[denis.sanga@hotmail.it]

Goodmorning,
I'm trying to write an UDF able to read the residual values with the simple extract of code found online:

...
for (nw=0; nw<MAX_EQNS; ++nw)
{
if (strlen(DOMAIN_EQN_LABEL(domain,nw))>0)
{
Message0("eq %s: %g\n", DOMAIN_EQN_LABEL(domain,nw), DOMAIN_RES(domain,nw)/DOMAIN_RES_SCALE(domain,nw));
}
}
...

When I compare the results printed in the log file from fluent and the residuals printed by udf there are many mistakes: in particular seems that the position "nw" of the residual label in the DOMAIN_EQN_LABEL(domain,nw) list is not the same of the corresponding residual number in the DOMAIN_RES(domain,nw)/DOMAIN_RES_SCALE(domain,nw) list.

For a simple case I obtain something like this:
nw=0 --> DOMAIN_EQN_LABEL=continuity; DOMAIN_RES=correspond to continuity as expected
nw=1 --> DOMAIN_EQN_LABEL=nothing; DOMAIN_RES=correspond to x_velocity
nw=2 --> DOMAIN_EQN_LABEL=x-velocity; DOMAIN_RES=correspond to y_velocity
nw=3 --> DOMAIN_EQN_LABEL=y-velocity;DOMAIN_RES=correspond to z_velocity
nw=4 --> DOMAIN_EQN_LABEL=z-velocity; DOMAIN_RES=correspond to k
nw=5 --> DOMAIN_EQN_LABEL=energy; DOMAIN_RES=correspond to omega
nw=6 --> DOMAIN_EQN_LABEL=nothing; DOMAIN_RES=correspond to energy
...
nw=8 --> DOMAIN_EQN_LABEL=k; DOMAIN_RES=-nan
nw=12 --> DOMAIN_EQN_LABEL=omega; DOMAIN_RES=inf


Can someone help me to understand how to read the correct residual number associated with the correct quantity?

Thank you very very much
Best regards
DS

[AlexanderZ]

I don't know the reason of such behavior of these macros.
But what I can suggest is:
you have both names of equations that you are interested in and you have names of residuals which are under other nw pointer.
So you may make a loop through the list of residual names and check whether the name is equal to the name you expected, or not.
So you will create our own order in the list, which corresponds to equation list.

hope, my idea is clear
best regards

[blackmask]

The index for the residuals is continuous whereas the index for the labels is not. You have to use the mapping to get the correct index for labels. The code below is for your reference.

DEFINE_ADJUST(res_list,domain)
{
    int eqn, cmpt, nw;
    real scaled_res;

    Message("Residuals for iteration %g\n",count2[nres-1]);
    for (eqn = 0; eqn < DOMAIN_NUMEQN(domain); ++eqn) {
        nw = DOMAIN_EQNS(domain, eqn);
        if(strlen(DOMAIN_EQN_LABEL(domain,nw))>0)
        {
            Message("[%d] %s equation:", nw, DOMAIN_EQN_LABEL(domain,nw));
            scaled_res = DOMAIN_RES(domain,eqn)/DOMAIN_RES_SCALE(domain,eqn);
            Message("%g\n", scaled_res);
        }
    }
}

[denis.sanga@hotmail.it]

Thank you all very very much. I just tried with a simple case the blackmask suggestions and works. I will try with more complex cases.

Thank you all really much
DS