[teethfish]

Hi, I want to modify the viscosity in K-W SST model in Fluent 13.0. But I met a problem that my UDF cannot initiate and it shows that:
Error:
FLUENT received fatal signal (ACCESS_VIOLATION)
1. Note exact events leading to error.
2. Save case/data under new name.
3. Exit program and restart to continue.
4. Report error to your distributor.
Error Object: #f

I have tried initialise without using "compute from" ....but the error was still there, I pasted my UDF here, if anone can help me with that?

#include "udf.h"

DEFINE_TURBULENT_VISCOSITY(user_mu_t,c,t)
{
real mu_t;
real a;
real rho;
real k;
real d;

a=(C_R(c,t)-0.5542)/997.6458;
rho = 0.5542+a*a*997.6458;

k = C_K(c,t);
d = C_D(c,t);
mu_t = M_keCmu*rho*k*k/d;
return mu_t;
}

[akm]

If you are working with k-w turbulence model, the dissipation rate C_D(c,t) will not be available in the UDFs as it is not calculated by the solver. Instead, specific dissipation rate is calculated C_O(c,t).

For the purpose of your udf calculation, you can covert \omega to \epsilon as under:
http://www.cfd-online.com/Wiki/Speci...ssipation_rate

[teethfish]

Wow, thank you for your help, you really helped me a lot!
And I have one more question, when i use this UDF:

DEFINE_TURBULENT_VISCOSITY(user_mu_t,c,t)
{
real mu_t;
real a;
real rho;
real k;
real d;

a=(C_R(c,t)-0.5542)/997.6458;
rho = 0.5542+a*a*997.6458;

k = C_K(c,t);
d = C_D(c,t);
mu_t = M_keCmu*rho*k*k/d;
return mu_t;
}

with realizable k-epsilon model, the results diverged, but it worked well with the RNG-kepsilon model, are they actually different?

Originally Posted by akm

If you are working with k-w turbulence model, the dissipation rate C_D(c,t) will not be available in the UDFs as it is not calculated by the solver. Instead, specific dissipation rate is calculated C_O(c,t).

For the purpose of your udf calculation, you can covert \omega to \epsilon as under:
http://www.cfd-online.com/Wiki/Speci...ssipation_rate

[micro11sl]

Hi teethfish,
Did you get your problem sorted?
I have some problems using the modified turbulent viscosity problem as well. I guess, although you hook a same udf to different turbulence models, there're something different more than the udf. The boundary condition, wall treatment, may be different. So a possible circumstance is the solution behavior is different between two distinct turbulence model with an identical udf.

I am still investigating this.

Sheng