[3014214149]

Hi, guys,

I met a strange problem.

I wrote a udf and used UDSI to store variable C_K and C_DUDX

CODE:

DEFINE_ADJUST(NNtraining, d)
{
Thread *t;
cell_t c;
real k;
real epsilon;
real C_mu;
C_mu=0.09;

thread_loop_c(t,d)
{
begin_c_loop(c,t)
{
k=C_K(c,t);
epsilon=C_D(c,t);
C_UDSI(c,t,0)=k;
C_UDSI(c,t,1)=C_DUDX(c,t);
}
end_c_loop(c,t)
}
}

In theory, when I look up contours. The results of Scalar-0 and turbulent kinetic energy should be the same. And the results of Scalar-1 and dX-velocity/dx should be the same. But in fact, they are different.

That's why? Could anybody help me?

[AlexanderZ]

to monitor variable you do probably want to use C_UDMI instead of C_UDSI (which is scalar)

if the difference still is here I can't predict

[3014214149]

Sad.

The difference is still here.

[pakk]

How big is the difference that we are talking about? Orders of magnitude, or after the sixth decimal?