[tsagaro]

Hi everybody,

I am simulating a system that includes a user defined scalar that is bounded (it cannot be higher than 1). Therefore, I have written the following DEFINE_ADJUST code, compiled (built and loaded) and hooked. No error is given. However, it does not work (the result is the same as with no DEFINE_ADJUST source code). I am using Fluent 13.0.
I would appreciate your help.



#include "udf.h"
DEFINE_ADJUST(limit_UDS,d)
{
Thread *t;
cell_t c;
{
begin_c_loop_c(t,d)
if(C_UDSI(c,t,0)>1.)
C_UDSI(c,t,0)=1.
end_c_loop(c,t)
}
}

[gearboy]

Originally Posted by tsagaro

Hi everybody,

I am simulating a system that includes a user defined scalar that is bounded (it cannot be higher than 1). Therefore, I have written the following DEFINE_ADJUST code, compiled (built and loaded) and hooked. No error is given. However, it does not work (the result is the same as with no DEFINE_ADJUST source code). I am using Fluent 13.0.
I would appreciate your help.



#include "udf.h"
DEFINE_ADJUST(limit_UDS,d)
{
Thread *t;
cell_t c;
{
begin_c_loop_c(t,d)
if(C_UDSI(c,t,0)>1.)
C_UDSI(c,t,0)=1.
end_c_loop(c,t)
}
}

It's dangerous to do so. Because during the iteration, Fluent will give the UDS value. In your adjust function, you force the UDS value to your value. And In Fluent's iteration period, Fluent set UDS to its value. What will be the ultimate value?

Therefore, forcing a UDS value in adjust function only should be used to rule temporary value which is unreasonable, like upper temperature limit 5000K in Fluent. You can't alway force the UDS to your value during the whole iteration. I mean you should force UDS value only when the value is out of reasonable range.

[tsagaro]

So, is there any way to limit the value of a UDS like temperature in the "limits" option in the "run calculation" dialog box?

[gearboy]

Originally Posted by gearboy

It's dangerous to do so. Because during the iteration, Fluent will give the UDS value. In your adjust function, you force the UDS value to your value. And In Fluent's iteration period, Fluent set UDS to its value. What will be the ultimate value?

Therefore, forcing a UDS value in adjust function only should be used to rule temporary value which is unreasonable, like upper temperature limit 5000K in Fluent. You can't alway force the UDS to your value during the whole iteration. I mean you should force UDS value only when the value is out of reasonable range.

I think you should read more about the UDF Manual. Here is the example.

#include "udf.h"
DEFINE_ADJUST(limit_UDS,d)
{
Thread *t;
cell_t c;
thread_loop_c(t, d) /*loops over all cell threads in domain*/
{
begin_c_loop(c, t) /* loops over cells in a cell thread */
{
if(C_UDSI(c,t,0)>1.)
{
C_UDSI(c,t,0)=1.
}
}
end_c_loop(c, t)
}
}

[tsagaro]

I am sorry but I forgot to write the

thread_loop_c(t, d)

line in my first message. So, my define_adjust is equal to yours but simply does not work.

In any case, thank you for your replies.

[gearboy]

Originally Posted by tsagaro

I am sorry but I forgot to write the

thread_loop_c(t, d)

line in my first message. So, my define_adjust is equal to yours but simply does not work.

In any case, thank you for your replies.

You can try using DEFINE_EXECUTE_AT_END. I think it will work.

[pablo]

Have the same problem.
Did you find some solution?
Do not why hooking define_adjust doesn't work.
I appreciate whatever clue to solve this...

best regards

[tsagaro]

No, I did not find the solution.

[antonk67]

Maybe it is not too straightforward, but consider the following:
1) If the scheme for transport equation is monotonous, the solution should be within limits set by boundary conditions, if we disregard the source term.
2) Accordingly, it is much better to formulate the source term in such a manner that it will preserve the solution within the necessary limits (or decrease the iteration parameter/time step).
3) If you adjust the solution, it will go out of limits upon the next iteration in a most unpleasant manner anyway due to (2).

Just take a look at http://www.cfd-online.com/Forums/flu...ds-values.html