I am trying to build a turbulence model that requires distance from the wall for each cell in the fluid.
I used C_WALL_DIST(c,t). The FLUENT interprets this without errors but solution initialization fails. If anyone else has written any UDF to find the wall distance, please let me know.
What is the error message you get during/after initialization?
I don't have much experience with UDF writing, and I never used it to create a turbulence model, but I found a kwsst turbulence model UDF on ResearchGate during a quick search (file is attached). It provides some turbulence source terms and uses the C_Wall_Dist mactro as well. Again, it's not my own file - I just forward what I found. Unfortunately, no information is provided on how to hook up the udf in a simulation and I did only a quick test with a simple case - successful compilation and initialization, but I got a "segmentation violation" error after startring the calcution. However, maybe it's helpful to resolve your problem anyway.
Best regards,
Stephan
Dear Stephan,
I too tried the same UDF a while ago to see if it works. I also get the same error (segmentation violation) both in the KW model you just sent and in the model that I am trying to build.
The C_WALL_DIST macro is no longer in the latest FLUENT UDF manual. I think they stopped using it.
The next obvious way is to find centroid of each fluid cell and then find the distance. This will be easy for 2D benchmark cases, however testing a 3D case with a lot of objects in the enclosure will not be practical with this method.
Please tell me if you or your colleagues have better suggestions.
Ok, in this case I just reproduced your bad experiences :-(
I'm using the latest Fluent version and think that it shouldn't compile if the macro was not longer available (I had some similar examples in the past, can't remember excatly what it was). Yes, the C_CENTROID macro could be an alternative. Perhaps, the ANSYS customer support could be of help, too...
Good luck.
Stephan
Hi Patrick,
Any luck in finding a solution to the problem you were having?
I also need to calculate wall distance and C_WALL_DIST is no longer available.
Thanks,
Ali
In case it is still an open question, here is how to define cell wall distance:
#define USE_UDM 0
#include "udf.h"
#include "prox.h"
void
cpySVartoUDM(Domain *domain, Svar sv, int udm)
{
size_t realsize = sizeof(real);
real *svpointr = NULL;
real *udmpoint = NULL;
Thread *thread = NULL;
Domain *supdom = DOMAIN_SUPER_DOMAIN(domain);
if (NULLP(supdom))
{
supdom = domain;
}
if (sg_udm nelements;
memcpy(udmpoint, svpointr, numbytes);
}
}
}
DEFINE_ON_DEMAND(ondmd_cwdtoudm0)
{
Domain *dom = Get_Domain(1);
Thread *thr = NULL;
Alloc_Storage_Vars(dom, SV_RTMP_0, SV_NULL);
Calc_Cell_Wall_Distance(dom, SV_RTMP_0);
cpySVartoUDM(dom, SV_RTMP_0, USE_UDM);
Free_Storage_Vars(dom, SV_RTMP_0, SV_NULL);
}
DEFINE_ON_DEMAND(ondmd_acwdtoudm0)
{
Domain *dom = Get_Domain(1);
Thread *thr = NULL;
Alloc_Storage_Vars(dom, SV_RTMP_0, SV_NULL);
Calc_Cell_Wall_Distance_Alternative(dom, SV_RTMP_0);
cpySVartoUDM(dom, SV_RTMP_0, USE_UDM);
Free_Storage_Vars(dom, SV_RTMP_0, SV_NULL);
}
Dear Subhasish,
where is Calc_Cell_Wall_Distance itself?
Dear Colleagues, please review this option:
BOUNDARY_FACE_GEOMETRY(f,t,A,ds,es,A_by_es, dr0);
delta = NV_DOT(A,dr0) / NV_MAG(A);
Thanx!
Hi Gregory,
The Calc_Cell_Wall_Distance is in FLUENT's own function libary; these functions are used in the regular C_WALL_DIST macro as well. However, in FLUENT 18.0 it does give an error when I compile the above function; it mentions the Calc_Cell_Wall_Distance as an unresolved external. It hence appears the library is not properly linked; any ideas on how to fix this?
Some addition; I got things to work (FLUENT 18.0) with the script referenced here: https://www.cfd-online.com/Forums/fluent-udf/63211-calculate-distance-wall-neighbourig-cell-centroid-2.html
It uses Calc_Cell_Wall_Distance_New to calculate the wall distance. Calc_Cell_Wall_Distance_Alternative also works with this script; the older Calc_Cell_Wall_Distance is not available anymore. The answers between the 2 methods hardly differ (I believe "new"is based on nearest node, and "alternative"on face center)
Do note the referenced script does not free temporary values as the above does. That's something to add, potentially.
Now, I am still puzzled as to why this script from CFD online does accept the Calc_Cell.. macro, whereas the script referenced above gives an unresolved external error.
One can compute the wall distance by displaying the contours of the 'Cell Wall Distance' at the boundaries.
Go to Results > Graphics > Contours > Contours of Mesh > Cell Wall distance > select surface > Compute
- Badges
- Science topic