[weirdsda]

Hi all:
I have been frustrated for a long time to try to make a UDF with a previous step variable.
I would like to access the previous step value of C_STRAIN_RATE_MAG(c,tc), for the current calculation. I tried UDMI and UDSI_M1 but I could not get it right.

I would really appreciate if anyone knows how to do it?

[kar1209]

I want to provide convective boundary condition at my domain outlet and I am trying to do that by relating the current time step U velocity for the cells in the cell thread on the outlet face with the U velocity of the cells from the previous time step. I tried using the F_U_M1 macro. But I am getting "segmentation error". Can anybody help me?

[sbaffini]

Pretty sure that M1 and M2 variables only exist for cell values, not face ones. You need to save current face variables in a F_UDMI and access at the next time step before updating them. There is a post hanging around about this very specific task (using UDMI to store variables from previous time steps).

[kar1209]

Thank you so much. I was able to find the thread regarding using UDMs for accessing variable values at previous time steps.

Previous time step variable to be used in UDF

I have one more doubt. I see that in all programs, F_UDMI is used inside the begin_f_loop(). In that case can a UDM location store multiple values corresponding to multiple faces in a face thread?

[sbaffini]

Yes, note that this:

F_UDMI(f,t,i)

is the way F_UDMI is accessed, where f is the local face index, t is the face thread and i is the (i-1)-th F_UDMI variable (i starts at 0).

[kar1209]

Originally Posted by kar1209

Thank you so much. I was able to find the thread regarding using UDMs for accessing variable values at previous time steps.

Previous time step variable to be used in UDF

I have one more doubt. I see that in all programs, F_UDMI is used inside the begin_f_loop(). In that case can a UDM location store multiple values corresponding to multiple faces in a face thread?

What I meant to say is that if I want to calculate the value at current time step using value at previous time step and then assign it to the faces using F_PROFILE(), can I use 2 different looping macros - one for calculating the next time step value in the current time step and another looping macro for assigning the values in the next time step?

for example:

/*for specifying u velocity*/

begin_f_loop(f,t)
{
F_PROFILE(f,t,i) = F_UDMI(1); /*values to be assigned to boundary face in the next time step*/
}
end_f_loop(f,t)
begin_f_loop(f,t)
{
F_UDMI(1) = (.....) + F_U(f,t); /*calculation in the current time step*/
}
end_f_loop(f,t)

The above code is valid only if UDM can store multiple values. Otherwise how do I go around this? Please help.