

DEFINE_LINEARIZED_MASS_TRANSFER(strain_source,cell,thread,from_index,from_species_index,to_index,to_species_index,d_mdot_d_vof_from,d_mdot_d_vof_to)
{

real strain_rate_mag = C_STRAIN_RATE_MAG(cell,thread);
real strain_rate_thresh = 100;

Thread *s1 = THREAD_SUB_THREAD(thread, from_index);
Thread *s2 = THREAD_SUB_THREAD(thread, to_index);
real vof_L = C_VOF(cell,s1);
real vof_S = C_VOF(cell,s2);

real m_dot, dstrain, m_source;
m_dot = 0.0;
m_source = 0.0;

if (strain_rate_mag > strain_rate_thresh)
{
    dstrain = strain_rate_mag - strain_rate_thresh;
    dstrain = MAX(dstrain, 1e-4);
    m_dot = 0.1;
    m_source = m_dot*vof_L;
    *d_mdot_d_vof_from = m_dot;
    *d_mdot_d_vof_to = -m_dot;
}

return m_source;
}

There are some lines in the code which I don't really understand by now and it also doesn't compile. So here are my questions:

1.) I read that cell and thread are given to the define macro by the solver. Do I still somehow have to define them before? How does it know the right thread? And does it automatically loop over all cells or do i have to but the macro inside a loop of some kind?

2.) I only want mass being transfered from s1 which has the ID 3 to s2 with the ID 4 and my secondary phases are not made up of multiple species. How do I tell that to the udf? Before the DEFINE_LINEARIZED_MASS_TRANSFER() line, should I include something like:

int from_index,from_species_index,to_index,to_species_index;
from_index=3;
from_species_index=-1;
to_index=4;
to_species_index=-1;

eg. I assume the line Thread *s1 = THREAD_SUB_THREAD(thread, from_index); to take the mixture thread pointer and give me a pointer to the first secondary phase. But therefore I need to it the ID of s1 somehow, don't I?

3.) When I click on build in the Compiled UDFs window I get the following text line feedback:

udf_names.c and user_nt.udf files in 3ddp are upto date.
(system "copy "C:\PROGRA~1\ANSYSI~1\v162\fluent"\fluent16.2.0\src\udf\makefile_nt.udf "libsel\win64\3ddp\makefile" ")
        1 file(s) copied.
(chdir "libsel")(chdir "win64\3ddp")# Generating ud_io1.h
strain_source.c
# Generating udf_names.c because of strain_source.obj
udf_names.c
# Linking libudf.dll because of makefile user_nt.udf udf_names.obj strain_source.obj
Microsoft (R) Incremental Linker Version 14.10.25017.0
Copyright (C) Microsoft Corporation.  All rights reserved.

   Bibliothek "libudf.lib" and Objekt "libudf.exp" are created.
strain_source.obj : error LNK2019: Reference to non-resolved external symbol "__imp__store_vof_norm" in function "_strain_source".
C:\PROGRA~1\ANSYSI~1\v162\fluent\fluent16.2.0\win64\3ddp\fl1620.lib : warning LNK4272:library computer type "x64" in conflict with destination computer type "X86"
libudf.dll : fatal error LNK1120: 1 non-resolved external

What confuses me about that feedback is that it states that the destination computer type is X86 but I'm using an x64. Furthermore, I don't know what "__imp__store_vof_norm" is.

Hints and help to any of my questions is highly appreciated!
Thanks, Thies

[deada]

A little update:
I have solved the question 3 (the x86 vs x64 issue) by following the instructions given here:
https://www.cfd-online.com/Wiki/Flue...ows_7_64bit.3F
I had to install the more recent Windows SDK instead of the .NET Framework 2.0 as suggested in the wiki and added the path C:\Program Files\Microsoft SDKs\Windows\v7.0\Include to the enviroment variables. Now I am able to compile the udf without any errors. However when I run the case without including mass transfer, everything is fine. But as soon as I activate it I get an segmentation fault. When I activated the mass transfer I also learned that I set the from_index and to_index manually in the GUI so I think question 2 is also solved.

But I still do not have an idea what causes the segmentation fault and if my udf does what I intend it to do

Error:  received a fatal signal (Segmentation fault).

Error:  received a fatal signal (Segmentation fault).
Error Object: #f

[deada]

Hi again,
by stepwise shortening of my .c file i found that the segmentation fault originated from the first line

real strain_rate_mag = C_STRAIN_RATE_MAG(cell,thread)

because I needed to use a subthread to thread as an input. Changing it to s1 and moving it below the THREAD_SUB_THREAD command solved the problem