[jpina]

Dear all,

I'm modeling a polymer flow, whose density is temperature and pressure dependance and, moreover, has a discontinuity in the density in the melting point. The graphic looks like this:



The UDF compiles and loads successfully and is the following:

#include "udf.h"
#include<string.h>
#include<stdio.h>
#include <math.h>
#include <stddef.h>
#include <stdlib.h>
#include <nrutil.h>

DEFINE_PROPERTY(cell_pvt2,cell,thread)
{
real T = C_T(cell, thread);
real P = C_P(cell,thread);
real B;
real V;
real V0;
real rho;
real B1s=8.54e-4;
real B2s=1.59e-7;
real B3s=2.99e8;
real B4s=1.71e-3;
real B1l=8.54e-4;
real B2l=5.62e-7;
real B3l=1.83e8;
real B4l=3.99e-3;
real Tg=144+273;

if (T>Tg) {
V0=B1l+B2l*(T-Tg);
B=B3l*exp(-B4l*(T-Tg));
}
else{
V0=B1s+B2s*(T-Tg);
B=B3s*exp(-B4s*(T-Tg));
}

V=V0*(1-0.0894*log(1+P/B));

rho=1/V;

if (P > 0.0001)
{
return rho;
}
else{
return 1000;
}
}

However, in the second iteration I give divergence problems and a floating point exception.

I'm modeling the inlet of a polymer into a cavity filled with air, having a 0 atm outlet and a pressure inlet that pushes the polymer.

I have tried the UDF without having the two regions but happens the same. I have tried different solvers and relaxing the UDF without success.

Somebody has a hint?

Thanks al ot!