[miguel.soto]

Hi everybody,

I´m triying to perform a conjugate heat transfer simulation, where a solid zone is modeled as an energy source, cooled by an air stream. The goal is to get the temperature distribution both at the solid and fluid zones.

I´ve realized that convergence of the energy residuals is too slow, likely due to the difference between thermal conductivity coefficients for air and aluminium.

I tried to start iterations disabling energy equation and then after the flow and turbulence are well converged, to enable the energy equation, but still facing problems due to the fact that when i set high under-relaxation factors for the energy, the solution starts to diverge.

CFX includes an option well known by its user´s called "solid time scale factor" , that allows to set different time steps for solid and fluid zone, thus improving energy solution convergence.

Does anybody knows if fluent includes an option similar to CFX "time scale factor?

Best Regards

Miguel

[vasava]

I think fluent does not have such support.

[miguel.soto]

Thanks for the answer Vasava,

That means fluent is less suitable than CFX to perform conjugate heat transfer simulations, as normally metals have higher thermal conductive coefficients than fluids, and "solid time scale factor" is very useful to increase convergence speed.

Remark: my mesh cotains 1.285.497 nodes, 6.295.780 faces, is expected to have a maximum solid temperature of 373K, within an ambient temperature of 300K. Each 1.000 iterations the maximum temperature at the solid zone increases 1ºC, by using an energy under relaxation factor of 0.9.
My computer takes 6 hours to perform 1.000 iterations, so the prediction is to get a reasonable result in about 18-20 days (if temperature increases linearly, everybody knows this asumption is not correct, but at least gives me an approach for the minimum calculation time), what is absolutely unfeasible.

[vasava]

I have not used CFX ever. But it is nice to know that CFX has such facility.